Grazer Philosophische Studien. Vol. 73 2006 (German Edition)

Grazer Philosophische Studien INTERNATIONALE ZEITSCHRIFT FÜR ANALYTISCHE PHILOSOPHIE

GEGRÜNDET VON Rudolf Haller HERAUSGEGEBEN VON Johannes L. Brandl Marian David Leopold Stubenberg

VOL 73 - 2006

Amsterdam - New York, NY 2006

Die Herausgabe der GPS erfolgt mit Unterstützung des Instituts für Philosophie der Universität Graz, der Forschungsstelle für Österreichische Philosophie, Graz, und wird von folgenden Institutionen gefördert: Bundesministerium für Bildung, Wissenschaft und Kultur, Wien Abteilung für Wissenschaft und Forschung des Amtes der Steiermärkischen Landesregierung, Graz Kulturreferat der Stadt Graz

The paper on which this book is printed meets the requirements of “ISO 9706:1994, Information and documentation - Paper for documents Requirements for permanence”. Lay out: Thomas Binder, Graz ISBN: 978-90-420-2232-4 ISSN: 0165-9227 © Editions Rodopi B.V., Amsterdam - New York, NY 2006 Printed in The Netherlands

INHALTSVERZEICHNIS

TABLE OF CONTENTS

Abhandlungen

Articles

Werner SAUER: Die Einheit der Intentionalitätskonzeption bei Brentano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

Tanja PIHLAR: Zur Theorie der Vorstellungsproduktion („Grazer“ Gestalttheorie I: France Weber) . . . . . . . . . . . . . . . . . . . . . . . . . .

27

Thane Martin NABERHAUS: Does Husserl Have an Argument against Representationalism? . . . . . . . . . . . . . . . . . . . . . . . . .

43

Torsten WILHOLT: Lost on the Way from Frege to Carnap: How the Philosophy of Science Forgot the Applicability Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

69

John PRESTON: Janik on Hertz and the Early Wittgenstein . . . . . .

83

Friedrich Christoph DOERGE: Re-Definition and Alston’s ‘Illocutionary Acts’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97

Michael VEBER: Not too Proud to Beg (the Question): Why Inferentialism Cannot Account for the A Priori . . . . . . . . . . . .

113

Anthony Robert BOOTH: Can there Be Epistemic Reasons for Action?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

133

Achim LOHMAR: Why Content Relativism Does Not Imply Fact Relativism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

145

Stephen HETHERINGTON: So-Far Incompatibilism and the So-Far Consequence Argument . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

163

Christoph JÄGER & Anne BARTSCH: Meta-Emotions . . . . . . . .

179

Diskussionen

Discussions

William H. HANSON: The Paradox of Nonbeing . . . . . . . . . . . . . .

205

Dan LÓPEZ DE SA: Flexible Property Designators . . . . . . . . . . .

221

Buchnotizen

Critical Notes

Alfred NORDMANN: Wittgenstein’s Tractatus: An Introduction. Cambridge: Cambridge University Press, 2005 (Marie McGINN)

231

Severin SCHROEDER: Wittgenstein: The Way Out of the Fly-Bottle. Cambridge: Polity, 2006 (Brian ARMSTRONG) . . . . . . . . . . . .

236

Willem A. deVRIES: Wilfrid Sellars. Chesham: Acumen, 2005 (Johannes HAAG) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

239

Geert KEIL and Udo TIETZ (eds.): Phänomenologie und Sprachanalyse. Mentis: Paderborn, 2006 (Paul LIVINGSTON) . . . . . . .

244

Wolfgang HUEMER: The Constitution of Consciousness: A Study in Analytic Phenomenology. New York: Routledge, 2005 (David R. CERBONE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

249

Fabrice CORREIA: Existential Dependence and Cognate Notions. München: Philosophia Verlag, 2005 (E. Jonathan LOWE) . . . . .

255

Albert CASULLO: A Priori Justification. New York: Oxford University Press, 2003 (Martin GRAJNER) . . . . . . . . . . . . . . . . .

259

R. Jay WALLACE, Philip PETTIT, Samuel SCHEFFLER, and Michael SMITH (eds.): Reason and Value. Themes from the Moral Philosophy of Joseph Raz. Oxford: Clarendon Press, 2004 (Herlinde PAUER-STUDER) . . . . . . . . . . . . . . . . . . . . . . . . . .

265

Eingelangte Bücher / Books Received . . . . . . . . . . . . . . . . . . . .

269

Grazer Philosophische Studien 73 (2006), 1–26.

DIE EINHEIT DER INTENTIONALITÄTSKONZEPTION BEI BRENTANO Werner SAUER Universität Graz Summary The objective of this paper is to refute the widely held view that in the wake of his so-called reistic turn Brentano subjected his notion of intentionality to a deep-going revision, viz., that he turned from an ontological account of the intentional object by way of identifying it with the thought-of-thing, i.e., the intentional correlate, or by way of attributing to it a peculiar sort of existence, to a non-ontological account thereof. It will be shown that neither the pre-reistic Brentano espoused anything of an ontological account of the intentional object in that he both distinguished it sharply from the intentional correlate and definitely rejected the idea of there being different sorts of existence, and it will be argued that the apparently ineradicable inclination to ascribe to the pre-reistic Brentano an ontological account of the intentional object stems from ignoring the Aristotelian background of Brentano’s thinking about relations.

Vorbemerkung Nach vorherrschender Meinung in der Brentano-Forschung hat die „reistische Wende“ oder „Abkehr vom Nichtrealen“ im Denken Brentanos auch seine Auffassung der Intentionalität der psychischen Phänomene tiefgehend affiziert: Ihr zufolge hatte Brentano in vorreistischer Zeit dem sogenannten intentionalen Objekt einen besonderen ontologischen Status zugeschrieben, um dann im Gefolge der reistischen Wende diese Auffassung fallen zu lassen und dergestalt seine Konzeption der Intentionalität gründlich zu revidieren. Da für jenen ontologischen Status des intentionalen Objekts in der Regel das sogenannte intentionale Korrelat, das Gedachte als solches oder einfach das Gedankending, wie wir es nennen können,1 herangezogen 1. Nach einer Bemerkung Brentanos im Brief an Anton Marty vom 1.3.1906 (ANR p. 147f.)

wird, sei gleich hier eine die Natur dieses Nichtrealen besonders klar herausstellende Passage aus dem vorreistischen Fragment ,Das Seiende im Sinne des Wahren‘ angeführt: Es ist nicht möglich, daß es ein A-Denkendes gibt, ohne daß es ein gedachtes A gibt, und umgekehrt. Aber man kann darum nicht sagen, der A-Denkende sei das von ihm gedachte A. Die beiden Begriffe sind nicht identisch, sondern korrelativ. Keinem kann in der Wirklichkeit etwas entsprechen, ohne daß dem andern etwas in der Wirklichkeit entspricht. Aber nur der eine ist der Begriff eines Wesenhaften [Realen], … der andere ist der von etwas, was nur, indem jenes gewirkt wird, als begleitendes Seiendes mit entsteht und fortbesteht, bis jenes endet … Sagt man: eben, indem man es dem wirklichen A entgegenstellt, gibt man zu erkennen, daß das gedachte A nichts Wahres und Wirkliches ist: so ist zu erwidern: Keineswegs! Es kann etwas recht wohl etwas Wahres und Wirkliches sein, ohne ein wirkliches A zu sein. Es ist … ein wirkliches gedachtes A. (WE p. 31)

Zu dieser Stelle, die wir einfach die Gedankending-Stelle nennen wollen, ist nur der terminologische Punkt anzumerken, daß Brentano hier, wie aus der Entgegensetzung von ,wirkliches A‘ und ,wirkliches gedachtes A‘ erhellt, ,wirklich‘ nicht wie anderswo (siehe Abschnitt I) für ,real‘, sondern für ,existierend‘ als Synonym verwendet. Hier wird gegen die vorherrschende Meinung geltend gemacht werden, daß Brentano in der vorreistischen Zeit ebenso wenig irgendeine ontologische Konzeption des intentionalen Objekts vertrat wie nach der reistischen Wende, und es wird gezeigt werden, daß die hartnäckige Tendenz, dem vorreistischen Brentano eine derartige Konzeption zuzuschreiben, aus einem Brentanos Lehre vom Relativen betreffenden Mißverständnis resultiert. I. In der berühmten Intentionalitätspassage in der Psychologie vom empirischen Standpunkt schreibt Brentano den psychischen Phänomenen als ihr Charakteristikum das zu, über den Ausdruck ,ens rationis‘, den er sonst auch selbst für das Nichtreale überhaupt verwendet (z.B. ANR p. 115). – Wie ,Gedankending‘ ganz allgemein für das intentionale Korrelat, werden wir, wie Brentano selbst, auch mit ,denken‘ und den damit gebildeten Ausdrücken in der Regel ganz allgemein das intentionale Verhalten meinen.

2

was die Scholastiker des Mittelalters die intentionale (wohl auch mentale) Inexistenz eines Gegenstandes genannt haben, und was wir, obwohl mit nicht ganz unzweideutigen Ausdrücken, die Beziehung auf einen Inhalt, die Richtung auf ein Objekt … oder die immanente Gegenständlichkeit nennen würden. Jedes [psychische Phänomen] enthält etwas als Objekt in sich … (PeS I p. 124f.)

Was die Auslegung der Intentionalitätspassage anlangt, so ist ein ausgesprochen merkwürdiges Phänomen zu beobachten: Wiewohl Brentano selbst ausdrücklich klarstellt, daß er den Ausdruck ,intentionale (mentale) Inexistenz‘ nicht als seinen eigenen terminus technicus verstanden wissen will, haben sich Kommentatoren vielfach gerade auf diesen Ausdruck als Schlüssel für das Verständnis der Intentionalitätspassage, ja der Intentionalitätsauffassung des vorreistischen Brentano überhaupt konzentriert. Blicken wir dazu zuerst auf Chisholm, wie es bei seiner herausragenden Stellung in der Einbringung des Brentanoschen Denkens in die neuere Philosophie nur angemessen ist. Nach Chisholms einflußreicher Auslegung der Intentionalitätspassage enthält dieselbe two different theses: one, an ontological thesis about the nature of certain objects of thought and of other psychological attitudes [which Brentano was later to abandon]; the other, a psychological thesis, implying that reference to an object is what distinguishes the mental or psychological from the physical.2

Das Problem, das Brentano nach Chisholm mit der ontologischen These lösen wollte, ist natürlich das, worauf das Denkende sich intentional beziehe, wenn das, was es denkt, nicht existiert: Worauf sich jemand bezieht, der ein Einhorn denkt, cannot be an actual unicorn, since there are no unicorns. According to the doctrine of intentional in-existence, the object of the thought about a unicorn is a unicorn, but a unicorn with a mode of being (intentional inexistence, immanent objectivity, or existence in the understanding) that is short of actuality but more than nothingness and that ... lasts for just the length of time the unicorn is thought about. 2. Roderick M. Chisholm, ,Intentionality‘, in Encyclopedia of Philosophy (ed. P. Edwards) IV, New York & London 1967, p. 201. Einfügung in eckigen Klammern aus der Parallelpassage in ,Brentano on Descriptive Psychology and the Intentional‘, Phenomenology and Existentialism (eds M.H. Mandelbaum & E.N. Lee), Baltimore 1967, p. 6. Nächstes Zitat: ,Intentionality‘, p. 201.

3

Diese bewußtseinsimmanente Entität wird von Chisholm sodann ohne weiteres mit dem gedankendinglichen Korrelat des Denkaktes gleichgesetzt; diese Identifikation dann auch später in der Einleitung zur Deskriptiven Psychologie, die, gemeinsam mit dem Mitherausgeber Wilhelm Baumgartner verfaßt, in diesem Punkt ohne Zweifel seine eigene Meinung wiedergibt: Die Lehre von der Intentionalität, wie sie hier [in DP] niedergelegt wird, ist im wesentlichen die der ersten Ausgabe der Psychologie vom empirischen Standpunkt (1874). Das intentionale Objekt ist stets „immanent“; es ist etwas Nicht-Reales oder Unwesenhaftes … Es ist ein nicht-reales Korrelat eines Denkens, das es als sein Objekt hat (DP p. xiii).

Demgemäß ist also das Gedankending Korrelat des Denkaktes und zugleich sein ihm immanentes Objekt. Es lohnt, zeitlich zurückzugehen. Eben diese Gleichsetzung des intentionalen Objekts mit dem intentionalen Korrelat nahm schon Alois Höfler 1905 auf dem Fünften Internationalen Psychologenkongreß in Rom vor. Doch Brentanos eigene Stellungnahme dazu war gänzlich negativ. Bei der Rede vom immanenten (intentionalen) Objekt, schreibt er am 17.3.1905 an Marty, ist es nicht meine Meinung gewesen, daß das immanente Objekt = „vorgestelltes Objekt“ sei. Die Vorstellung hat nicht „vorgestelltes Ding“, sondern „das Ding“, also z.B. die Vorstellung eines Pferdes nicht „vorgestelltes Pferd“, sondern „Pferd“ zum (immanenten, d.h. allein eigentlich Objekt zu nennenden) Objekt. (ANR p. 119f., WE p. 87f.)

Weiters stellt er dann in dem Brief die in seiner Sicht bestehende Übereinstimmung in der Sache mit Aristoteles heraus (siehe Abschnitt III), um darauf auf die absurden Konsequenzen der von Höfler vorgenommenen Identifikation aufmerksam zu machen: Das „gedachte Pferd“ als Objekt genommen, wäre Gegenstand der inneren Wahrnehmung, die das Denkende wahrnimmt, wenn dies mit dem Gedachten ein Paar Korrelative bildete, da Korrelative ohne einander nicht wahrnehmbar sind. Das, was als primäres Objekt empfunden oder vom Verstand … gedacht wird, ist aber doch nicht Gegenstand der inneren Wahrnehmung. Entweder müßte ich der primären Vorstellungsbeziehung gar kein Objekt ... zugeschrieben haben, oder ich konnte … [es] nicht = „gedachtes Objekt“ gleichgesetzt haben. Ich protestiere also gegen die mir angedichtete Albernheit. (ANR p. 120f., WE p. 88f.)

4

Da der reale Denkakt und das nichtreale Gedankending Korrelate sind und für Korrelate gilt, „daß das Eine nicht bloß nicht ohne das Andere sein, sondern auch nicht ohne dasselbe erkannt werden kann“ (VE p.45), ist ebenso wie der reale Denkakt sein nichtreales Korrelat Objekt der inneren Wahrnehmung; deren Gebiet, so der vorreistische Brentano, „zeigt also reale und nicht-reale Phänomene“ (DP p.131; Hervorhebung W.S.). Wäre nun das intentionale Objekt identisch mit dem intentionalen Korrelat, so müßte, da dieses Korrelat Objekt der inneren Wahrnehmung ist, alles, worauf wir uns intentional beziehen, Objekt der inneren Wahrnehmung d.h. des sekundären Bewußtseins sein, so daß wir in diesem Fall nichts zum primären Objekt der intentionalen Bezugnahme und somit überhaupt kein primäres Bewußtsein hätten: Ein Pferd oder ein Einhorn wären, weil Objekte des primären Bewußtseins, unmöglich zu denken. Der getreue Oskar Kraus hatte seine Schwierigkeiten mit Brentanos Protest gegen die ihm angedichtete Albernheit. In einer Anmerkung zur Gedankending-Stelle schreibt er: Hier ist also expressis verbis ausgesprochen, daß jedem Denken das Gedachte als Korrelat entspricht. Und obwohl Brentano immer betont hat, daß wir nicht ein vorgestelltes Pferd, sondern ein Pferd vorstellen, hat er doch die Korrelation zwischen dem Pferdevorstellenden und dem vorgestellten Pferd gelehrt. Ich bemerke dies darum, weil in dem unten abegdruckten Briefe vom 17.3.1905 diese Lehre ihm so fremd geworden ist, daß es ihm fraglich erscheint, sie je vorgetragen zu haben. (WE p. 177 Anm. 33)

Daraus spricht pure Konfusion gepaart mit schierer Ratlosigkeit. Brentano bestreitet in dem Brief mit keinem Wort, früher die Lehre vom intentionalen Korrelat, sondern vielmehr, besagte Gleichsetzung vertreten zu haben, und in der Gedankending-Stelle wiederum setzt er das intentionale Korrelat mit keinem Wort zum intentionalen Objekt in Bezug. Es ist nur Kraus selbst, der einen solchen herstellt: Und so ergibt sich das Durcheinander, daß er einerseits Brentanos Selbstinterpretation in dem Brief folgt, andererseits aber die Gedankending-Stelle als Beleg dafür nimmt, daß Brentano früher doch nicht wirklich das gelehrt habe, was er in dem Brief gelehrt zu haben behauptet, um sich dann reichlich hilflos in einen angeblichen Anflug von Vergeßlichkeit auf seiten Brentanos zu flüchten. Offenbar sah sich Kraus, wiewohl durch des Meisters ipsissima verba in keiner Weise dazu genötigt, irgendwie unter dem Zwang stehen, das intentionale Objekt beim vorreistischen Brentano mit dem intentionalen Korrelat zusammenzubringen. In einer weiteren Anmerkung in Wahrheit

5

und Evidenz nimmt er dann die Gleichsetzung explizit vor. In einer Kritik an Christoph Sigwarts Konzeption der Existenz aus 1889 sagt Brentano, deren Verkehrtheit erweise sich recht einfach auch an folgendem Satze: ein wirklicher Zentaure existiert nicht, ein vorgestellter Zentaure aber existiert, und zwar so oft, als ich ihn vorstelle. (WE p.48)

Indem hier ,wirklich‘ für ,real‘ steht und vom intentionalen Objekt keine Rede ist, paßt das also vollkommen zur Gedankending-Stelle. Kraus aber kommentiert: Diese Stelle ist heute historisch interessant. Es geht aus ihr mit aller Klarheit hervor, daß Brentano 1889 die Lehre von der Intentionalität als mentale Inexistenz des Objektes verstanden hat und den vorgestellten Zentauren als Korrelat des „Zentaurvorstellens“ existieren ließ. (WE p. 183 Anm. 51)

Auch ein ausdrücklicher Rückverweis auf die Gedankending-Stelle und seine Anmerkung zu ihr („vgl. … S. 31 und Anm. 33“) fehlt nicht. In Chisholms ontologischer Deutung des intentionalen Objekts beim vorreistischen Brentano liegt also nichts anderes vor als eine Bekräftigung der schon von Kraus vertretenen Auffassung. Die Kraus-Chisholm-Deutung, wie wir sie demnach nennen können, ist gleichsam so etwas wie eine communis opinio geworden, die sich z.B. auch bei Barry Smith findet: Ihm zufolge haben beim vorreistischen Brentano Sachen wie „experienced colours and sounds as existing in the mind as parts of consciousness … a diminished sort of existence, an existence ,in the mind‘“, so daß die Intentionalität eine Beziehung „between two mental entities“3 sei, nämlich zwischen dem realen Denkakt und seinem gedankendinglichen Korrelat: Every mental phenomenon … includes something as object within itself. Consciousness is an intentional relation … But what is the relation between act and object? The act, Brentano says, is real, its object (the horse insofar as it is thought, the redness insofar as it is seen) is non-real.

Konnte Brentano wirklich gesagt haben, was er dann so vehement als eine ihm angedichtete Albernheit zurückwies? Sehen wir uns also die Sache näher an.

3. Barry Smith, Austrian Philosophy. The Legacy of Franz Brentano, Chicago & La Salle, Ill., 1994, p. 44. Nächstes Zitat: p. 55.

6

II. Spricht man von einer ontologischen Auffassung des intentionalen Objekts, so sind hinsichtlich dessen, was an ihr das spezifisch Ontologische ausmachen und so ihre Bezeichnung rechtfertigen kann, diese beiden Möglichkeiten zu unterscheiden: (A) Denke ich ein Ding A, das, wenn es existiert, eine bewußtseinsunabhängige Entität ist, so ist das intentionale Objekt das psychisch modifizierte Gegenstück von A, also das Gedankending gedachtes-A, d.h. das intentionale Korrelat. (B) Ein- und dasselbe Ding kann in zwei verschiedenen Weisen existieren: einmal bloß „als Objekt des Denkens“, zum anderen auch „in Wirklichkeit“, so daß jedes Objekt des Denkens im Sinne der ersten Existenzweise existiert, das eine oder andere auch im Sinne der zweiten. Der Unterschied ist klar: (A) besagt, daß das Objekt des Denkens eine besondere Art von Entität sei, (B) hingegen, daß es auf eine besondere Art existiere, kurz, (A) betrifft in ,Es gibt ein F ‘ das ,F ‘, (B) hingegen das ,es gibt‘.4 Kraus und Chisholm, ebenso auch Smith, amalgamieren nun (A) und (B) derart, daß sie das Gedankending aus (A) mit der ersten Existenzweise aus (B) zu (C) Ein Gedankending zu sein impliziert, in der besonderen Existenzweise der bewußtseinsimmanenten Existenz (intentionalen Inexistenz) zu existieren, und umgekehrt verbinden; die Kraus-Chisholm-Deutung bedarf also keiner eigenen Erörterung über die von (A) und (B) hinaus. Beginnen wir mit (B). An der gegen Sigwarts Konzeption der Existenz gerichteten Stelle macht Brentano keinen Unterschied zwischen dem ,existiert‘ in ,Ein wirklicher [realer] Zentaur existiert nicht‘ und dem in ,Ein vorgestellter Zentaur existiert‘, desgleichen in der Gedankending-Stelle keinen zwischen ,wirklich‘ – dort, wie gesagt, nicht für ,real‘, sondern für ,existierend‘ stehend – in ,wirkliches gedachtes-A ‘ und in ,wirkliches A ‘. 4. Eine präzise Formulierung des Unterschieds, aber nicht in besonderer Hinsicht auf Brentano, findet sich bei A. N. Prior, Objects of Thought, Oxford 1971, p. 127.

7

Man kann es nicht besser sagen, als es in diesen Worten zur Stelle gesagt worden ist: On Brentano‘s account, … the difference between the actual A that is thought about and the thought-about-A is not that one has a mode of being different from that of the other. Both, he says, are actual. The difference is that the A is a different sort of entity from that of a thought-about-A.5

Und derselbe Befund ergibt sich bereits auch für 1874 daraus, wie in der Psychologie vom empirischen Standpunkt der Satz ,Ein Zentaur ist eine Erfindung der Poeten‘ als Existenzialsatz interpretiert wird.6 In dieser Hinsicht ist also Brentanos Position genau die, welche später dann in Quine einen Vorkämpfer finden sollte: Ganz analog dazu, daß das reale A sich nicht in der Existenzweise vom gedachten-A unterscheidet, sondern eben als ein Reales von einem gedankendinglichen Nichtrealen, verwirft Quine die Auffassung, daß die Anwendung des Wortes ,existiert‘ auf Abstrakta wie Klassen oder Zahlen einerseits und die auf Konkreta wie materielle Dinge andererseits zwei verschiedene Gebrauchsweisen des Wortes ,existiert‘ seien, als eine grundlose Verwandlung des Unterschieds zwischen verschiedenen Arten von Entitäten in einen Unterschied verschiedener Weisen der Existenz.7 Die Distinktion, um die es hier geht und die sich auch durch die beiden Fragen, was A ist und ob A ist, ausdrücken läßt, ist ein Eckstein in Brentanos ontologischem Denken. Ihre Basis ist die Unterscheidung von Aristoteles zwischen dem on kath’ hauto (ens per se), dem Seienden im eigentlichen Sinn d.h. gemäß den Kategorien, und dem on hôs alêthes, dem Seienden im Sinne des Wahren (Metaphysik [Met.] Δ.7). Die Quintessenz 5. Otis T. Kent, ,Brentano and the Relational View of Consciousness‘, Man and World 17 (1984), p. 32. 6. Brentano verteidigt hierbei seine These, daß alle Urteile letztlich Existenzialurteile seien, gegen Mill, für den ihre Verneinung aus einem Satz wie diesem erhellt, „A centaur is a fiction of the poets; where it cannot possibly be implied that a centaur exists, since the proposition itself expressly asserts that the thing has no real existence“ (A System of Logic I.iv.1). Brentanos Erwiderung darauf ist die: In der Tat, der „Satz ,ein Zentaur ist eine Erfindung der Poeten‘ verlangt … nicht, daß ein Zentaur existiere, vielmehr das Gegenteil. Allein er verlangt, um wahr zu sein, wenigstens, daß etwas anderes existiere, nämlich eine Fiktion der Poeten … [W]enn es keinen von den Poeten fingierten Zentauren gäbe, so wäre der Satz falsch; und seine Bedeutung ist tatsächlich keine andere als … ,es gibt einen von den Poeten fingierten Zentauren‘“ (PeS II p. 61 Anm.). Die Stelle ist in sich klar, nicht minder als die aus der Sigwart-Kritik oder die Gedankending-Stelle. 7. W. V. O. Quine, ,On What There Is‘, From a Logical Point of View, ²1961, repr. New York & Evanston 1963, p. 3; Word and Object, Cambridge, Mass., 1960, § 27, p. 131.

8

dieser zweiten Weise des Seienden ist in der Deutung durch Thomas von Aquin die, daß das in dieser Weise Seiende das ist, in bezug worauf die Bejahung der ,An est?‘-Frage eine wahre Aussage ist. In diesem Sinn kann auch von einer Privation wie der Blindheit gesagt werden, daß sie ist, indem nämlich die Aussage, daß etwas blind ist, wahr ist; das ens per se hingegen antwortet darüber hinaus auch auf die ,Quid est ?‘-Frage als Frage nach der quidditas d.h. der essentia.8 Das Seiende im Sinne des Wahren kann demnach nach Thomas ebenso gut das Seiende im Sinne des Existierenden genannt werden, wie es Brentano nun tatsächlich nennt.9 Der oben aus seiner Sigwart-Kritik zitierten Stelle mit dem realen Zentauren, der nicht existiert, und dem gedachten-Zentauren, der existiert, läßt er unmittelbar diese Worte folgen: Wem hier nicht der Unterschied des on hôs alêthes, d.h. im Sinne des Existierenden, vom on im Sinne des Dinglichen (Wesenhaften) klar wird … (WE p. 48)

Und ganz konform damit heißt es zu Beginn des Fragments ,Das Seiende im Sinne des Wahren‘: Aristoteles, da er die Bedeutungen des Seienden schied, unterschied das Seiende im Sinne des Wesenhaften … und Seiendes im Sinne des Wahren, tatsächlich Gegebenen [= Existierenden] … [Dessen] Begriff steht dem des Wesenhaften nicht so gegenüber, daß der eine den andern ausschlösse. Im 8. Siehe De Ente et Essentia cp. 1, Summa Theologiae [ST ] Ia qu. 48 art. 2 ad 2, und Quaestiones disputatae [Qu. disp.] de Malo qu. 1 art. 1 ad 19. 9. Brentanos Rezeption des Aristotelischen on hôs alêthes ist ganz vom Vermittler Thomas geprägt. In seiner Dissertation findet er bei Aristoteles zwei Varianten desselben. Nach der einen, für die er sich auf Met. E.4 (1027b18–23) beruft, wird es „von dem ganzen, fertig ausgesprochenen Urteile prädiziert“ (MB p. 34), d.h., es manifestiert sich im ,ist‘ in ,Es ist so, daß p‘, gleichgültig, ob ,p‘ eine Bejahung oder eine Verneinung ist. Nach der anderen, die er in Met. Δ.7 (1017a31–35) findet, ist es „das Sein der Kopula“ in der „wahren, affirmativen Behauptung“ (p. 37). Aus diesem Text die zweite Variante zu extrahieren, ist freilich ein recht künstlicher Zug. Klar ausgesprochen findet sie sich vielmehr bei Thomas, der das Seiende im Sinne des Wahren explizit an der compositio propositionis, deren Zeichen die Kopula ist, festmacht, wovon man über die Lehre vom existenzialen Gehalt der bejahenden Kopula zum Seienden im Sinne des Existierenden gelangt. Im späten Rückblick, hierbei gemäß seiner Urteilslehre nicht mehr von der Kopula sondern vom existenzialen ,ist‘ sprechend, sagt dann Brentano selbst, er sei von Thomas dazu gebracht worden, das ,ist‘ in den Sätzen ,Ein Baum ist‘ und ,Daß ein Baum ist, ist‘ “für gleichmäßig funktionierend zu halten” (ANR p. 291f.; cf. auch WE p. 129, 162f.). Dem späten Brentano geht es um die Erklärung dessen, wie er zur Annahme des Nichtrealen der Urteilsinhalte, wie daß ein Baum ist, gekommen war, in unserem Zusammenhang dagegen darum, wie bei ihm das Seiende im Sinne des Wahren zum Seienden im Sinne des Existierenden wurde, das als solches keine Differenzierung in Reales und Nichtreales kennt.

9

Gegenteil … Alles was ist, ist, insofern es ist, ein Seiendes im Sinne des Wahren, tatsächlich Gegebenen (WE p.30),

oder in der Ausdrucksweise des Vortrags ,Über den Begriff der Wahrheit‘ aus 1889: Seiendes in diesem Sinn ist alles, ob Reales oder Nichtreales, wofür „die bejahende Beurteilungsweise die passende ist“ (WE p. 24). In einem Wort, das Seiende im Sinne des Wahren oder des Existierenden ist eindeutig. Das bejahende Existenzzeichen, das existenziale ,ist‘ oder dergleichen (,es gibt‘‚ ,existiert‘), das in Brentanos Urteilslehre die bejahende Kopula als Zeichen der bejahenden Aussage ablöst, ist eindeutig, weil es Ausdruck des Anerkennens ist und dieses keine verschiedenen Weisen kennt („realhaftes“ versus „nichtrealhaftes“): Der Unterschied zwischen ,Reales ist‘ und ,Nichtreales ist‘ liegt gänzlich in den beiden Termen ,Reales‘ und ,Nichtreales‘. In diesem Sinn bemerkt Brentano 1889 gegen Windelband, wenn dieser sich wundere, wie er in Existenzialsätzen über Reales und über Nichtreales das existenziale ,ist‘ für gleichbedeutend nehmen könne …: so kann ich nur erwidern, daß, wer hierin nicht eine einfache Konsequenz meiner Auffassung vom Urteil erblickt, diese Lehre kaum erfaßt haben dürfte (WE p. 41),

worauf er an der Stelle damit fortfährt, daß es Aristoteles gar nicht ein[fällt], das „estin“, welches den Ausdruck der Vorstellung zum Ausdruck des Urteils ergänzt, und das „on hôs alêthes“, wie er es nennt, ähnlich wie das „on“ im Sinne einer Realität in verschiedene Kategorien … zu zerlegen,

d.h. verschiedene Weisen desselben zu statuieren: Denn ist das Existenzzeichen eindeutig, so sind es auch die Terme ,Existierendes‘ und ,Existenz‘, für dessen Sinn Brentano die „Reflexion auf das bejahende Urteil“ als Quelle namhaft macht (WE p. 45; barbarisch aber passend ,Anzuerkennensein‘ für ,Existenz‘ in PeS II p. 89). So gibt es beim vorreistischen Brentano sowenig wie beim reistischen einen Platz für (B), somit keinen Platz für eine Existenz „short of actuality but more than nothingness“ oder „a diminished sort of existence“, was in des vorreistischen Brentano eigenen Worten eine einfache Konsequenz seiner Urteilslehre ist. Es gilt also auch schon für den vorreistischen Brentano, was der reistische sagen wird: nämlich, „unfähig“ zu sein, einer Unterscheidung von Existenzweisen „überhaupt irgendwelchen Sinn abzugewinnen“ (PeS II p.137). In Brentanos Brief an Marty vom 17.3.1905 geht es aber auch gar nicht 10

um (B), sondern um (A). Werfen wir dazu zunächst einen kurzen Blick auf die Psychologie vom empirischen Standpunkt, und zwar auf folgendes Argument: Der Begriff Ton ist kein relativer Begriff. Wäre dies der Fall, so würde nicht das Hören ein sekundäres, sondern mit dem Tone zugleich das primäre Objekt des psychischen Aktes sein (PeS I p. 185).

Brentano argumentiert hier folgendermaßen: (i) Der Ton ist primäres Objekt des psychischen Aktes Hören-desTones, ein für Brentano zu jeder Zeit wahrer Satz. Wäre nun der Begriff des Tones ein relativer Begriff, so würde das heißen: (ii) Der Ton (als Objekt des Hörens-des-Tones) ist dasselbe wie der gehörte-Ton, so daß folgen würde: (iii) Der gehörte-Ton ist primäres Objekt des psychischen Aktes Hörendes-Tones, woraus für das reale Korrelat, das Hören-des-Tones, aufgrund des für Korrelate Geltenden sich (iv) Das Hören-des-Tones ist primäres Objekt des psychischen Aktes Hören-des-Tones ergeben würde: Also ist der Begriff des Tones kein relativer Begriff, d.h. (ii) ist falsch. Der ganze Unterschied zwischen diesem Argument und dem im Brief an Marty liegt in dem, daß Brentano dort, bei Substituierung des Ton-Beispiels, die wahre Prämisse (i) durch die in vorreistischer Zeit ebenfalls wahre Prämisse (i*) Der gehörte-Ton ist sekundäres Objekt des psychischen Aktes Hörendes-Tones ersetzt, um über (ii) weiterzugehen zu

11

(iii*) Der Ton (als Objekt etc.) ist sekundäres Objekt des psychischen Aktes Hören-des-Tones,

so daß der psychische Akt Hören-des-Tones kein primäres Objekt hätte. Beidemale geht es um die Herausstellung der absurden Konsequenzen der falschen Identitätsbehauptung (ii), unter Zugrundelegung einmal des primären, dann des sekundären Bewußtseins. Das mag für den Brentano von 1874 hinreichen.10 Für die spätere vorreistische Zeit ist das für unseren Punkt Wesentliche die Klassifikation der verschiedenen Weisen des Teil-Seins in der Deskriptiven Psychologie. Nach dieser Klassifikation zerfallen die Teile, die sich auf dem psychischen Gebiet finden, zunächst in ablösbare (in Wirklichkeit abtrennbare) und in distinktionelle (nur begrifflich abtrennbare) Teile. Bei den distinktionellen Teilen wiederum treten die distinktionellen Teile im eigentlichen Sinn den distinktionellen Teilen im modifizierenden Sinn bzw. „durch modifizierende Distinktion zu gewinnenden Teilen“ (DP p. 25) gegenüber. Unter den distinktionellen Teilen im eigentlichen Sinn unterscheidet Brentano nochmals mehrere Arten, von denen hier nur zu nennen sind die logischen Teile und die Teile des intentionalen Korrelatenpaares, welch letztere „zwei rein distinktionelle Teile“ sind, von denen „der eine real ist, der andere nicht“ (p. 21f.). Das intentionale Objekt dagegen ist die auf psychischem Gebiet vorfindliche Art der distinktionellen Teile im modifizierenden Sinn, und d.h., ihm kommt dieselbe Weise des Teil-Seins zu wie auf nicht-psychischem Gebiet dem im gewesenen-Ton enthaltenen Ton (mehr im übernächsten Absatz). Die Kluft zwischen intentionalem Objekt und intentionalem Korrelat zeigt sich also schon daran, daß in diesem Klassifikationsschema die beiden Sachen verschieden lokalisiert werden und nicht einmal unter dieselbe Gattung fallen. Brentano sieht offenkundig keinerlei Gefahr der Konfundierung des mereologischen Ortes des intentionalen Objekts mit dem des intentionalen Korrelats. Die Gefahr, die er sieht, ist vielmehr die, daß man es zu den logischen Teilen schlagen könnte. Die logischen Teile sind die Teile, die sich in der Differenzierung der Gattung zur Art zeigen, und es mag naheliegen, z.B. Empfinden als Gattung zu nehmen, die durch Farbe zu Sehen differenziert wird, so daß das intentionale Objekt zugleich Differenz wäre und, da wie Empfinden nicht Farbe so auch Farbe nicht Empfinden 10. Ein weiteres Beispiel wäre Brentanos Argumentation gegen die Doktrin des esse est percipi, PeS I p. 129–32; es ist im wesentlichen dieselbe, die später dann Russell in The Problems of Philosophy (1912), ch. 4, gegen den Idealismus vorbringen wird.

12

einschließt, zwei gegenseitig distinktionelle logische Teile vorlägen. Das kann nun freilich nicht der Fall sein, denn logische Teile sind einseitig distinktionelle Teile, was nichts anderes besagt, als daß die Differenz die Gattung als Teil einschließt, so daß sie, wie Brentano nach Aristoteles (Met. Z.12) geltend macht, gleich der Art ist; und darüber hinaus, wäre Farbe Differenz von Sehen, so müßte ein Akt des Sehens selbst farbig sein (cf. DP p. 26). Den Fingerzeig für die richtige Lokalisierung des intentionalen Objekts findet Brentano in der Bemerkung des Aristoteles, daß das Sehende (to horôn, d.h. hier: das Sinnesorgan, to aisthêtêrion) gewissermaßen (hôs) gefärbt sei (De Anima III.2 425b22f.). Von daher deutet Brentano, wie schon gesagt, das Verhältnis des psychischen Aktes zu dem in ihm enthaltenen Objekt in Entsprechung zum Verhältnis des gewesenen-Tones zu dem ihn ihm enthaltenen Ton. Ein gewesener-Ton ist kein Ton, sondern aufgrund des modifizierenden Attributs des Gewesenseins ein Nichtreales, das aber den Ton insofern in sich enthält, als wir nicht verstehen können, was ein gewesener-Ton ist, ohne zu verstehen, was ein Ton ist. Da nun aber der Ton kein eigentlicher distinktioneller Teil des gewesenen-Tones ist – sonst wäre dieser ein Ton –, kann er auch nicht durch ein eigentliches Distinguieren aus dem gewesenen-Ton gewonnen werden, “sondern nur durch ein modifizierendes Distinguieren” (DP p. 19), das sozusagen die Modifikation des Tones zum gewesenen-Ton wieder rückgängig macht: Und gerade so wird auch die Farbe aus dem Sehen durch modifizierende Distinktion als Teil gewonnen werden können … Und was vom realen Glied der intentionalen Beziehung gilt, das gilt auch von seinem nicht realen Korrelat. „Gesehene Farbe“ enthält gewissermaßen Farbe in sich, nicht als distinktionellen Teil im eigentlichen Sinn, sondern als einen durch modifizierende Distinktion daraus zu gewinnenden Teil. (p. 27)

Das intentionale Objekt ist also nicht nur als ein distinktioneller Teil im modifizierenden Sinn klassifikatorisch vom intentionalen Korrelat geschieden, es ist darüber hinaus auch noch selbst ein solcher Teil des letzteren. Das Klassifikationsschema, das sich gezeigt hat, in Übersicht:

13

ablösbare Teile

distinktionelle Teile

im eigentlichen Sinn

im modifizierenden Sinn, u.a.

logische Teile, ..., Teile des intentionalen Korrelatenpaares (Hören-des-Tones, gehörter-Ton)

Ton im Ton in den Teilen gewesenen-Ton, …, des intent. Korelatenpaares Hören-des-Tones und gehörter-Ton (intent. Objekt).

Verwechslung oder Identifizierung des intentionalen Objekts mit dem intentionalen Korrelat ist ausgeschlossen, und Brentanos Protest in dem Brief an Marty gegen die ihm zugeschriebene Albernheit ist vindiziert. III. Thomas unterschied zwischen einem esse naturale und einem esse intentionale, und der Einfluß des Aquinaten auf den frühen Brentano bedarf keines besonderen Nachweises: So mag man spekulieren, ob nicht hinter der Rede von der intentionalen Inexistenz in der Psychologie vom empirischen Standpunkt noch etwas vom Thomasischen esse intentionale steckt und von daher Licht auf die Intentionalitätsauffassung des frühen Brentano im Sinne einer ontologischen Deutung fallen könnte. Einen Versuch in diese Richtung hat Linda McAlister unternommen. Sie weist die Position (A) überzeugend zurück und argumentiert, sich hierin vom Haupttrend in der ontologischen Deutung absetzend, daß von der genannten Thomasischen Distinktion her so etwas wie (B) sich als die Intentionalitätsauffassung des frühen Brentano erweise.11 Zwar ist nach dem Bisherigen schon absehbar, daß das Ergebnis auf Sand gebaut sein muß, es ist aber auch zu zeigen, wie es auf Sand gebaut ist. Der Ausgangspunkt von McAlisters Darlegungen ist der folgende Passus in dem Brief an Marty vom 17.3.1905: Aristoteles sagt auch, daß die aisthêsis [Wahrnehmung] das eidos [Form] ohne die hylê [Materie] aufnehme … Hat er nicht wesentlich gedacht wie wir? (WE p. 88; für Aristoteles cf. De Anima II.12 424a18f.) 11. Linda McAlister, The Development of Franz Brentano’s Ethics, Würzburg & Amsterdam 1982, p. 23ff. Seitenangaben zu den Zitaten im folgenden Absatz im Text.

14

Ihr Kommentar dazu ist dieser: As Brentano seems to have understood the doctrine, though the immanent object in the case of someone seeing or thinking of a horse, is a horse, nonetheless, as immanent object it has a different ,mode of being‘ than it has when it is completely unrelated to a mind as object of a mental act (p. 25).

Übereinstimmung (“similarity”, p. 25) damit findet sie im Kommentar des Aquinaten zu De Anima, dafür die folgende Stelle anführend, an der Thomas das Aristotelische Aufnehmen der Form ohne Materie durch den Sinn kommentiert: Der Sinn nimmt die Form ohne Materie auf, denn die Form hat im Sinn Sein (esse) in anderer Weise als im Sinnending; im Sinnending hat sie nämlich natürliches Sein (esse naturale), im Sinn dagegen hat sie intentionales und mentales Sein (esse intentionale et spirituale). (In De An.,12 lib.II, lec.24, n.553)

An diese Stelle anknüpfend, charakterisiert sie dann das, was sie Brentanos “original doctrine of intentionality” (p.29) nennt, folgenderweise: So we are dealing here with two different modes of being which objects may have: actual existence and intentional existence … I may, for example, think about a white horse and there may actually be a white horse which I am thinking about. If so, this horse has actual existence, but it also has intentional existence because when I think about him he acquires, in addition to his actual existence, a kind of existence in my mind which he did not have previous to anyone‘s thinking about him. And what is more, even if there were no actually existing white horses at all, a white horse could still be the object of my thought, for a white horse begins to exist intentionally the instant I begin to think of it (p. 25).

Das wäre demnach die Lehre der Intentionalitätspassage in der Psychologie vom empirischen Standpunkt: In seiner Abhängigkeit vom Denkakt gleicht das weiße Pferd qua intentionales Objekt dem gedankendinglichen Korrelat, es ist aber nicht ein gedachtes-(weißes Pferd), das einfachhin existiert, sondern ein weißes Pferd einfachhin, das intentional existiert. Zunächst einmal, um damit zu beginnen, ist auf eine gewisse Gewaltsamkeit in der ganzen Argumentationslinie hinzuweisen. Weder spricht die Stelle in dem Brief an Marty von etwas, das mit Thomas ein esse intentionale 12. S. Thomae Aquinatis in Aristotelis librum de Anima commentarium (ed. M. Pirotta), Turin 1925.

15

und dazu noch des Objekts anstatt der Form zu nennen wäre, noch sagt Brentano darin irgendetwas über eine Übereinstimmung mit Aristoteles allein in früherer Zeit, denn die in der rhetorischen Frage am Ende der Stelle ausgesprochene Übereinstimmung im Wesentlichen schließt doch ganz offenbar auch die Zeit der Abfassung des Briefes ein. Und was die Art dieser Übereinstimmung betrifft, so kann Brentano auch nicht sagen wollen, er hätte damals, als er die Intentionalitätspassage niederschrieb, die Lehre von der Aufnahme der Form ohne Materie als solche anerkannt, was eine ganz absurde Selbstinterpretation gewesen wäre: Worauf als das wesentliche Gemeinsame er hinweisen will, ist doch offenbar dies, daß auch Aristoteles, indem er die psychische Bezugnahme auf ein materielles Ding durch die Aufnahme der Form ohne Materie des Dinges, nicht des wahrgenommenen-Dinges oder gedachten-Dinges erklärt, das Vorstellende das Ding und nicht das vorgestellte-Ding vorstellen d.h. zum Objekt der psychischen Bezugnahme haben läßt; und auch nur so wird es verständlich, wie Brentano unmittelbar nach dieser Stelle mit der in Abschnitt I zitierten, in der er gegen die ihm angedichtete Albernheit protestiert, fortsetzen kann. – Aus dieser Stelle via Thomas’ Kommentierung der Aufnahme der Form durch den Sinn etwas über „a different mode of being“ des Pferdes als Objekt herauszulesen und das auf die Intentionalitätspassage in der Psychologie vom empirischen Standpunkt zu projizieren, ist also eine ganz abwegige Gewaltsamkeit. Um zur Sache selbst zu kommen, mag es dienlich sein, mit der Lehre des Aquinaten zu beginnen, daß Gottes essentia dasselbe sei wie sein esse (welches Wort wir am besten unübersetzt lassen). Heißt das also, daß die Erkenntnis dessen, daß ein Gott ist, zugleich die Erkenntnis des göttlichen Wesens ist? Thomas verneint natürlich, und zwar indem er wie beim ens auch beim esse zwei Weisen unterscheidet: Vom esse wird in zwei Weisen gesprochen. In der einen bedeutet es den Akt des Seins (actus essendi), in der anderen die Komposition der Aussage, zu der die Seele im Verbinden des Prädikats mit dem Subjekt gelangt. Das esse in der ersten Weise genommen, können wir freilich Gottes esse nicht erkennen, sowenig wie seine essentia. Nur nach der zweiten Weise können wir Gottes esse erkennen: Wir wissen nämlich, daß die Aussage, die wir über Gott bilden, wenn wir sagen: Es ist ein Gott, wahr ist (ST Ia qu. 3 art. 4 ad 2).

So ist da ein esse Gottes, das wir nicht erkennen können, und ein anderes, das wir erkennen, wenn wir die Aussage ,Es ist ein Gott‘ als wahr erken-

16

nen. Dieses zweite esse ist also das, was das Existenzzeichen ausdrückt, während das erste etwas ist, das bei Gott mit der essentia identisch ist und bei nichtgöttlichen Wesen mit ihr eine Zusammensetzung bildet. Dem gemäß hat esse im ersten Sinn nur das, was auch eine essentia hat d.h. auch auf die ,Quid est?‘-Frage antwortet, so daß die Unterscheidung beim esse zu der beim ens genau parallel läuft: Während nach der ersten Weise als esse der Akt des Seienden (actus entis), insofern es Seiendes ist, bezeichnet wird … [und so] nur eben jenen Dingen zugeschrieben wird, die in den zehn Kategorien enthalten sind,

wird nach der zweiten Weise esse allem zugeschrieben, worüber eine [wahre] Aussage gebildet werden kann, sei es ein Seiendes [gemäß den Kategorien], sei es die Privation eines solchen; wir sagen nämlich auch, daß die Blindheit ist (Quaestiones quodlibetales [Quodl.] IX qu. 2 art. 2 co).

Die Unterscheidung von esse naturale und esse intentionale ist eine die erste Weise des esse, esse als actus entis oder essendi betreffende, worum es also im weiteren geht. Aristoteles führt in Met. Δ.8 unter den Weisen, in denen von der ousia, der Substanz, gesprochen wird, nach der ousia im Sinne des eigentlichen Subjekts das an, was die einem solchen Subjekt immanente Ursache des Seins (aition tou einai) für dasselbe ist, wie z.B. für das Lebewesen die Seele (1017b14–16), d.h. seine Form: Denn die Seele ist das Prinzip des Lebens, und „das Leben ist für die Lebewesen das Sein“ (De Anima II.4 415b13f.); vivere viventibus est esse, wie Thomas oft wiederholt (z.B. Quodl. IX qu. 2 art. 2). Dieses Sein ist energeia, Akt, im strikten Sinn zum Unterschied von der in ihrem Vollzug immer unvollendeten kinêsis (cf. Met. 4.6 1048b18ff.) und vor einer anderen solchen energeia des Lebewesens wie z.B. dem Sehen als sein substanzieller Akt ausgezeichnet. Dieses Aristotelische Sein als energeia ist der Hintergrund von Thomas’ Lehre vom esse als actus essendi, das, wie gesagt, bei Gott mit der essentia identisch ist und bei nichtgöttlichen Wesen mit ihr eine Zusammensetzung bildet, in der des weiteren esse zu essentia sich so verhält wie die Form zur Materie, nämlich als Akt zur Potenz (sicut actus ad potentiam: ST Ia qu. 3 art. 4 co), und durch die jeweilige Form zum jeweiligen besonderen esse spezifiziert ist, so eben beim Lebendigen durch die Seele zu Vivere.13 Das Verhältnis 13. Zu dieser Lehre vom esse, das bei Gott esse einfachhin und bei den nichtgöttlichen Dingen durch ihre Form zu einem je besonderen esse spezifiziert oder limitiert ist, cf. bes. Qu. disp. de

17

zwischen dem derartigen esse und dem im Sinne des Existenzzeichens ist dann das: Ein Seiendes gemäß den Kategorien hat esse im ExistenzzeichenSinn – antwortet positiv auf die ,An est ?‘-Frage – dadurch, daß es esse im Sinne des actus essendi hat (est per hoc quod habet esse: Summa contra Gentiles I cp. 22).14 Nun müssen wir noch sehen, wie Thomas das Haben von solchem esse versteht. In diesem Punkt stellt er die per se bestehenden Substanzen allem Übrigen gegenüber, dem solches esse zukommen kann: Dieses esse wird in zwei Weisen zugeschrieben, und zwar einerseits so wie dem, was eigentlich (proprie) und wahrhaft esse hat oder ist. Und solcherart wird es einzig der per se subsistierenden Substanz zugeschrieben … Alles aber, was nicht per se, sondern in einem Anderen und mit einem Anderen subsistiert, seien es Akzidenzien oder substanzielle Formen, oder auch irgendwelche Teile, hat esse nicht so, daß es selbst wahrhaft ist, sondern es wird ihm esse in anderer Weise zugeschrieben, nämlich als dem, wodurch irgendetwas ist; so wie von der Farbe Weiß (albedo) esse ausgesagt wird nicht, weil etwa sie selbst in sich subsistiere, sondern weil durch sie irgendetwas esse hat derart, daß es weiß ist (quod habet esse album) (Quodl. IX qu. 2 art. 2 co).

Katzen z.B. sind per se Bestehendes, Substanzen im Sinne des Subjekts oder erste Substanzen in der Diktion der Kategorienschrift des Aristoteles: Von einer Katze zu sagen, sie habe esse, heißt von ihr zu sagen, daß sie ein so und so geartetes Leben lebt; von der Katzenform zu sagen, sie habe esse, drückt dasselbe nur in anderer, die nicht eigentlich Subjekt seiende Form als Subjekt herausstellender Bezeichnungsweise aus; und desgleichen dann für ein Akzidenz wie die Farbe Weiß, für die esse zu haben besagt, daß irgendetwas weiß, oder genauer, aktual weiß ist. Nun ist aber dies, daß die Farbe Weiß solches esse so hat, daß etwas weiß ist, eben das, daß sie esse naturale hat (ST Ia qu. 56 art. 2 ad 3), und in Entsprechung dazu besteht dann ihr Haben von esse intentionale im sensitiven Körper darin, daß dieser aktual Weißes wahrnehmend ist. So kommt im Erkenntnisakt zweierlei zur Geltung, nämlich einerseits die Form, vermöge welcher andererseits dem Erkenntnisakt ein Objekt gegeben ist: Potentia qu. 7 art. 2 co und ad 9. 14. In neuerer Zeit hat besonders Peter Geach das Thomasische esse im Sinne des actus essendi verteidigt und dabei insbesondere auch den Zusammenhang dieses esse mit dem Gebrauch von ,sein‘ in ,anfangen zu sein‘, ,fortfahren zu sein‘ und ,aufhören zu sein‘ herausgestellt: G. E. M. Anscombe & P. T. Geach, Three Philosophers, Oxford 1961, p. 88ff.

18

Eine derartige [Prinzip des Erkennens im Erkennenden seiende] Form kann in zwei Weisen betrachtet werden. Einmal dem esse nach, das sie im Erkennenden hat: In dieser Hinsicht macht sie den Erkennenden aktual erkennend. Zum anderen dem Bezug (respectus) nach, den sie zu dem Ding hat, dessen Ähnlichkeit (similitudo) sie ist: In dieser Hinsicht aber bestimmt sie die Erkenntnis hin zu einem bestimmten Erkennbaren. (Qu. disp. de Veritate qu. 10 art. 4 co; Text etwas umgestellt)

Nach Thomas ist das Erkannte in gewisser Weise im Erkennenden (In De An. lib.I, lec.4, n.43; lib.II, lec.12, n.377), nämlich so, daß die im Erkennenden esse intentionale habende Form die Ähnlichkeit dessen ist, was das Erkennende zum Objekt hat, so daß in ihr das Objekt in obliquo mit zur Gegebenheit kommt. Um in dieser Weise im Erkennenden zu sein, bedarf also das Objekt nicht selbst wieder irgendeines esse intentionale, indem die solches esse habende Form oder species als das aktual erkennen Machende in einem damit den Bezug zum Objekt liefert, so daß das Denken eines Dinges eben das Denken eines Dinges ist: In Thomas’ prägnanter Formel: Die species ist nicht das, was, sondern das, wodurch etwas erkannt wird (cf. z.B. In De An. lib.III, lec. 8, n. 718; ST Ia qu. 85 art. 2 co). Brentanos Rede von der intentionalen Inexistenz des Objekts in der Intentionalitätspassage ist also nicht einmal der sprachlichen Formulierung nach mit Thomas’ Analyse des Denkaktes in Einklang zu bringen, vielmehr entstammt sie offensichtlich späteren scholastischen Quellen.15 Umso weniger ist ein begrifflicher Einklang gegeben. Das Begriffsgefüge in Brentanos Psychologie von 1874 ist nicht mehr das metaphysische Begriffsgefüge, in das das esse-Haben der Form gehört, wie denn auch in der Psychologie vom empirischen Standpunkt diese spezifisch Thomasische Lehre vom esse als Akt, und eo ipso die vom esse intentionale, ausdrücklich als Irrweg, und zwar als ein auf der Verkennung der Natur des Existenzialurteils beruhender Irrweg hingestellt wird (PeS II p. 75–77). Und so ist also McAlisters Versuch verfehlt, der ontologischen Deutung der Intentionalitätsthese, wie sie Brentano 1874 formuliert, sozusagen Thomasische Nahrung zuzuführen, nicht nur in dem Brentano betreffenden Ergebnis, sondern schon vom ganzen Ansatz her, indem bei Thomas das esse im Sinne des actus essendi mit Existenz im Sinne des Existenzzeichens sowie das intentionale Objekt mit der esse intentionale habenden Form bzw. species konfundiert wird, mit dem Scheinergebnis 15. Cf. dazu Klaus Hedwig, ,Der scholastische Kontext des Intentionalen bei Brentano‘, Grazer Philosophische Studien 5 (1978).

19

einer besonderen, intentionalen Existenz desselben: Was sich zunächst als abwegige Gewaltsamkeit in der Interpretation jener Stelle in dem Brief Brentanos an Marty vom 17.3.1905 darstellte, hat sich auch in der Sache als ganz unbegründet erwiesen. IV. So hat sich die ontologische Deutung der Intentionalitätsthese beim vorreistischen Brentano sowohl nach (A) wie nach (B) als unhaltbar erwiesen, und das aus letztlich doch recht offen daliegenden Gründen. Woher also diese hartnäckige Tendenz, dem vorreistischen Brentano eine ontologische Intentionalitätsauffassung zuzuschreiben, sogar aus dem von ihm ohne besondere Festlegung gebrauchten Term ,intentionale Inexistenz‘ weitläufige ontologische Folgerungen zu ziehen, während für den reistischen unisono – und natürlich richtig – das Gegenteil gesagt wird? Die Antwort darauf hat natürlich mit dem Problem des Einhorn- oder Zentaur-Denkenden zu tun: Die Intentionalität soll eine Beziehung sein, und wie soll sie bestehen können in dem Fall, da keine zweite Entität da ist? Nun ist es der charakteristische Zug der Relationenlehre des späten Brentano, daß er die so selbstverständlich erscheinende Annahme, das Bestehen einer (zweistelligen) Beziehung erfordere die Existenz zweier Entitäten, als allgemeingeltendes Merkmal des Relationalen verwirft und auf besondere Arten desselben beschränkt. Der locus classicus dafür befindet sich in der ersten der von Brentano zur zweiten Auflage der Psychologie vom empirischen Standpunkt von 1911 hinzugefügten Abhandlungen. Dafür, daß z.B. eine Beziehung von Ursache und Wirkung bestehe, muß sowohl das, was verursacht, als das, was verursacht wird, existieren … Ganz anders ist es dagegen bei der psychischen Beziehung. Denkt einer etwas, so muß zwar das Denkende, keineswegs aber das Objekt seines Denkens existieren … So ist denn das Denkende das einzige Ding, welches die psychische Beziehung verlangt … Man könnte darum zweifeln, ob hier wirklich etwas Relatives vorliege, und nicht vielmehr etwas in gewissem Betracht einem Relativen Ähnliches, was man darum als etwas „Relativliches“ bezeichnen könnte. Die Ähnlichkeit besteht darin, daß, wie derjenige, der ein Relativ im eigentlichen Sinne denkt, auch der, welcher eine psychische Tätigkeit denkt, in gewisser Weise zugleich zwei Objekte denkt, das eine sozusagen in recto, das andere in obliquo (PeS II p. 133f.).

20

Die Erwägung, die Denkbeziehung als bloß Relativliches vom eigentlichen Relativen abzugrenzen, ist eine Übergangserscheinung, denn in der Folgezeit wird Brentano eben das, was das Gemeinsame ist, als das Wesen des Relativen herausstellen und im Denken von etwas in recto und zugleich von etwas anderem in obliquo den „einheitlichen Begriff für alles Relative“ finden (KL p.169). Und so pflegt man in dieser Unterscheidung von Modus rectus und Modus obliquus das Instrument zu sehen, mit dem sich Brentano von der ontologischen Intentionalitätsauffassung habe lösen können. Denn ursprünglich, so Kraus, folgte Brentano der „herrschenden Lehre“, daß das Bestehen einer Beziehung stets „die Existenz zweier Termini“ verlange, so daß er die Denkbeziehung als Beziehung „zwischen demjenigen, der irgendetwas vorstellt und einem dem Geiste irgendwie innewohnenden Objekte“ faßte und damit, solange er daran festhielt, zur Sicherung des zweiten Terminus auf „die Lehre ,von der mentalen Inexistenz‘ des Objektes“ festgelegt gewesen sei (PeS I p. xxivf.): Dann aber habe er erkannt, daß für die Denkbeziehung „nur die Existenz des einen Terminus gefordert ist, keinesfalls aber die Existenz des zweiten Terminus, auf den der psychisch Tätige sich bezieht“ (p. xxvi). Gleich so auch Alfred Kastil, der andere Brentano-Kronschüler, und so fort bis in die Gegenwart.16 Aber nachdem wir gesehen haben, daß die vermeintliche Lösung des Problems mit dem nichtexistierenden zweiten Relatum der Denkbeziehung beim vorreistischen Brentano weder in der Gestalt von (A) noch in der von (B) existiert, wird wohl auch das Problem selbst nur ein vorgebliches sein: Und es zeigt sich, daß es in der Tat nur auf einem Mißverständnis von Brentanos Auffassung des Relationalen beruht. Wenn wir von relationalen Sachverhalten sprechen, so denken wir an etwas von der Form aRb: an ein geordnetes Paar von Dingen und ein zweistelliges Prädikat, das auf die Paarglieder zutrifft, so daß das eine zum anderen in der von dem zweistelligen Prädikat ausgedrückten Beziehung steht. Diese Sichtweise auf Brentano zu übertragen, heißt aber, den Aristotelischen Hintergrund seines Denkens über relationale Sachverhalte zu 16. Für Kastil cf. dessen lange Anmerkung in VE p. 262f. Anm. 15. McAlister: „Brentano then believed that a necessary condition for there being a relation, xRy, is that both x and y exist“, so daß er gemeint habe, daß „in the case of, for example, someone thinking about a centaur, xRy is a genuine relation because, even though the centaur does not have actual existence, he does have another mode of existence, i.e. intentional inexistence, but existence nonetheless“ (a.a.O. p. 27f.). Ebenfalls zu nennen wäre John Haldane, ,Brentano’s Problem‘, Grazer Philosophische Studien 35 (1989), nach dem Brentano „[o]riginally … held that ,X thinks of Y‘, and ,X is larger than Y‘, require the same analysis, viz., that … two terms … are involved“ (p. 11).

21

ignorieren. Aristoteles spricht vom Relativen, dem pros ti als dem zu etwas sich Verhaltenden, wie Brentano übersetzt (KL p. 166). Nehmen wir den Sachverhalt, daß Simmias größer ist als Sokrates. Wie wir über Relationen zu denken gewohnt sind, sind Simmias und Sokrates die betreffenden Relata, d.h. die Glieder eines geordneten Paares, auf welche das zweistellige Prädikat ,ist größer als‘ zutrifft, so daß Simmias in der Beziehung des Größer-Seins zu Sokrates steht. Aber unsere Relata sind im Aristotelischen Kategorienschema Substanzen und keine Relativa. Solche sind vielmehr Größeres und Kleineres, Herr und Knecht und dergleichen (Cat. cp. 7 6a38–b2, b29–33). D.h. Aristoteles analysiert unseren Fall so, daß von den beiden einstelligen Prädikaten ,ist größer-als-Sokrates‘ und ,ist kleiner-als-Simmias‘ das erste von Simmias und das zweite von Sokrates wahr prädizierbar ist, was weiters besagt, daß hier zwei Entitäten vorliegen, „für die zu sein dasselbe ist wie zu etwas irgendwie sich zu verhalten“ (8a32), nämlich ein Größeres-als-Sokrates, das dem Subjekt nach Simmias, und ein Kleineres-als-Simmias, das dem Subjekt nach Sokrates ist: Hier haben wir das Relative und sein Korrelat, von denen Aristoteles sagt, was dann Brentano wieder geltend machen wird, nämlich daß sie sich der Existenz und der Erkenntnis nach gegenseitig involvieren (7b15–22, 8a35–b15). Bei der Koexistenzbedingung kannte Aristoteles freilich eine Ausnahme, eben die Denkbeziehung (7b22–8a12), bei der nach Met. Δ.15 nur das eine Glied des Korrelatenpaares, das Denkende, ein genuines Relatives, das zweite hingegen mehr nur dem Namen nach ein solches ist, „denn das Denkbare (dianoêton) bezeichnet, daß es ein Denken von ihm gibt“(1021a31). In dieser Sicht des relationalen Sachverhalts zerlegt also die ontologische Analyse der Form aRb diese für einen Fall wie den, daß Simmias größer ist als Sokrates, in das Korrelatenpaar a(Rb) und b(R*a), wobei R* konvers zu R ist und (Rb), (R*a) die Formen der betreffenden monadischen Bestimmungen von a und b sind. Nun paßt das aber nicht für die Denkbeziehung, denn wenn wir a(Rb) das Denkende und b(R*a) das Gedachte repräsentieren lassen, so würde das von (R*a) repräsentierte Prädikat ,wirdgedacht-von-a‘ auf b zutreffen und also die Existenz von b verlangt sein. Die Ausnahme, die die Denkbeziehung nach Aristoteles ist, stellt sich so dar, daß die Form des Gedachten vielmehr so etwas ist wie (bR*)a, was aber nur zum Prädikat ,b-wird-gedacht-von‘ führt, das sich bloß durch die sprachliche Passivform vom Prädikat ,denkt-b‘ unterscheidet:17 So 17. Anders gesagt: Die Form (bR*)a des Gedachten drückt aus, daß a(Rb), wenn das Denkende repräsentierend, nicht die Existenz von b impliziert. Zur Relationenlehre Brentanos siehe auch Stephan Körner, ,Über Brentanos Reismus und die extensionale Logik‘, Grazer

22

ist (bR*) nicht mehr als nur ein ens linguae, als welches das Gedachte an der angeführten Aristoteles-Stelle jedenfalls implizit gefaßt wird. Ebenso knapp wie erhellend dazu der späte Brentano: Bei der Unterscheidung eines A-Denkenden von einem B-Denkenden kommt man naturgemäß dazu, … namhaft zu machen, was er denkt und worauf, wie man zu sagen pflegt, er sich denkend bezieht. Daraufhin behandelt man sprachlich den Fall so, als ob es sich um eine Relation zwischen zwei Dingen handelte, und kehrt diese Relation um, indem man den Unterschied der aktiven und passiven Verbalform benützt, obwohl es sich um kein Tun und Leiden handelt. So erscheint denn sprachlich das, was einer denkt, ganz so behandelt, als wäre es wie er selbst, und es kommt zu dem Gebrauche des Seienden im uneigentlichen Sinne (KL p.15),

d.h. dazu, daß man sagt, „ein gedachtes Ding sei, und zwar sei es in einem Denkenden“ (p. 14). Auf diesem Weg ist beim vorreistischen Brentano das Gedachte zum nichtrealen Glied des intentionalen Korrelatenpaares geworden, das dergestalt die Koexistenzbedingung erfüllte, so daß er sagen konnte, daß sie wie die Erkenntnisbedingung „überall gilt“ (VE p.45). So also ist es zu verstehen, wenn der vorreistische Brentano sagt, „bei jeder Beziehung finden sich … zwei Korrelate“ (DP p. 21), was bei der Denkbeziehung eben nicht impliziert, daß auch zwei Relata existieren. Es ist nun leicht, das hinter der ontologischen Deutung der Intentionalitätsthese beim vorreistischen Brentano stehende Mißverständnis herauszustellen: Es ist einfach die Verwechslung der Korrelate mit den Relata. Und ist das einmal geschehen, dann ist die Koexistenzbedingung für die Denkbeziehung – die beim vorreistischen Brentano wirklich, aber eben für die Korrelate galt – tatsächlich nur dadurch sicherzustellen, daß man in der Form aRb des relationalen Sachverhalts entweder im Sinne von (A) Philosophische Studien 5 (1978), §3. – Zu beachten ist freilich, daß die Formen a(Rb) usw. nicht genuin Brentanosche sind, sondern prädikatenlogische Gegenstücke von solchen. Nach der Brentanoschen Urteilslehre besteht der Satz ,Simmias ist größer-als-Sokrates‘ logisch nicht aus einem singulären Term und einem davon syntaktisch grundverschiedenen Prädikat, ,ist größer-als-Sokrates‘, sondern aus zwei „Namen“, die eine Namenskonjunktion zum komplexen Namen ,der ein Größeres-als-Sokrates seiende Simmias‘ verknüpft, den das Existenzzeichen ,ist‘ zum Satz ergänzt. So steht Brentano in seinem logischen Denken der sogenannten Ontologie Leśniewskis weitaus näher als der Prädikatenlogik Frege-Russellscher Provenienz. Zu Leśniewskis Ontologie siehe Czeslaw Lejewski, ,Zu Leśniewskis Ontologie‘, Ratio 1 (1957). Für einen Aufbau der Leśniewski-Ontologie auf der Brentanoschen Basis von Namenskonjunktion und Existenzzeichen anstelle von Leśniewskis singulärer Kopula siehe Peter Simons, ,A Brentanian Basis for Leśniewskian Logic‘, Philosophy and Logic in Central Europe from Bolzano to Tarski, Dordrecht 1992.

23

gedachtes-b für b substituiert oder im Sinne von (B) für b eine spezielle Existenzweise einführt (oder gar beides amalgamiert). Wurde oben gesagt, es sei der charakteristische Zug der Relationenlehre des späten Brentano, daß sie für das Bestehen einer Beziehung nicht die Existenz zweier Entitäten verlange, so sehen wir jetzt genauer, wie das zu verstehen ist: nämlich so, daß es Relativa ohne Korrelat geben kann. Von Brentanos eigenem Standpunkt her ist das die große Änderung nach der reistischen Wende und der Einführung der Unterscheidung von Modus rectus und Modus obliquus des Denkens. Was er dagegen schon immer gehabt hatte, waren Relativa ohne ein Paar von Relata: eben die intentionalen Relativa wie das Zentaur-Denkende oder jede äußere Wahrnehmung, gibt es doch auch für den Brentano von 1874 keine Sinnesqualitäten, keine Farben, Töne usw. in der physischen Welt (cf. PeS I p. 13f.). V. Wenn Brentano 1911 erwägt, die Denkbeziehung als bloß Relativliches von den eigentlichen Relativa abzugrenzen, so hatte er das auch früher bereits tatsächlich getan, denn im Anhang VI der Deskriptiven Psychologie, vermutlich aus der Zeit um 1901 stammend, lesen wir, daß die „sog. psychischen Beziehungen … keine eigentlichen Relationen sind“ (DP p. 163). Der Grund dafür findet sich in einem Brief von Kraus an Brentano kurz nach dessen reistischer Wende ausgesprochen, nämlich, daß es „sich beim Psychischen um eine einseitige reale Relation, und somit um etwas ganz Eigenartiges“ handle (6.10.1904, ANR p. 119). Brentano selbst verneinte zu der Zeit überhaupt, daß der Denkakt vermöge seiner Intentionalität etwas Relatives sei: „Sie wollen die intentionale Beziehung nicht gelten lassen“, schreibt Kraus im selben Brief (p. 118). Offenbar hatte Brentano zu dieser Zeit die Unterscheidung von Modus rectus und Modus obliquus noch nicht zur Verfügung, mit der er dann den Entfall des nichtrealen Denkkorrelats kompensieren und den Denkakt erneut als Relatives, wenngleich zunächst noch im uneigentlichen Sinn, verstehen konnte (so schon in der Abhandlung ,Vom Objekt‘ aus 1906, ANR p. 340). Also mußte es in dieser Zwischenzeit so sein, daß mit einem A-Denkenden als solchem, wie Brentano im September 1904 an Marty schreibt, „keine Relation gegeben [ist], außer wenn sowohl ein A-Denkendes als A selbst oder dgl. besteht“ (ANR p. 115). Das sind demnach die Stationen: Zunächst war der Denkakt ein echtes

24

Relatives, weil er ein nichtreales Korrelat hatte; sodann gegen Ende der vorreistischen Zeit ein uneigentliches Relatives, weil er ein nichtreales Korrelat hatte; darauf in den der ersten reistischen Zeit als solcher überhaupt kein Relatives; und endlich aufgrund seiner recto/obliquo-Struktur wieder ein Relatives, vorerst noch in eher uneigentlichem Sinn und schließlich im ganz eigentlichen Sinn, womit sich der Kreis schließt, aber so, daß das Gedachte zu dem ens linguae reduziert worden ist, das es bei Aristoteles eigentlich schon war: Wer sagt, ein A-Denkender sei und ein gedachtes A sei, sagt ganz und gar dasselbe, und der letzte so wenig als der erste, daß A selbst sei (,Vom Objekt‘, ANR p. 339),

heißt es jetzt in diametralem Gegensatz zu dem über das Denkende und das Gedachte in der Gedankending-Stelle Gesagten. Aber alle diese Stationen in der Deutung der Intentionalität des Denkaktes betreffen Wandlungen in Brentanos Lehre vom Relativen in Verbindung zuerst mit der Anerkennung und dann der Verwerfung des Nichtrealen und affizieren nicht die innere Struktur des Denkaktes selbst, die einmal in Begriffen des distinktionellen Teils im modifizierenden Sinn und dann wieder in Begriffen der Unterscheidung von Modus rectus und Modus obliquus analysiert wird: Und so können wir damit schließen, daß für die Tendenz, dem vorreistischen Brentano eine wesentlich andere Intentionalitätsauffassung zuzuschreiben als dem reistischen, nämlich eine sogenannte ontologische, bei Brentano selbst keinerlei Grundlage vorzufinden ist.

VERWENDETE WERKE BRENTANOS ANR = Die Abkehr vom Nichtrealen (hrsg. von F. Mayer-Hillebrand), Neudruck Hamburg 1977. DP = Deskriptive Psychologie (hrsg. von R. M. Chisholm & W. Baumgartner), Hamburg 1982. KL = Kategorienlehre (hrsg. von A. Kastil), Neudruck Hamburg 1985. MB = Von der mannigfachen Bedeutung des Seienden nach Aristoteles, Neudruck Hildesheim 1960. PeS = Psychologie vom empirischen Standpunkt I–II (hrsg. von O. Kraus), I: Neudruck der Ausg. von 1924, Hamburg 1973, II: Neudruck der Ausg. von 1925, Hamburg 1971.

25

VE = Versuch über die Erkenntnis (hrsg. von A. Kastil), 2. Aufl. (hrsg. von F. MayerHillebrand) Hamburg 1970. WE = Wahrheit und Evidenz (hrsg. von O. Kraus), Neudruck Hamburg 1974.

26

Grazer Philosophische Studien 73 (2006), 27–41.

ZUR THEORIE DER VORSTELLUNGSPRODUKTION („GRAZER“ GESTALTTHEORIE I: FRANCE WEBER) Tanja PIHLAR Graz & Kranj Summary In the following discussion, we are dealing with Weber’s theory of the production of presentations, as presented in his article “The Problem of the Production of Presentations”. In this article, published in 1928, Weber offers an essential modification of a version of the theory of objects which had been developed by the Graz school (and was closely linked with the theory of higher-order objects). According to Weber, the production of presentations consists in a primary transition from passive to corresponding active presentations (so there is active as well as passive presentation). Weber distinguishes several types of production of presentations: psychophysical, content, act, intentional, and surrogate production, all of which can be divided into many subtypes. Most interesting in this connection is his theory of intentional presentation. In the 1928 article, Weber postulates non-intentional presentations, on which intentional presentations are based. He distinguishes four levels of intentionality: non-intentional presentation, on the lowest level, is followed by presentational intentionality, isolative, and rational intentional presentation. Weber’s 1928 article is of considerable importance for an understanding of his subsequent philosophical development.

1. Einleitung Im vorliegenden Aufsatz soll die Theorie der Vorstellungsproduktion erörtert werden, die der slowenische Philosoph und Meinong-Schüler France Weber (1890–1975) in seiner umfangreichen Abhandlung „Problem predstavne produkcije“ [„Das Problem der Vorstellungsproduktion“] (1928)1 entwickelt hat. Diese Abhandlung, die Weber selbst als 1. Zitiert nach der deutschen Übersetzung: „Über das Problem der Vorstellungsproduktion. Aus dem Slowenischen übertragen von cand. phil. Franz Žekar“, 114 S. Das bisher unpublizierte

„die allererste Anbahnung einer neuen und erst zu schaffenden gesamten Aktivitätspsychologie (nicht ,Aktpsychologie‘)“ (PVP, 114) bezeichnet, fällt in die Anfangsphase seiner „personalistischen“ Periode (etwa bis 1934), in welcher er sich vor allem mit dem Verhältnis zwischen dem Erlebnissubjekt und dessen Erlebnissen beschäftigte und seine Philosophie der Persönlichkeit entfaltete. Webers Theorie der Vorstellungsproduktion (VP) verdankt viel der Grazer Schule der Gegenstandstheorie und steht in enger Beziehung zur Lehre von den Gegenständen höherer Ordnung.2 Wie bekannt ist, gibt es nach Meinong Gegenstände verschiedener Konstitutionsebenen, nämlich sowohl Gegenstände, die selbständig existieren können (etwa Töne, Farben) als auch Gegenstände, die andere Gegenstände notwendigerweise voraussetzen (etwa eine Melodie, eine Verschiedenheit). Die letzteren werden daher durch eine besondere „innere Unselbständigkeit“ (Meinong 1899; GA II, 386) gekennzeichnet, die Meinong auch mit „Unfertigkeit“ (ebd.) bezeichnet, welche in ihrer Natur liegt. Die Verschiedenheit von Rot und Grün kann zum Beispiel nicht ohne die Farben Rot und Grün bestehen, auf welchen sie notwendigerweise aufbaut. Sie wird als Gegenstand höherer Ordnung bzw. als Superius bezeichnet, während Gegenstände, die ihr als Voraussetzung dienen, Inferiora heißen. Ein Gegenstand höherer Ordnung ist etwas Neues und Eigenständiges und kann nicht als ein bloßes Kollektiv aus Inferiora aufgefasst werden. Hierbei gilt, dass ein Inferius auch allein, ohne entsprechenden Superius vorkommen kann, während ein Superius seine Inferiora immer benötigt. Ordnungsreihen bleiben nach Meinongs Auffassung offen, insofern es weitere Gegenstände von höherer und höchster Ordnung gibt, während sie nach „untenhin“ mit Infima abgeschlossen sind – dies ist das sogenannte Prinzip der obligatorischen Infima (siehe Meinong 1921; GA VII, 15). An dieser Stelle stellt sich die Frage, wie man einen Gegenstand höherer Ordnung, wie etwa eine Verschiedenheit oder eine Melodie erfassen kann. Gegenstände niederer Ordnung sind nach Meinong real – er nennt Manuskript befindet sich im Weber-Nachlass (Signatur Z.1.), welcher in der „Forschungsstelle und Dokumentationszentrum für Österreichische Philosophie“ in Graz aufbewahrt ist. Hier wird die Abkürzung PVP benutzt. Die Originalabhandlung in slowenischer Sprache wurde in der Zeitschrift des Wissenschaftlichen Vereines für Geisteswissenschaften in Ljubljana 1928 veröffentlicht. 2. Über die Vorgeschichte dieser Theorie siehe Wolfgang G. Stock (1995), 458ff. Die Produktionstheorie wurde von der Grazer Schule sowohl theoretisch als auch experimentell erforscht. Dazu siehe Wolfgang G. Stock (1995), 462ff; Mauro Antonelli (1994), 36ff; Andrea Zemljič (1993), 27ff.

28

„real“ jene Gegenstände, für die gilt, dass sie „falls sie nicht wirklich existieren, ihrer Natur nach jedenfalls existieren könnten“ (siehe Meinong 1899; GA II, 394). Gegenstände höherer Ordnung sind hingegen in der Regel ideal; als solche können sie bloß bestehen, ihnen kommt keine Existenz zu. Dazu bemerkt Meinong, dass nur reale Gegenstände niederer Ordnung wahrgenommen werden können, während Gegenstände höherer Ordnung, indem sie ideal sind, keinen Gegenstand der Sinneswahrnehmung ausmachen. Meinong behauptet, dass für das Entstehen bestimmter Vorstellungen eine Tätigkeit des Subjektes erforderlich sei, mit deren Hilfe das Vorstellungsmaterial verarbeitet wird. Ein Beispiel: Wenn wir uns die Verschiedenheit der Farben Blau und Grün vorstellen, müssen wir uns gleichzeitig sowohl Blau als auch Grün vorstellen, wobei beide Vorstellungen „in eine bestimmte Realrelation treten“ (1899; GA II, 398). Als Ergebnis dieses Vorgangs wird eine neue Vorstellung, nämlich die der Verschiedenheit zwischen Blau und Grün gebildet. Der beschriebene psychische Vorgang wurde in der Grazer Schule „Vorstellungsproduktion“ genannt.3 Wie Ameseder in seiner Schrift „Über Vorstellungsproduktion“ (1904) zusammenfassend feststellt, sind für das Entstehen von Superiusvorstellungen zwei Faktoren erforderlich (siehe Ameseder 1904, 487f.), nämlich sowohl Vorstellungen von entsprechenden Inferiora als auch etwas zusätzlich Psychisches. Dadurch werden Superiusvorstellungen „hervorgerufen“ bzw. „gebildet“ – Ameseder spricht in diesem Fall von „Produktion“ (siehe Ameseder 1904, 488). Als deren Ergebnis werden produzierte Vorstellungen genannt; diesen sind Elementarvorstellungen gegenüberzustellen. Produzierte Vorstellungen stellen im Vergleich zu Elementarvorstellungen etwas Neues dar, obwohl sie auf diesen aufgebaut sind. Außerdem ist wichtig, im Auge zu behalten, dass im obigen Beispiel die Verschiedenheitsvorstellung produziert wird, nicht jedoch die Verschiedenheit selbst, wie Ameseder in der erwähnten Abhandlung ausdrücklich hervorhebt. Wie wir sehen werden, hat Weber die „Grazer“ Produktionstheorie in der Abhandlung von 1928 in mehrfacher Hinsicht stark modifiziert. Diese Modifikationen beginnen schon bei seiner Auffassung von VP, welche er als ein primäres Übergehen vom passiven zum entsprechenden aktiven Vorstellen charakterisiert. Er unterscheidet fünf Grundarten der 3. Meinong spricht in seiner Abhandlung von 1899 noch über „Fundierung“: „Im Hinblick auf diesen Sachverhalt nenne ich den eben skizzierten Vorgang Fundierung, genauer Fundierung der betreffenden Superiora durch ihre Inferiora […]. Fundierung leistet insofern für Vorstellungen idealer Gegenstände dasselbe wie Wahrnehmung für Vorstellungen realer Gegenstände […]“ (GA II, 399).

29

VP, und zwar psychophysische, inhaltliche, aktmäßige, surrogative und intentionale Produktion. Von besonderem Interesse ist in diesem Kontext Webers Theorie der Intentionalität, welche er in der genannten Abhandlung vertritt. 1928 meint er nämlich, dass intentionale Erlebnisse nicht einfach vorgegeben sind, sondern dass sie produziert werden. In diesem Zusammenhang unterscheidet er vier Intentionalitätsstufen: Dem nichtintentionalen Vorstellen auf der untersten Ebene folgen bzw. sind auf ihm aufgebaut: ein Vorstellen mit präsentierender Intentionalität, ein isolatives und ein rationales intentionales Vorstellen. Wie wir sehen werden, hat Weber in der erwähnten Abhandlung auch einige seiner Grundbegriffe bzw. einige seiner Theorien mit einer neuen Interpretation versehen, wie z.B. das sekundäre Vorstellen, das Ernst- und Phantasievorstellen sowie seine Empfindungstheorie. Weber hat jedoch diese neue Produktionstheorie in seinen späteren Werken – soweit uns bekannt ist – nicht mehr aufgegriffen, sie stellt sozusagen eine bloße „Episode“ in seinem Denken dar. Da Weber manche dieser Theorien auch nicht ausführlich ausgearbeitet hat, darf man aus ihnen nicht allzu große Konsequenzen ziehen. Trotz dieser Tatsache ist die hier untersuchte Abhandlung von 1928, welche bis zum heutigen Tage (weitgehend) unbeachtet geblieben ist4, in mancher Hinsicht für das Verständnis der weiteren philosophischen Entwicklung Webers von Bedeutung. Unsere Absicht ist in erster Linie zu zeigen, in welche „merkwürdige“ Richtung sich die Grazer Theorie der Vorstellungsproduktion in Slowenien entwickelt hat. 2. Die Vorstellungsproduktion: Definition Weber hat die ursprüngliche VP-Theorie in mehreren früheren Schriften kurz erörtert, insbesondere in seinen Büchern Sistem filozofije [System der Philosophie] (1921), Očrt psihologije [Grundriss der Psychologie] (1924) und Estetika [Ästhetik] (1925). Ausführlicher geht Weber in der Abhandlung „Problem predstavne produkcije [Das Problem der Vorstellungsproduktion]“ (1928) auf sie ein, modifiziert aber seine früheren Ansichten bedeutsam. Hier wollen wir zunächst seine frühe VP-Theorie darstellen. Weber behauptet in seinen Frühschriften (1921, 1924, 1925), dass alle Vorstellungen ihrer Natur nach indifferent sind, sie sind weder posi4. Diese Abhandlung ist bisher kaum behandelt worden. Hinweise auf sie finden wir in: Wolfgang G. Stock (1995), 483; Anton Terstenjak (1972), 37ff; Ludvik Bartelj (1983), 83ff.

30

tiv noch negativ; außerdem sind sie – im Gegensatz zu Gedanken und Strebungen – durchaus passiv. Diesen Standpunkt hat Weber in der Folge stark modifiziert: In der Abhandlung von 1928 behauptet er, dass nicht alle Vorstellungen vollständig passiv sind, vielmehr gibt es auch ein aktives Vorstellen. Im Einklang damit wird die VP folgendermaßen definiert: Sie ist als ein ursprüngliches oder primäres Übergehen vom passiven (oder relativ passiven) Vorstellen zum zugeordneten aktiven Vorstellen zu verstehen. Demgemäß werden zunächst Vorstellungen produziert, indem wir sie „ursprünglich oder primär gewinnen“ (PVP, 1). Diese dienen ihrerseits als psychologische Voraussetzung jener Vorstellungen, zu welchen wir auf eine andere Weise gelangen, etwa mittels Vorstellungsassoziation oder mittels Vorstellungsdetermination. Mit Webers Worten ausgedrückt, macht „die umschriebene produktive oder ursprüngliche Vorstellungsentwicklung die unumgängliche Grundlage des gesamten übrigen Vorstellungsflußes […]“ aus (PVP, 2). Darüber hinaus hat Weber seine frühere Auffassung der VP wesentlich umgestaltet: 1928 behauptet er, dass die beschriebene psychische Aktivität (wie Vergleichen, Gestalten) kein besonderes Erlebnis darstellt, sondern als notwendiger Teil des aktiven Vorstellens anzusehen ist. Für diese Behauptung führt er zwei Argumente an: (1) Einerseits steht die Gestaltvorstellung mit der psychischen Aktivität des Gestaltens in viel engerem Zusammenhang als dies für Gestaltvorstellungen bezüglich ihrer Vorstellungsfundamente gilt. Weber meint nämlich, dass zwischen Gestaltvorstellungen und den Vorstellungen ihrer Fundamente eine einseitige Abhängigkeit besteht. Man kann sich z.B. keine Melodie ohne die ihr zugehörigen Töne vorstellen, während die Vorstellungen einzelner Töne keine Melodievorstellungen voraussetzen. Andrerseits sind aber Gestaltvorstellungen und die ihnen zugehörige psychische Aktivität des Gestaltens voneinander abhängig, da es weder Gestaltvorstellungen ohne gleichzeitige psychische Aktivität gibt, noch psychische Aktivität ohne synchrone Gestaltvorstellungen möglich ist. Daraus ergibt sich für Weber, dass diese Aktivität kein eigenständiges Erlebnis darstellen kann, sondern lediglich als notwendige Komponente des Gestaltvorstellens aufzufassen ist. (2) Zur Bekräftigung der Behauptung, dass die psychische Aktivität des Gestaltens bzw. des Vergleichens nicht als ein nicht-präsentierendes reales Erlebnis qualifiziert werden darf, führt er die Tatsache an, dass wir uns auf Grund derselben Fundamente ganz verschiedene Gestalten vorstellen können. Dies lässt sich nicht mit Hilfe der Vorstellungen selbst erklären, da

31

durchaus nicht klar ist, wie man von inhaltlich gleichen Vorstellungen (an der Basis) zu inhaltlich verschiedenen Vorstellungen (höherer Ordnung) gelangt. Die Webersche Lösung lautet: Die erwähnte Tatsache kann mit Hilfe der psychischen Aktivität des Vergleichens selbst erklärt werden, d.h. dem Vorgang des Vergleichens kommt auch die Präsentationsfunktion zu. Wenn diese Aktivität de facto ohne jeden Inhalt wäre, könnte sie keinesfalls zu inhaltlich verschiedenen Vorstellungen führen. Aus alledem ergibt sich, dass die betreffende psychische Aktivität einen notwendigen Teil der Gestalt- und Relationsvorstellungen ausmacht. In seinen Frühschriften hat Weber vor allem zwei Produktionsarten behandelt, deren Ergebnis Gestalt- und Vergleichungsvorstellungen bilden.5 In der Abhandlung von 1928 fügt er zahlreiche neue Arten von VP hinzu, welche sich in der Vorstellungsaktivität voneinander unterscheiden. Ausgehend von einer Analyse der Struktur des Vorstellens hält Weber inhaltliche, aktmäßige und intentionale VP auseinander, welche ihrerseits wiederum in mannigfache Unterarten zerfallen. Außerdem postuliert er eine sogenannte psychophysische und eine surrogative VP, deren Ergebnis Empfindungen bzw. Raum- und Zeitvorstellungen sind.6 Zwischen den genannten Arten der VP bestehen nach Weber verschiedene entwicklungsmäßige Beziehungen (siehe PVP, 106ff), worauf hier jedoch nicht näher eingegangen werden kann. In Übereinstimmung mit der Grazer Schule behauptet Weber, dass auf der gegenständlichen Seite den „produzierten“ Vorstellungen ideale Gegenstände höherer Ordnung entsprechen. Wie bereits ausgeführt, bauen diese Gegenstände notwendigerweise auf Gegenständen niedrigerer Ordnung auf. Ähnlichkeit ist z.B. ein Gegenstand höherer Ordnung, der etwa auf zwei Vierecken als realen Gegenständen niedrigerer Ordnung basiert. Hierbei handelt es sich um eine Fundierung; Weber selbst spricht vom Verhältnis der Fundation. In der Abhandlung von 1928 führt er indes ein weiteres Verhältnis an, nämlich das Verhältnis der gegenständlichen Teilhabe (griechisch PHW|FHLQ), welches nach ihm von der Fundierung grundverschieden ist. So meint er, dass ein Teilhabe-Verhältnis ausschließlich zwischen Gegenständen gleicher Ordnung besteht, wobei ein bestimmter Gegenstand nicht auf einem anderen aufbaut, sondern vielmehr in ihm „eingeschlossen“ ist. Ein Beispiel: Das allgemeine Viereck ist in einem 5. Die Grazer Schule unterscheidet in der Regel folgende Produktionsarten: Gestalt-, Ähnlichkeits-, Verschiedenheits-, Lage- und Verbindungsvorstellungsproduktion. Siehe dazu Rudolf Ameseder (1904), 506ff. 6. Weber unterscheidet 5 Hauptarten der VP, die insgesamt in 14 Unterarten zerfallen.

32

bestimmten einzelnen Viereck eingeschlossen bzw. es ist „,Teilnehmer‘“ (PVP, 14) des zugeordneten einzelnen Vierecks. Ein Fundierungsverhältnis liegt bei jeder aktmäßigen, intentionalen, surrogativen und teilweise auch bei der inhaltlichen VP vor; ein Teilhabe-Verhältnis hingegen tritt bei inhaltlicher VP auf. 3. Die intentionale Vorstellungsproduktion7 In der Abhandlung über VP widmet sich Weber schließlich auch der Intentionalitätstheorie und behandelt sie umfassend in einem eigenen Abschnitt. Weber erörtert zunächst das Erlebnisschema Inhalt – Akt – Gegenstand, das er von Meinong übernommen hat, und führt diesbezüglich einige weitere Distinktionen ein. Bei Vorstellungen kann man also zwei psychische Komponenten unterscheiden: Inhalt und Akt. Ihnen kommen verschiedene Funktionen zu: Während vom Inhalt abhängt, was wir uns präsentieren, gibt der Akt Antwort auf die Frage, wie bzw. auf welche Weise wir uns etwas präsentieren; dabei unterscheidet Weber zwischen ernstmäßigem und phantasiemäßigem Vorstellen. Auch bei Weber sind Inhalt und Akt des Vorstellens miteinander real untrennbar verbunden; wir können sie nur mittels logischer Analyse getrennt behandeln, sie sind nämlich logische Bestandteile des Vorstellens. Der Gegenstand andrerseits ist in der Regel nicht-psychischer Natur und vom Vorstellen unabhängig. Diese Ausführungen finden sich bereits in Webers Frühschriften; in der genannten Abhandlung legt er eine neue Analyse der Intentionalität vor. Weber meint nun, dass die Intentionalität des Vorstellens mit der Inhalts- sowie mit der Aktkomponente des Vorstellens allein noch nicht gegeben ist bzw. sie ist von der inhaltlich-aktmäßigen Struktur des Vorstellens zu trennen. Zur Bekräftigung dieser Behauptung führt er mehrere Argumente an. Er sagt, dass die Intentionalität des Vorstellens erst mit der ganzen Vorstellung vorhanden ist, da sie „beide Fragen [betrifft], auf was die Vorstellung gerichtet und wie sie auf dieses [et]was gerichtet ist“ (PVP, 78). Demgemäß ist sie als besondere reale und nicht bloß logische Komponente des Vorstellens zu betrachten. Ferner gilt, dass wir erst mittels der Intentionalität zur inhaltlich-aktmäßigen Struktur des Vorstellens gelangen können: Insofern wir mittels Vorstellungen auf verschiedene 7. Siehe Tanja Pihlar (2004), 266–268.

33

Gegenstände gerichtet sind, können wir darauf schließen, dass diese Vorstellungen dadurch selbst verschieden sein müssen bzw. dass sie verschiedene Inhaltskomponenten besitzen (analog gilt für die Aktkomponente: Aus der Tatsache, dass man sich mit Vorstellungen auf unterschiedliche Weise auf dieselben Gegenstände beziehen kann, ergibt sich, dass sie verschiedene Aktkomponenten aufweisen müssen). Ein weiteres Argument folgt aus der Tatsache, dass der Inhalt-Akt-Struktur des Vorstellens auf der gegenständlichen Seite kein Gegenstand als solcher entspricht. Weber meint in der Abhandlung von 1928, dass jeder Gegenstand – analog den Erlebnissen – aus zwei Komponenten besteht (siehe PVP, 79; Anm. 22), und zwar aus einer Gegenstands- und einer Faktizitätskomponente. Demgemäß ist dem Inhalt auf der gegenständlichen Seite eine Gegenstandskomponente und dem Akt eine Faktizitätskomponente zugeordnet. Dies kann man schematisch auf folgende Weise darstellen: Inhalt – – – – – Gegenstand Erlebnis

Gegenstand Akt – – – – – Faktizität

Weber behauptet, dass die inhaltlich-aktmäßige Struktur des Vorstellens einen Gegenstand nur potentiell und noch nicht aktuell präsentiert; wir können uns einen Gegenstand erst dann tatsächlich präsentieren, wenn wir darauf intentional gerichtet sind. Außerdem meint er, dass die inhaltlichaktmäßige Struktur des Vorstellens von der Intentionalität des Vorstellens real und erfahrungsmäßig abtrennbar ist, was bedeutet, dass es wohl auch nicht-intentionale Vorstellungen gibt, welche lediglich die inhaltlichaktmäßige Struktur besitzen und auf keinen bestimmten Gegenstand hinweisen. Derartige Vorstellungen sind demgemäß als nicht-intentionale zu bezeichnen. Dies steht genau genommen im Gegensatz zu Brentano, welcher die Existenz gegenstandsloser Erlebnisse nicht zulässt. Weber verfolgt diese seine Ansicht auch im Buch Philosophie (1930) weiter, in dem er zwischen vor- und nachintentionalen nicht-intentionalen Erlebnissen8 unterscheidet. Welche Erlebnisse sind nun nach Weber nicht-intentional? Er meint, dass zu den nicht-intentionalen Erlebnissen die „primitivsten“ und 8. Siehe auch Seppo Sajama & Matti Kamppinen (1987).

34

„ursprünglichen“ Vorstellungen zu zählen sind, welche er mit einfachsten Empfindungen gleichsetzt (da die meisten Empfindungen bereits intentionalen Charakter besitzen). Als Beispiel für solche Empfindungen führt Weber jene an, welche wir im Tiefschlaf erleben sowie sogenannte „Organempfindungen“ (PVP, 81f ). Nach Weber sind derartige nichtintentionale Empfindungen auch entwicklungsmäßig primär und gehen allen anderen Erlebnissen voraus. Ein Neugeborenes z.B. kann – laut Weber – nur potentiell empfinden; sein Empfinden ist zwar inhaltlichaktmäßig bestimmt, jedoch ist es noch auf keinen bestimmten Gegenstand gerichtet. Dazu behauptet er, dass ein intentionales Erlebnis sich zu einem nicht-intentionalen wandeln kann. Ein Erlebnis kann einen intentionalen Charakter bekommen bzw. verlieren, wie z.B. in dem Falle, wenn ich Konturen und Farben eines Baumes sehe, jedoch noch keine Vorstellung von diesem Baum habe, oder ich kann mir einen Baum noch weiter vorstellen, obwohl ich ihn in der Tat nicht mehr sehe. Solche Erlebnisse haben lediglich die inhaltlich-aktmäßige Struktur, sind indes nicht-intentional. In diesem Zusammenhang spricht er vom potentiellen oder vegetativen Vorstellen (dieser Ausdruck ist hier als Terminus technicus gebraucht). Ein derartiges Vorstellen hat keinen Gegenstand, ihm entspricht auf der gegenständlichen Seite kein Gegenstand, sondern „ein (positives) Nichts“.9 Das vegetative Vorstellen ist seiner Natur nach ein passives Erlebnis, auf welchem intentionale Vorstellungen aufbauen, deren notwendige Grundlage es dann bildet. Außerdem meint Weber zu dieser Zeit, dass intentionale Vorstellungen nicht einfach vorgegeben sind, sondern produziert werden – in diesem Zusammenhang spricht er von intentionaler VP. Hierbei unterschiedet er vier Intentionalitätsstufen: (a) ein nicht-intentionales oder bloß potentiell präsentierendes Vorstellen, (b) ein präsentierendes intentionales Vorstellen, (c) ein isolatives intentionales und (d) ein rationales intentionales Vorstellen. Dabei gilt, dass jedes intentionale Vorstellen höherer Stufe vom jeweiligen niedrigeren Vorstellen abhängig ist bzw. auf ihm aufbaut: Isolatives Vorstellen baut unmittelbar auf präsentierendem intentionalem Vorstellen und rationales Vorstellen auf isolativem und präsentierendem intentionalem Vorstellen auf, während präsentierendes intentionales Vorstellen auf nicht-intentionalem Vorstellen fußt. Außerdem unterscheiden sich die erwähnten Erlebnisse ihrem Charakter nach: Sie sind entweder 9. Nach Weber ist das Nichts als positiv zu bezeichnen, insofern auf ihm andere Gegenstandskategorien aufbauen, welche er als „Was“, „Das“ und „Darum“ bezeichnet.

35

passiv oder aktiv, genauer gesagt: Jedes Erlebnis auf niedrigerer Intentionalitätsstufe ist als passiv und das entsprechende Erlebnis auf höherer Stufe ist als ein aktives Erlebnis zu betrachten, wobei es sich nach Weber um das Produzieren des Vorstellens handelt, das heißt, um das Übergehen von passiven zu entsprechenden aktiven intentionalen Vorstellungen. Wenden wir uns den folgenden Intentionalitätsstufen zu: Das präsentierende Vorstellen Auf dieser Stufe findet sich das, was Weber „aktuelles Präsentieren“ nennt: Ein Gegenstand wird tatsächlich präsentiert – wenn wir einen Ton hören, haben wir die Vorstellung vom Ton, wenn wir eine Farbe sehen, haben wir die Vorstellung von der Farbe usw. Ein Gegenstand wird hierbei zwar vorgestellt, jedoch wird er nicht als etwas Selbstständiges erfasst, das man von anderen Gegenständen unterscheiden könnte. Deshalb bezeichnet ihn Weber mit dem Indefinitpronomen „was“ [das „Was“]. Das isolative Vorstellen Diese Intentionalitätsstufe lässt sich mit einem Beispiel verdeutlichen (PVP, 83): Wenn wir die Fassade eines Gebäudes, welches mit Ornamenten bemalt ist, betrachten, erleben wir hierbei verschiedenartige Vorstellungen. Zunächst ist unser Vorstellen nicht-intentional, wie das Vorstellen einzelner Linien, Schatten, Farben usw. Danach wird dieser Vorstellungsgegenstand sozusagen „gleichmäßig“ (ebd.) präsentiert; dies bedeutet, dass unser Vorstellen zwar als intentional zu bezeichnen ist, jedoch die Fassade (sowie ihre Details) dabei nicht als selbstständige Erscheinung erfasst wird. Dies geschieht erst auf der nächsten Intentionalitätsstufe, in einem Augenblick, wenn „unser Präsentieren der gesamten Fassade gewissermaßen wogend, stürmisch [wird], … [sobald] wir nämlich einige ihrer Teile vorstellungsmäßig sozusagen aus der Fülle anderer loszulösen und [sie] in diesem Sinne … als eigene, selbsteigene Phänomene zu betrachten beginnen, was wir gewöhnlich auch äußerlich mit den Ausdrücken, wie ,das!‘, ,dieses!‘, ,jenes!‘ u.s.w. kundgeben“ (PVP, 84). Unter isolativem Vorstellen ist also ein Vorstellen zu verstehen, welches uns die Gegenstände als selbstständige Entitäten, die sich von ihrem Hintergrund abheben, präsentiert. Einen derartigen Gegenstand bezeichnet Weber mit dem Wort „das“ [das „Das“].

36

Das rationale Vorstellen Rationales (oder auch begriffliches) Vorstellen fußt auf beiden vorangehenden Vorstellungsarten, nämlich auf präsentierenden und auf isolativen Vorstellungen. Nach Weber ist ein derartiges Vorstellen „,begründet‘“ oder „,motiviert‘“ (PVP, 86), was bedeutet, dass wir mit dem Vorstellen zugleich wissen, warum wir uns einen Gegenstand auf eine bestimmte Weise vorstellen – hier handelt es sich „um einen besonderen und innerlich wahrnehmbaren Aspekt“ (ebd.) der Intentionalität der Vorstellung. Als Beispiel dafür wird die mathematische Definition eines Kreises als „,einer solchen geschlossenen Linie, welche an allen ihren Punkten von einem Punkte, welcher [sich] außerhalb [von] ihr … befindet, gleich entfernt ist‘“ (PVP, 85f ) angegeben. Wenn ich also einen Kreis betrachte, habe ich folgende Vorstellungen: a) Mein Vorstellen ist zunächst bloß potentiell präsentierend, was bedeutet, dass es zwar eine inhaltlich-aktmäßige Struktur besitzt, aber noch nicht auf den Kreis selbst gerichtet ist. b) Die zweite Vorstellung ist präsentierend intentional, insofern ich mir den Kreis in der Tat vorstelle, ihn aber noch nicht als etwas Selbstständiges erfasse. c) Ich kann nun den Kreis als selbstständige Entität erfassen und ihn zugleich klar von anderen Gegenständen unterscheiden – mein Vorstellen ist hierbei als isolatives zu bezeichnen. d) Mein Vorstellen ist rational, da ich schließlich auch den Grund kenne, warum ich mir den Kreis gerade so und nicht anders vorstelle – und zwar „,weil‘ der Kreis so und so beschaffen ist, ,darum‘ stellen wir ihn uns so und nicht anders vor“ (PVP, 86). Für das rationale Vorstellen ist demnach charakteristisch, dass es „diese Erscheinungen nicht nur zu seinen ,Objekten‘ oder ,Gegenständen‘, sondern zugleich zu seinen ,Gründen‘ oder ,Motiven‘ hat“ (ebd.). Deshalb bezeichnet Weber solche Gegenstände mit dem Ausdruck „darum“ [das „Darum“]. Dazu meint er, dass die Intentionalität bei Vorstellungen sozusagen „,kumuliert‘“ (siehe PVP, 105): Die Aktivität des isolativen Vorstellens z.B. umfasst zugleich die Aktivität des präsentierenden intentionalen Vorstellens, oder schließt ein, wie Weber es ausdrückt, dass „jedes erfahrungsgemäss gegebene intentional höhere Erlebnis bereits selbst zugleich das beigeordnete niederere Erlebnis ist und nicht umgekehrt“ (PVP, 96).

37

IV. Die neue Einteilung der Erlebnisse Auf Grund der eben skizzierten unterschiedlichen Stufen von Intentionalität legt Weber eine neue Einteilung der Erlebnisse vor, welche von der (überkommenen) Klassifikation der Erlebnisse in vier Grundklassen, d.h. in Vorstellungen, Gedanken, Gefühle und Strebungen, durchaus unabhängig ist. Nach dieser Einteilung werden die Erlebnisse zunächst in zwei Klassen geteilt, und zwar in animalische und in geistige.10 Animalische Erlebnisse zerfallen ferner in nicht-intentionale oder vegetative Erlebnisse und in präsentierende intentionale Erlebnisse; geistige Erlebnisse dagegen spalten sich in isolative und rationale intentionale Erlebnisse auf. Diese Einteilung lässt sich schematisch folgendermaßen darstellen: Erlebnisse animalische nicht-intentionale oder vegetative

geistige

intentional präsentierende

isolative intentionale

rationale intentionale

Unter diesen Erlebnissen sind vielfache Beziehungen zu konstatieren: Die geistigen Erlebnisse sind von den animalischen abhängig, was bedeutet, dass es einerseits Subjekte gibt, welche ausschließlich animalische Erlebnisse besitzen, sowie andrerseits Subjekte mit animalischen und geistigen Erlebnissen. Ähnliche Relationen lassen sich auch innerhalb der beiden Erlebnisklassen feststellen: Die präsentierenden intentionalen Erlebnisse sind von den nicht-intentionalen abhängig und ebenso sind die rationalen intentionalen Erlebnisse von den isolativen abhängig. Nach Webers Meinung ist diese neue Einteilung der Erlebnisse als die grundlegende Klassifikation aufzufassen und völlig unabhängig von den bekannten Einteilungen der Erlebnisse in drei oder vier Hauptklassen. Dafür führt Weber folgende Gründe an: Erlebnisse besitzen nach der „Grazer“ Einteilung eine „architektonische Struktur“, wobei Erlebnisse höherer Ordnung notwendigerweise auf jenen niederer Ordnung aufbauen. Dabei sind die Erlebnisse niederer von solchen höherer Ordnung real-numerisch verschieden, was für Erlebnisse nach Webers neuer Einteilung nicht gilt. 10. Die Unterscheidung zwischen animalischen und geisteigen Erlebnissen finden wir bereits in Veber (1927), 15ff.

38

Wie schon erwähnt, akkumuliert und umfasst damit jedes Erlebnis auf höherer Stufe alle relevanten Erlebnisse auf niedrigeren Stufen: Isolatives intentionales Vorstellen ist zugleich als intentional präsentierend und als potenziell präsentierend zu bezeichnen. Nach Weber besteht zwischen vegetativ-animalischen und geistigen Erlebnissen ein real-funktioneller Unterschied (siehe PVP, 102; Anm. 30): Vegetativ-animalische Erlebnisse sind ohne geistige Erlebnisse möglich, während geistige stets vegetativanimalische Erlebnisse voraussetzen; zwischen beiden besteht also ein realfunktioneller Unterschied. Die neue Einteilung der Erlebnisse ermöglicht auch eine Differenzierung der Subjekte, die dann in vegetative, animalische und geistige Subjekte geschieden werden, welche ihrerseits voneinander qualitativ, ja sogar unüberbrückbar verschieden sind. Weber hat in seiner Abhandlung von 1928 die „Grazer“ Produktionstheorie in mehrfacher Hinsicht wesentlich modifiziert: In der Auffassung der VP selbst, im Verhältnis der Teilhabe, in Zuerkennung von verschiedenen Intentionalitätsstufen, durch die Einführung von mannigfachen Produktionsarten usw. Es ist jedoch fraglich, ob derlei Differenzierungen weiterer Produktionsarten de facto zugunsten seiner „Innovationen“ sprechen und nicht eher zu unnötigen Entwicklungen führen. Seine darin niedergelegten Theorien können in mehreren Punkten kritisiert werden, wie z.B. hinsichtlich der Behauptung, dass bereits Empfindungen produziert werden, welche von den zeitgenössischen Theorien sehr abweicht. Es ist im Auge zu behalten, dass manche dieser Lehrmeinungen, welche Weber speziell in dieser Abhandlung entwickelt hat, in seinen späteren Schriften nicht mehr vorkommen. Andrerseits finden wir darin auch fruchtbare Ansätze, wie z.B. die Behauptung, dass es neben intentionalen auch nicht-intentionale Vorstellungen gibt, und dass sie verschiedene Intentionalitätsstufen aufweisen.11

11. Die vorliegende Publikation stützt sich auf Forschungsergebnisse der FWF-Projekte M680 und M776. Für Hinweise aller Art und sprachliche Korrekturen bedanke ich mich bei Wolfgang L. Gombocz, Ulf Höfer und Johann C. Marek.

39

LITERATUR Rudolf Ameseder (1904): „Über Vorstellungsproduktion“, in: Alexius Meinong (Hg.): Untersuchungen zur Gegenstandstheorie und Psychologie, Leipzig: Barth, 481–508. Mauro Antonelli (1994): Die experimentelle Analyse des Bewusstseins bei Vittorio Benussi, Amsterdam – Atlanta: Rodopi. Ludvik Bartelj (1983): Globinski razum in stvarnost II, Ljubljana: Selbstverlag. Alexius Meinong (1899): „Über Gegenstände höherer Ordnung und deren Verhältnis zur inneren Wahrnehmung“, Zeitschrift für Psychologie und Physiologie der Sinnesorgane 21, 182–272. Nachdruck in: Alexius Meinong Gesamtausgabe, Bd. II, 377–471. — (1921): „A. Meinong“, in: Raymund Schmidt (Hg.): Die deutsche Philosophie der Gegenwart in Selbstdarstellungen, Bd. I., Leipzig: Meiner, 91–150. Nachdruck in: Alexius Meinong Gesamtausgabe, Bd. VII, 1–62. Tanja Pihlar (2004): „Zur Theorie der Intentionalität bei France Weber“, in: Erfahrung und Analyse. Beiträge des 27. internationalen Wittgenstein Symposiums, 8.–14. August 2004, hg. von Johann Christian Marek und Maria Elisabeth Reicher, Bd. XII, 266–268. Maria E. Reicher (2001): „Die Grazer Schule der Gegenstandstheorie“, in: Th. Binder u.a. (Hg.): Bausteine zu einer Geschichte der Philosophie an der Universität Graz, Amsterdam/New York: Rodopi (Studien zur Österreichischen Philosophie 33), 173–207. Seppo Sajama & Matti Kamppinen (1987): A Historical Introduction to Phenomenology, London/New York/Sydney: Croom Helm. Wolfgang G. Stock (1989): „Die Bedeutung der Theorie der Vorstellungsproduktion der Grazer Schule für die kognitive Wissenschaft“, Acta Analytica 5, 45–63. — (1995): „Die Genese der Theorie der Vorstellungsproduktion der Grazer Schule“, Grazer Philosophische Studien 50, 457–490. Anton Terstenjak (1972): „Franz Webers philosophisches Gedankengut im Umriß“, in: ders. (Hg.), Vom Gegenstand zum Sein. Von Meinong zu Weber, München: Trofenik, 15–64. France Veber (1921): Sistem filozofije, Ljubljana: Kleinmayr & Bamberg. — (1924): Očrt psihologije, Ljubljana: Zvezna tiskarna in knjigarna. — (1927): Idejni temelji slovanskega agrarizma, Ljubljana: Kmetijska tiskovna zadruga. — (1928): „Problem predstavne produkcije“, Znanstveno društvo za humanistične vede v Ljubljani 4, 139–253. [Deutsche Übersetzung: „Über das Problem der

40

Vorstellungsproduktion. Aus dem Slowenischen übertragen von cand. phil. Franz Žekar“, unpubliziertes Manuskript im Weber–Nachlass, Signatur Z.1.] — (1930): Filozofija, Ljubljana: Jugoslovanska knjigarna. — (1985): Estetika, Neuauflage, Ljubljana: Slovenska matica. Andrea Zemljič (1993): „Leben und Werk von Stephan Witasek (1870–1915). Fallstudie zur Ausdifferenzierung der Psychologie aus der Philosophie“, in: Internationale Bibliographie zur Österreichischen Philosophie (IBÖP 1976/1979), bearbeitet v. Th. Binder u.a., Amsterdam – Atlanta: Rodopi, 1–99.

41

Grazer Philosophische Studien 73 (2006), 43–68.

DOES HUSSERL HAVE AN ARGUMENT AGAINST REPRESENTATIONALISM? Thane Martin NABERHAUS Mount St. Mary’s University (USA) Summary It is often said that by rejecting the representationalist model of mind, phenomenology makes a decisive advance over empiricism. Yet despite such pronouncements, the arguments Husserl uses to refute representationalism have received scant critical attention, and upon examination many turn out to be obscure. I argue here that some of Husserl’s best known anti-representationalist arguments fail. I end the paper, however, by suggesting that if these unsuccessful arguments are paired with certain methodological considerations taken from Husserl’s mature philosophy, they may provide adequate grounds for a rejection of representationalism after all.

While representationalism, as a thesis in the philosophy of mind, has been gaining in popularity recently, it is probably fair to say that the older, strictly epistemological position of the same name (also commonly referred to as “representative realism” or “indirect realism”) is regarded by most philosophers today as untenable.1 Once the dominant model of the mind’s relation to the world, representationalism in this sense was subjected to a long succession of withering criticisms by philosophers otherwise as different as Reid, Hegel, Heidegger, Wittgenstein, and Rorty. Like these thinkers, Edmund Husserl, the founder of phenomenology, put the rejection of representationalism at the center of his philosophy. Indeed, it is generally 1. In current discussions in the philosophy of mind, “representationalism” (sometimes “representationism”) is a thesis that concerns the extent to which phenomenal properties may be understood to have representational or intentional content. I will not be engaging these debates here. My concern instead is with the epistemological view that stems from Descartes and Locke, was defended a generation ago by Jackson (in Jackson 1977), and has been defended more recently by BonJour (see BonJour 1999, 2001). Given the contemporary philosophical scene, BonJour’s attempt to resuscitate representationalism cannot but appear as benighted to many philosophers, a fact that is undoubtedly due in no small measure to the influence of Rorty 1979.

agreed, by phenomenologists at least, that one of phenomenology’s decisive advances over the empiricism it emerged from is that it breaks out of an “immanentistic” model of mind according to which the mind has to do only with its own “ideas”, “appearances”, or “representations”. Despite the centrality of the rejection of representationalism to phenomenology, the arguments that Husserl uses to refute representationalism have received scant critical attention. Indeed, for many of those commenting on Husserl, his grounds for rejecting representationalism are thought to be so self-evident and compelling as to require little further examination. Yet close inspection reveals that many of Husserl’s anti-representationalist arguments are much more obscure than they at first appear. In this paper I will attempt to show that a group of Husserlian arguments that have frequently been taken to deal the deathblow to representationalism are in fact unsuccessful. My purpose, however, is not merely critical, nor is my aim to pave the way for a resurrection of Cartesian idea-epistemology, which Husserl was right, I believe, to reject. Instead, in exhibiting the failure of these arguments, I hope to point toward what I think is the proper way for phenomenology to demonstrate the inadequacy of representationalism and thereby to defend one of its foundational principles. Accordingly, at the end of the paper I will outline what I think a successful Husserlian argument against representationalism might look like. I begin by describing the representationalist view of Husserl’s philosophical mentor Franz Brentano, which forms the background for his thinking about intentionality (section 1). I then turn to a series of antirepresentationalist arguments offered by Husserl and argue that they are unsuccessful (section 2). Next, I examine a philosophical problem that presents a challenge to any theory of intentionality: the problem of nonveridical intentionality or “objectless presentations” (section 3). After outlining a representationalist solution to this problem (section 4), I analyze two arguments deployed by Husserl against this solution, and conclude that these arguments, too, fail (section 5). I end the paper by suggesting that although the grounds that Husserl adduces in his early works for rejecting representationalism are weak, if his arguments are paired with considerations taken from his later philosophy — considerations pertaining to the very nature of the phenomenological enterprise — they may provide adequate cause for such a rejection after all (section 6).

44

1. Brentano’s representationalism Representationalism is a view which aims to reconcile the commonsense belief that, especially in perception, we are somehow aware of (and hence in some important sense have knowledge of ) mind-independent entities with another claim that older representationalists typically took for granted and more recent representationalists have explicitly defended, viz., that what we immediately experience are not such mind-independent entities but rather “ideas”, sense-data, or some other aspect of mental states or their contents. For the representationalist, the relation between that which is immediately experienced and the entities of which we are indirectly aware through these experiences is one of representation, where this notion is typically thought of as involving depiction or encoding, though the exact nature of the relation is often left unclarified. Often the representative relation is held to be a mostly faithful one (in some sense of “faithful”), and this claim is ordinarily justified on the basis of the further assumption that the relevant subjective experiences are caused by the external entities they represent. The main features of representationalism, then, are (a) it involves a relation between (certain aspects of ) subjective experiences which are immediately experienced and mind-independent entities which are not, (b) the relation involved is one of representation, which is usually thought of as a kind of depiction or encoding, (c) this representative relation is held to be by and large accurate, and (d) the mind-independent entities in question are held to be key elements in the etiology of the correlative subjective experiences. “Representationalism”, however, names not a natural kind but rather a family of views, and not all of those whom we might want to call representationalists embrace all four of these features with equal enthusiasm. Now, since Husserl’s views on these matters took shape under the influence of his philosophical mentor Franz Brentano, any discussion of his position vis-à-vis representationalism must begin with the latter. Brentano’s thought takes its inspiration from Aristotle, Descartes, Locke, Comte, and Mill, and although he does not agree with (c), his position otherwise fits neatly into the representationalist model described above. Like Comte and Mill, Brentano was a proponent of empiricism, which he understood to be a form of psychology. Psychology, for Brentano, is a discipline comprising two subdisciplines: descriptive and genetic psychology. Descriptive psychology has the task of determining, by means of introspection (or “inner

45

perception”), the “elements of human consciousness and their modes of connection” (Brentano 1982, 1). The business of genetic psychology, on the other hand, is to “specify the causal conditions which attend the particular appearances” (ibid.).2 Like most of Brentano’s work, his theory of intentionality, for which he is best known, falls within the domain of descriptive psychology. Intentionality, indeed, is for him the central descriptive feature of conscious life. What is intentionality? We must begin with a distinction. According to Brentano, conscious appearances, or, as he calls them, “phenomena”, are divided into two main classes: psychic and physical phenomena. By “psychic phenomena” Brentano has in mind mental states, processes, or “acts” such as thinking, perceiving, judging, remembering, hoping, and desiring. These are characterized, in his famous words, by what the Scholastics of the Middle Ages called the intentional (and no doubt also mental) inexistence of an object, and what we would call, though not wholly unambiguously, relation to a content, direction toward an object … or immanent objectivity. (Brentano 1924, 124–25)

In other words, every psychic phenomenon is directed toward some “content” or “object”. But what does Brentano mean by “content” and “object”? As is evident in this passage, Brentano makes no careful distinction between these two notions. The terms “content” and “object” are used, virtually interchangeably, to refer to whatever it is that the psychic phenomenon is directed toward or “about”: “In a presentation [Vorstellung] something is presented, in a judgment something affirmed or denied, in love loved, in hate hated, in desire desired, and so on” (Brentano 1924, 125). This ambiguity can easily lead to confusions, but one important point should not be lost sight of: when Brentano says that a psychic phenomenon is “directed toward an object”, he does not mean that the act is directed toward a thing in the actual, mind-independent world. According to the passage quoted above, a psychic phenomenon is characterized by the “intentional inexistence” of its object, and for Brentano this means, quite literally, that this object exists in the mental act (though not, to be sure, in the spatial sense of “in”). 2. I should state at the outset that I will only be considering the Brentano of the early and middle periods. The reason for this restriction is that it is only the early and middle Brentano who influenced Husserl. Late in his life Brentano came to reject some of the doctrines attributed to him here.

46

Brentano holds that in the paradigmatic case of intentional relatedness, the case of perceptual presentations, the contents or objects of the psychic phenomena are sensible qualities such as shapes, colors, sounds, and warmth. Such qualities, in fact, are precisely what he places under the heading of “physical phenomena”. Physical phenomena, that is to say, are the contents or objects of psychic phenomena. But like psychic phenomena, physical phenomena are just that: phenomena, that is, appearances (Erscheinungen), which means that they have no existence in mind-independent reality (cf. Jacquette 1990/91, 177; Hedwig 1990/91, 51). Indeed, while psychic phenomena, according to Brentano, are “true in themselves” (Brentano 1924, 28) in the sense that there is no gap between how they appear and how they really are, physical phenomena, as will be explained in more detail in a moment, are “mere” appearances in the sense that that which appears in them is in reality quite different from how it appears. In psychic phenomena, in other words, seeming and being coincide, but the same does not hold for physical phenomena: in them there is an ontological gap between appearance and what appears. Indeed, as Brentano frequently stresses, physical phenomena, the objects of intentional acts, would exist even if nothing external to the mind corresponded to them (Brentano 1982, 22; 1930, 87; cf. Brentano 1924, 124 n; 1925, 9 n). It has been hard for interpreters of Brentano to accept this point. Barry Smith complains of a “seemingly unshakeable tendency” on the part of some commentators to “twist Brentano’s words” on the matter of the immanence of the intentional object (Smith 1994, 42). In making this comment Smith may well have had in mind Dagfin Føllesdal, who claims that Brentano “rejects the attempt to save the theory of intentionality by saying that the object should be part of our own consciousness” (Føllesdal 1982, 84–85). There is simply no textual evidence to support this claim, and the evidence for the contrary claim is ample.3 (One passage in the Psychology which makes it clear that Brentano regards all phenomena, including physical phenomena, as mental appearances and not as items of the mind-independent world is one in which he criticizes the theories of several “prominent psychologists” who seem to use the term “phenomenon” in precisely this latter sense: “And thus it turns out that what they call ‘physical appearances’ in fact do not appear to us, indeed that we have 3. Again, I am restricting this claim to the Brentano of the early and middle periods. However, although there is no space for me to argue it here, I believe that even the late Brentano held that mental acts are intentionally directed toward mentally immanent objects, not the mind-independent world.

47

no presentation of them whatsoever — certainly a curious misuse of the name ‘phenomenon’!” [Brentano 1924, 110].) But how, then, are psychic and physical phenomena related to the nonmental, mind-independent world? Brentano’s answer is that it is precisely the task of genetic psychology to determine this relationship. Genetic psychology seeks to establish correlations between the psychic and physical phenomena described by descriptive psychology, on the one side, and the actual worldly conditions that are causally responsible for these mental experiences, on the other. For Brentano, then, the mind is not intentionally related to things in the actual world. Our mental acts are intentionally directed toward physical phenomena, which in turn are produced in us by “things” (Brentano’s preferred term for them) in the actual world through their exerting a certain causal “influence [Einwirkung]” (Brentano 1924, 28) on us. As a result, physical phenomena are for Brentano mere “signs” of the worldly things that causally give rise to them in us and are not adequate representations of those things. Physical phenomena, Brentano tells us, give us knowledge of reality “only in a very incomplete sense”. “That which truly exists in and of itself does not enter into our appearances”, he writes, “and that which does appear is not true. The truth of physical phenomena is, as they say, a merely relative truth” (ibid.). It should be clear that although Brentano does not accept the notion that our subjective experiences give us an accurate representation of the mind-independent world, his view otherwise conforms quite closely to the representationalist theory outlined above. It is this theory that Husserl takes aim at in his early work on intentionality. 2. Some Husserlian arguments against representationalism The argument against representationalism that Husserl seems to favor most is his charge that it entails a “false doubling” of the intentional object. The phrase is taken from the 1894 essay “Intentional Objects”, which is devoted in large part to attacking the representationalist view. Framing his argument using the Brentanian language of affirmative and negative existential judgments (that is, judgments that affirm or deny the existence of something), Husserl claims that as long as a presentation is veridical (“corresponds to truth”) its intentional object “can be … none other than the true object” (Husserl 1979, 308). There is no difference,

48

he argues, between the “intentional object” and the real object in the mind-independent world: “The same object that is merely presented in the presentation is set forth as truly existing in the corresponding affirmative judgment. Whether we merely present Berlin or judge it to exist: in any case it is Berlin itself [that we mean]” (ibid.). Husserl evidently considers this argument especially important and compelling, for he repeats it, with emphasis, in a well-known and widely admired passage in the Logical Investigations. “It need only be said to be acknowledged”, he writes there, “that the intentional object of a presentation is the same as its real or (as the case may be) external object and that it is absurd to distinguish between the two” (Husserl 1984, 439; cf. Husserl 1976, 111–13, 207–08). Husserl’s claim, then, is that the presented object is nothing distinct from the real object at which the presentation aims. What we intend, present, or mean in an intentional act is not some mentally immanent object that in turn stands in for an object in the real world. We intend the real object directly. Since Husserl is claiming that the presented or intended object is the same as the real object, we might call this the “identity thesis”. Is this argument sufficient to undermine representationalism? It seems clear that acceptance of the identity thesis would constitute a sufficient reason to reject representationalism. But why should we accept the identity thesis? Husserl has offered us no reasons to do so. Brentano asserts that there is, in addition to the mentally immanent intentional object, a mind-independent thing corresponding to it, a thing that we hypothesize to exist but are not intentionally related to. Husserl denies that this is so, holding that there is only one, identical object associated with the mental act. But what is his evidence for this denial? We have nothing, it appears, but the unsupported assertion that there is only one object, not two — an assertion that is unlikely to impress the representationalist. In short, contrary to Husserl’s strenuous insistence in the Logical Investigations that it is “absurd” to distinguish between the intentional object and the real object, there seems to be no obvious absurdity in the hypothesis of such a duality of objects, and the false doubling “argument” does nothing to unsettle that hypothesis. Husserl has not succeeded in showing what is “false” about the representationalist doubling of the object of a presentation; he has instead simply begged the question at issue. Still, it might be thought that the identity thesis is so inherently reasonable that the accusation of question begging is mere carping. Is it not simply implausible to suggest that mental acts are directed toward mentally

49

immanent objects? Does that suggestion not amount to an attempt to resuscitate the long since discredited “idea theory”? John Haldane expresses just this intuition. Since, on the “theory of ideas”, any relationship between the intentional (that is, mentally immanent) object and something beyond it is “external to the intentionality of the thought”, he says, the worldly object becomes “superfluous”. True, the representationalist schema has the advantage that it helps to explain cases of nonveridical intentionality, where nothing extramental corresponds to the thought. (Indeed, as we will see below, this is perhaps the strongest argument in favor of representationalism.) Nevertheless, says Haldane, the representationalist view “clashes with a deeper and surer supposition, namely, that some of one’s thoughts bear upon the world and not upon ideas”. What this consideration suggests, Haldane concludes, is that “far from it being the real object that is superfluous it is the idea theorists’ mental objects, or entia rationes, that are uncalled for — and indeed are unhelpful” (Haldane 1989, 4–5). This appeal to realist intuitions certainly has a measure of plausibility, but as an argument it is scarcely more effective than Husserl’s dogmatic assertion of the identity thesis. After all, representationalists from Descartes to Locke to Brentano were apparently unmoved by such intuitions (the pull of which they, too, as ordinary human beings, presumably felt), and we can imagine that they would have been unfazed by a Johnson-style refutation of representationalism (“I refute Brentano thus!”).4 We would do well, then, to see whether there might not be another way to argue against representationalism. Does Husserl have any other tricks up his sleeve that might help him out here? In the fifth Logical Investigation Husserl offers an argument that appears to supply the desired refutation. The argument occurs at the beginning of the “Appendix to paragraphs 11 and 20” and is directed against what Husserl calls the “image theory” and the “sign theory” — perhaps the two most important species of representationalism. Husserl describes the image theory as the view which says, “‘Outside’ is, at least under certain circumstances, the thing itself; in consciousness there is an image which does duty for it” (Husserl 1984, 436). The sign theory is exactly similar to 4. Moreover, a defender of Brentanian representationalism would certainly want to point out that although for him our mental states are not intentionally related to the mind-independent world, some of them do stand in a relation of signification with that world, inasmuch as physical phenomena are regarded as signs of worldly, material things. There is, therefore, an important sense in which these states do “bear upon the world”, for Brentano.

50

this view except that it substitutes a mentally immanent sign of the “thing itself ” for the image theory’s mentally immanent image. Husserl argues that neither the image theory nor the sign theory can explain how a mentally immanent object could represent a nonmental object. As Bernhard Rang has shown, the argument, which Husserl develops in greater detail in Ideas I (cf. Husserl 1976, § 43), proceeds by way of a reflection on what it means for something to represent or stand in as proxy for something else. The key to Husserl’s argument is the observation that for the image/sign theory, the represented object is in principle unavailable to consciousness. Given that this is so, it makes no sense, Rang explains, to talk of the (consciously accessible) mental object as a representation, image, or sign of the (consciously inaccessible) real object, since it belongs to the very idea of something being a representation of something else that both items, ideally at least, be available to him who takes the one to be a representative of the other. I cannot take a portrait, even one of a historical person, as a portrait, without there being the ideal possibility of comparing it to the person it is supposed to be a portrait of (even if for me that comparison is factually impossible). It makes no sense to take thunder as a sign of an impending storm without having experienced storms in the past and without the assumption that it will be possible to experience storms in the future (Rang 1975, 116–121). Even the hypothetical supposition (made by Descartes, Locke, and Brentano) that the mental image or sign is in some way similar to the mind-independent object will do us no good here, Husserl argues, since for consciousness, which by assumption is only in possession of the image, [this similarity] means nothing … The similarity between two objects, however great it may be, does not make the one to be an image of the other. It is only through the ability of a presenting ego to make use of a similar as an image-representation of a similar, to have the one intuitively present and in its place to mean the other, that an image can become an image at all. (Husserl 1984, 436)

Thus the image theory and the sign theory founder on their supposition that something that is available to consciousness could stand in as representative for something that is not merely de facto but in principle unavailable to consciousness. To suppose this is to violate the very sense of what it is for something to represent something else. This argument also sheds some light on Husserl’s obscure claim in the passage quoted above that it is “absurd” to distinguish between the inten-

51

tional object and the real object. The German word translated as “absurd” is widersinnig — literally, “contrary to sense”. In light of the argument we have just examined, this claim can be interpreted as meaning that the representationalist supposition that there are two objects — one available to consciousness, one not — associated with a (veridical) presentation, and that the one represents the other, is “contrary to the sense” of what it is for something to represent something else. Perhaps, then, Husserl’s “false doubling” argument is not so question begging after all. The foregoing considerations certainly do seem to show that the notion that a mentally immanent object could be taken as a representative of a nonmental, consciously inaccessible object is nonsensical. Nevertheless, a defender of representationalism might object that the argument assumes that representationalism holds that the mental representative of the nonmental thing in the world is taken as a representative (image, sign, proxy) of it, whereas that is not what it in fact holds. Thus Brentano’s sign theory, for example, might be defended against Husserl’s argument by saying that although intentionally inexisting physical phenomena are in fact signs of mind-independent things, natural, unreflective consciousness does not take them to be signs. After all, representationalism does not purport to be a description of how things seem to us, i.e., of how we experience them. It is rather a philosophical theory about experience. The same point tells against Husserl’s anti-representationalist claim in the Ideas that, contrary to the representationalist doubling of objects, “only one [object] is found to be present and possible” (Husserl 1976, 207). For representationalism does not claim that we experience two objects (a mentally immanent object and a real object that it represents); it claims that that is how things in fact are.5 It appears, then, that the argument from the Logical Investigations that we have been considering is question begging in the same way the false doubling argument from “Intentional Objects” is. Like this latter argument, it assumes that since only one world appears and can appear to us, the representationalist postulate of a relationship between the world of our experience and a world of things lying beyond our experience is inescapably nonsensical. But this assumption involves a misunderstanding, since representationalism does not dispute that we experience only one world of objects. It simply tries to explain the origin of our experiences by pos5. A similar objection might be made to the well-known argument, presented by Husserl in “Intentional Objects” (Husserl 1979, 306), in the “Appendix” (Husserl 1984, 437), and later in the Ideas (Husserl 1976, 111), that representationalism leads to an infinite regress.

52

tulating another, unexperienced world of objects. So far, Husserl would not seem to have given us any reason to reject that postulate. 3. The problem of objectless presentations Let us, then, for the sake of argument at least, assume that representationalism is a viable theory of veridical presentation, that is, that it successfully explains how we present things to ourselves (in perception and otherwise) that really do exist as we present them to be. That assumption by itself does not make representationalism a viable theory of mental activity, for there is a formidable challenge which has yet to be examined, one that any theory of intentionality must meet. It comes in the form of a philosophical problem concerning nonveridical intentionality — presentations of things that do not exist as we present them to be. This problem, which in its essentials dates back to Plato (cf. Theaetetus, 189a) and which came to occupy an important place in English-language philosophy through the work of Bertrand Russell (cf. Russell 1905), was known to Brentano, Husserl, and their contemporaries under the title of the problem of “objectless presentations”. It can be seen as arising from a conflict between two intuitively plausible theses, one held by Brentano and the other held by Bernard Bolzano. Consider Brentano’s claim that every psychic phenomenon is characterized by the “direction toward an object”, a claim we might call the “intentionality thesis”. The intentionality thesis seems hard to deny, whatever else we might think about Brentano’s representationalism. But what then should we say about presentations like “the present king of France”, “a golden mountain”, “a round square”, and “the largest prime number”? Towards what objects are such presentations directed? It seems clear that in some sense the objects of such presentations, whether for factual or necessary reasons, do not exist, and this led Bolzano to claim that some presentations are “objectless”, that is, that they simply lack objects. If Bolzano is right and there are objectless presentations, however, it would seem that Brentano’s intentionality thesis must be abandoned. Thus we have two incompatible theses, each with strong intuitive pull (cf. Philipse 1987, 306; Schuhmann 1991, 49–50). The importance of this paradox lies in the fact that any adequate theory of intentionality must be able to explain both veridical acts of intentionality — intentional acts that “hit their target” — and nonveridical acts. A theory of intentionality that can explain how I can have a veridical perception

53

of a yellow traffic signal, for example, but cannot explain how it is possible that I can perceptually present to myself a traffic signal as yellow when it is in fact red must be regarded as seriously inadequate. For presentational mistakes, whether they be perceptual hallucinations, memorial errors, or judgmental failures of reference, are an important part of conscious life, and any theory that fails to do them justice is as good as useless. Let us say, therefore, that an adequate theory of intentionality must satisfy what we can call the nonveridical intentionality requirement, that is, the requirement that it be able to explain nonveridical intentions. What, then, we might wonder, does Brentano have to say about the problem of objectless presentations? The answer, unfortunately, is very little. Although his theory of intentionality leads directly to the problem of nonveridical intentionality, Brentano never seems to have tackled the issue head-on. On the contrary, as Michael Dummett remarks, “in his writings, what promise to be discussions of it always slide off into some other topic, leaving what seems to us the central problem unresolved” (Dummett 1993, 28).6 Indeed, it is hard to see how the problem of objectless presentations could be resolved, if we begin with Brentano’s representationalist theory, and if we accept his thesis that every mental act is directed toward an object. If all mental acts are directed toward mentally immanent objects, how are we to distinguish those presentations that are thought to reach their targets successfully from those that do not? There is surely an important difference between my presentation of the current king of France and my presentation of the current queen of England, for instance, and yet Brentano’s theory seems to have no way of capturing it. Perhaps, one might say, it could be accounted for by appealing to causal explanations derived from genetic psychology in order to explain why our presentations sometimes fail to correspond to reality. Yet such explanations, while they might seem like a plausible way of accounting for hallucinations, mirages, and the like hardly seem capable of accounting for the difference between, say, “the smallest positive prime number” and “the largest prime number”. The problem is a deeper one than that, and it takes us into the heart of Brentano’s theory of intentionality.7 6. It is worth emphasizing this point in view of the widely shared misconception that Brentano’s distinction between psychic and physical phenomena was designed precisely to address the problem of nonveridical intentionality. As Linda McAlister shows (in McAlister 1977), it wasn’t. 7. Nonveridical intentionality isn’t just a problem for Brentano, however, or even for repre-

54

4. A representationalist solution to the problem of objectless presentations Unlike their teacher, a number of Brentano’s students were concerned with the problem of objectless presentations. The first to recognize its significance seems to have been Kasimir Twardowski, who discussed the issue at length in his influential book On the Doctrine of the Content and Object of Presentations (1894). This discussion, and Twardowski’s proposal for solving the problem of objectless presentations, in turn generated a spark of interest in the issue among the other members of the Brentano school and led to a number of other proposed solutions. To understand these proposals, we must first know something about an innovation introduced by two Brentanists that was designed to cope with another problem bequeathed to Brentano’s students by their teacher: the ambiguous relation between the “content” and “object” of a presentation. In place of this ambiguity, Alois Höfler and Alexius Meinong proposed a sharp distinction. A presentation’s content, as they defined it, (a) “exists mentally ‘in’ us” and (b) is a “quasi-image” or “sign” of the presentation’s object. The presentational object, for its part, (c) exists “in itself ” and “independent of thought” and (d) is that entity “towards which our presenting and judging are as it were directed” (Höfler and Meinong, 1890, § 6; quoted in Twardowski 1982, 4). On the one hand, therefore, Höfler and Meinong endorse Brentano’s representationalist claim that the content of a standard, veridical presentation is a “sign” of the corresponding really existing, mind-independent thing. At the same time, however, they differ crucially from their teacher in assigning to this mind-independent thing the role of the direct object of the presentation: it, and not the content, is that “toward which” our presenting and judging acts are “directed”. In other words, the intentional relation is now understood not as an exclusively intramental relation, as Brentano had understood it, but instead as holding between the mind and the extramental world. We have, then, a substantially different version of representationalism from the Brentanian version. Whereas Brentano’s position, which we might call “mind-directed representationalism”, holds that intentional acts aim at mentally immanent objects, Höfler and Meinong, whom we might call “world-directed representationalists”, claim that mental acts are intentionally directed toward extramental objects by means of a mentally immanent content. sentationalism generally. Direct realism, for example, faces similar difficulties. For if we assume that our mental acts are directly related to objects in the mind-independent world, what are we intentionally related to when what we are directed toward fails to exist?

55

In On the Doctrine of the Content and Object of Presentations, Twardowski endorses the world-directed representationalism of Höfler and Meinong, as well as another terminological innovation they introduced, namely, their labeling of the content as an “immanent object” (Twardowski 1982, 4). Although the main or “primary” object of a veridical presentation is a mindindependent thing, Twardowski tells us, the content may be regarded as the “secondary object” (ibid., 18) of the presentation. Every presentation, he writes, therefore has a “double object”, and the activity of presentation accordingly “moves in two directions” (ibid., 14). Twardowski clarifies his view by comparing the activity of presenting to the activity of painting. Just as we say that the object of the painting activity — what is being painted — is in one sense the painting and in another sense the landscape, so, too, the object of the presentational activity is in one sense the mental content and in another sense the mind-independent, real thing. Talk of the “presented object”, therefore, is just as ambiguous as talk of a “painted landscape”. For there are really two “presented objects”: the immanent, secondary object (the content) and the mind-independent, primary object. These distinctions and terminological innovations provide fresh resources for addressing the problem of objectless presentations. Among the strategies developed by Brentano’s students to deal with that problem, the best known is undoubtedly Meinong’s postulation of mind-independent but nonexistent objects or “objectives” (nonexistent round squares, golden mountains, and the like) which serve as the targets of nonveridical intentions. In seeking to preserve Brentano’s thesis that every presentation is directed toward an object, Meinong in effect simply sacrifices the Bolzanian intuition that some presentations lack objects. Every presentation has both a mentally immanent content and a mind-independent object, he proposes, but in some cases this latter object is a special kind of nonexistent object (cf. Smith 1994, 125–26). This suggestion provides a neat way of handling the problem of objectless presentations, yielding as it does a unified theory of intentionality in which the basic structure of the intentional relation is the same for veridical and nonveridical intentions. Of course, this benefit comes at the cost of introducing entities with a rather queer ontological status, a cost many would consider too high. Indeed, for most philosophers, the fact that Meinongianism is committed to an ontology of nonexistent objects is reason enough to dismiss it as unacceptably cumbersome at best, and a philosophical nonstarter at worst. Fortunately, the distinctions drawn by Höfler, Meinong, and Twardowski permit the construction of another solution to the problem of objectless

56

presentations, one that has no such questionable ontological commitments.8 Consider the distinction between the presented object as the intentional content and as the real object of the presentation — between what Twardowski calls the “secondary” and “primary” objects of the presentation. Using this distinction, it would seem that we can make short work of the problem of objectless presentations. For by means of it we can say both that (a) every presentation has an object, since every presentation has at least a secondary object (that is, a content), and that (b) some presentations are objectless, since not every presentation has a real or primary object corresponding to that content. Instead of postulating extramental but nonexistent objects to account for nonveridical intentionality, this solution holds that in the case of nonveridical presentations, the content of the presentation — a mentally immanent object of some sort, such as a mental image or sign — serves as the target of the intentionally directed mental act. Whereas a veridical presentation is directed toward its primary, worldly object, when that object fails to exist the intention is “redirected” to the presentation’s content. In this way, it would seem, the paradox of objectless presentations is easily dissolved. Since this proposal for solving the problem of objectless presentations substitutes the content of a nonveridical presentation for its object, we might call it the “content as ersatz object solution”. Because it is free of commitment to ontologically questionable entities, the content as ersatz object solution seems like an appealing approach to the conundrum of nonveridical intentionality. Indeed, based as it is on a modified version of Brentano’s representationalist view, it might be thought to constitute one of the strongest arguments in favor of representationalism. It should be emphasized that although this solution is based on the “world-directed” rather than the “mind-directed” version of representationalism, it nevertheless depends crucially on the postulation of mentally immanent objects. If the Twardowskian “secondary objects” seem, for world-directed representationalism, to play an almost invisible and indeed even perhaps unnecessary role in veridical presentations, they become 8. Outside the Brentano school, of course, there have been other proposals for solving the problem of nonexistent objects, most notably Russell’s theory of definite descriptions. This approach is now widely recognized as having certain limitations (for example, in regard to fictional objects) and in any event would seem to have little to offer toward the construction of a theory of the mind’s (or subject’s) relation to the world. For an analysis of some limitations of Russell’s “deflationist” approach, see Cartwright 1960.

57

absolutely indispensable for the explanation of nonveridical perception. It is, in fact, precisely the postulation of such immanent objects that makes the content as ersatz object solution seem like such an attractive way of escaping the paradox of nonveridical intentionality. Indeed, if the Meinongian route is rejected, it can seem, as Husserl once suggested, like the only way to solve the problem of objectless presentations: A presentation without a presented object is not conceivable; thus there are no objectless presentations. On the other hand, however, not all presentations have real objects corresponding to them; that is, there are objectless presentations. The contradiction can only be avoided, it seems, by distinguishing between the presented and the real object: there is no presentation without an immanent object, but there are presentations without real objects. (Husserl 1979, 420)

Husserl nevertheless believes that this attempt to find a way out of the difficulty is a dead end. His main argument against this approach, however, is unsuccessful — or so at least I shall argue in the next section. 5. Two unsuccessful arguments It is not necessary here to go into the details of Twardowski’s solution to the problem of objectless presentations. What is worth noting, however, is that although his solution shares certain things in common with the content as ersatz object solution, he does not accept that solution and mounts an argument against it. It will be useful briefly to examine that argument, not least because Husserl accepts it and makes it his own in developing his rejection of representationalism. Twardowski’s attack on the content as ersatz object solution begins with the observation that the content of any “objectless” or nonveridical presentation, as of any presentation whatever, exists just as surely as the presentation itself does. But if that is true, he reasons, then the content of such a presentation cannot serve as the presented object (that is, as what the presentation is directed toward). For unlike the content, the presented object, ex hypothesi, does not exist. In other words, since the content of a presentation always exists (that is, whether or not the presentation has a real object corresponding to it), and since what a nonveridical presentation is directed toward does not exist, it follows that such a presentation cannot be directed toward its content. 58

Twardowski puts this point by saying that a true negative existential judgment — the true judgment that something does not exist — must be about the actual or primary object of the underlying presentation, even though that object does not exist. If we judge, for example, that “An obliquely angled square does not exist”, this judgment cannot be about the content of the underlying presentation, since this content, unlike the obliquely angled square, exists just as much as the presentation itself does. The conclusion we must draw, Twardowski says, is that where nonveridical presentations are concerned, it is not the content that is intended in the associated negative existential judgment, but rather the (admittedly nonexistent) object: the obliquely angled square is not the “something presented” in the sense in which the presentational content is presented, for the content exists. The obliquely angled square is rather “what is presented” in the sense of the presentational object, whose existence in this case is admittedly denied, but which nevertheless is presented as an object. (Twardowski 1982, 24)

However strange it may seem to say that what is presented in such a case is something that does not exist, this is preferable to saying that what is presented in such cases is the presentation’s content. For “the existence of the obliquely angled square, understood as the content of the presentation, cannot be denied”, Twardowski reasons; “the mental content … exists in the truest sense of the word” (ibid.). If this reasoning is correct, it follows that the content as ersatz object solution is incapable of solving the Brentano-Bolzano paradox, for that solution is committed precisely to the idea that nonveridical presentations present their contents.9 Husserl, for his part, clearly agrees with this reasoning. Arguing against the image theory, for example, he writes that the alleged mental image cannot be what is meant, noting that this mental image “after all always exists” (Husserl 9. It is worth reiterating that Twardowski makes this argument, since there is a tendency among commentators to attribute the content as ersatz object solution to him. Schuhmann gives the clearest expression of this misinterpretation: “From here Twardowski then solves the Bolzanian problem of presentations for which there is either de facto (as in the case of a golden mountain) or for essential reasons (as in the case of a round square) no corresponding object. These presentations, too, have a content (an intentional object), but the objects corresponding to them do not really exist. Thus these presentations are not objectless in every sense; rather, their objects are precisely merely presented and not, in addition, real. … He counters the Bolzanian objection of objectless presentations with the distinction between the intentional and the real object” (Schuhmann 1991, 53–54). I read Rang (in Rang 1975, 113–14) as attributing the content as ersatz object solution to Twardowski as well.

59

1979, 306; cf. 309–10). Or again, in an 1896 review of On the Doctrine of the Content and Object of Presentations, he writes: “it could again only be the object itself which is presented, and not the presented object as such [that is, the content], which indeed exists even in the case of contradictory presentations [like “round square”], and which we therefore surely do not intend when we, for example, truly deny that it exists” (ibid., 353 n). Moreover, Husserl argues that Twardowski’s own position falls prey to the very objection to the content as ersatz object solution just discussed, something he could obviously only do if he understood and agreed with Twardowski’s argument (cf. ibid., 330). As far as Husserl is concerned, then, Twardowski’s attack on the content as ersatz object solution is perfectly legitimate. But Twardowski’s argument is flawed, and it is so for reasons that are familiar to us from our discussion of veridical intentionality above. The key point it overlooks is that like the representationalist theory of veridical presentation examined above, the content as ersatz object solution is designed to explain nonveridical presentation, not to describe our experience of presenting nonexistent objects. It proposes that whatever our experience might suggest the direct object of our nonveridical presentations to be, what we are in fact presenting is not an actual, mind-independent object, but instead merely the content of the presentation itself. Hence, what we are in fact presenting, on this suggestion, is something that does exist, even if it seems to us that it does not. It is important to recall that for the representationalist, first-person experience is not a trustworthy guide to the actual, full structure of a presentation, and thus it is not sufficient simply to claim that what we are presenting in a nonveridical presentation is something that does not exist. We may think we are presenting something that does not exist, but according to the content as ersatz object solution, we are wrong. Twardowski’s argument, then, begs the question against the content as ersatz object solution in just the same way that Husserl begs the question against the representationalist account of veridical intentionality. Now, while Husserl agrees with Twardowski’s argument, his main attack on the content as ersatz object solution stems from a different kind of consideration, and it might be thought that this attack is more successful than the Twardowskian argument he endorses. This second Husserlian argument is directed against what he describes as a “popular” solution to the problem of objectless presentations, one that is based on the image theory, which we examined in section 2 above. It is easy to see, however, that this solution is simply a version of the content as ersatz object solution.

60

Recall that the basic tenet of the image theory is the claim that having a presentation involves the having of a mental image or depiction of the presented object. Now, if we accept this claim, a resolution of the paradox of objectless presentations might seem to lie close at hand. For as Husserl says, “just as in general an image can exist while that which it depicts does not, so too here [sc. in the case of presentations]” (Husserl 1979, 304). The idea, then, is that the content of a presentation is a mental image, and that it serves as the direct object of a presentation when the real object is lacking. The mental image, in other words, serves as an ersatz object. After presenting several arguments that specifically target the thesis that every presentation involves the having of a mental image, Husserl proceeds to develop a more general argument that applies to the content as ersatz object solution in any form. Even if the notion that every presentation involves the having of a mental image could be made convincing, he argues, the postulation of mental images — and the same reasoning would apply, mutatis mutandis, to other mentally immanent objects such as signs — can offer no assistance in solving the problem of objectless presentations. The basic problem, Husserl argues, is that the content as ersatz object solution equivocates on the phrase “presented object”, taking it at one moment to be an immanent object and at another moment to be the real thing that is supposedly depicted by that object. But the Brentano-Bolzano paradox, Husserl claims, arises from the fact that it is the same presented object that, in the case of a presentation like “a round square”, seems both to exist and not to exist (ibid., 305). Thus, separating the presented object into two objects (in the case of the image theory, the mental image and the really existing object depicted by it) does nothing to resolve the paradox. For in fact, says Husserl, there is only one presented object, namely, the real, mind-independent object, whether it exists or not: The same Berlin that I present also exists, and the same Berlin would no longer exist, should a judgment be decreed upon it as upon Sodom and Gomorrah. It is the very same centaur Chiron about which I presently speak and which I thereby present that does not exist. (Ibid., 305–06)

Contra the image theory, Husserl is claiming, what we intend or aim at (meinen) in a presentational act is not a mental image, but rather the real object itself, even when that object does not exist. “In having an image, [that is,] this subjective content, in the presenting act”, he writes, we do not aim at the image but rather at the corresponding “external” object — the very object that so frequently does us the injustice of not exist-

61

ing. It is the object itself that we present; we judge about it; it is to it that joy and sadness, wishes and volitions are related. (Ibid., 306.)

“Hence the difficulty persists undiminished”, Husserl concludes, for “the two-sided talk of presentations concerns the same object, and the entire illusion of a solution derives from the fact that we occasionally understand the term ‘presentation’ to mean ‘having an image of something,’ which however is not what is at issue here” (ibid.). Indeed, he says, “[t]he popular aura of this solution could be found seductive, if it weren’t for the fact that the same problem is contained in the analogy [with images] that was supposed to be solved” (ibid., 304). In short, the charge is that the content as ersatz object solution conflicts with the fact (and Husserl takes it to be a fact) that intentional acts are directed toward objects in the real world. It is not difficult to see that this argument is little more than a version of the false doubling argument used against the representationalist account of veridical intentionality. Indeed, in his treatment of the problem of objectless presentations in “Intentional Objects”, Husserl does not distinguish cases where the object of the presentation exists from those where it does not; he apparently takes the underlying point to be the same, whether the presentation is veridical or nonveridical. Many of Husserl’s interpreters (among them Rang 1975, 114; Philipse 1987, 306; Schuhmann 1991, 55–56; and Zahavi 1992, 106–07) have followed him in thinking that this argument delivers the coup de grâce to the content as ersatz object solution. But just as the false doubling argument is unsuccessful in refuting the representationalist account of veridical intentionality, the present argument fails to show that the content as ersatz object solution is untenable as an account of nonveridical intentionality. Husserl argues, for example, that it is “intrinsic to the sense” of the Brentano-Bolzano paradox that in the case of an objectless presentation “the same object which is presented seems both to exist and not to exist” (Husserl 1979, 305). But this interpretation of the problem of objectless presentations simply rules out the content as ersatz object solution by fiat: it assumes without argument that the problem of objectless presentations cannot be solved by appealing to a mental image or some other kind of mentally immanent object as the ersatz target of a nonveridical intention. And in doing so, it begs the central question at issue. The whole point of the content as ersatz object solution is to say that the direct object of a nonveridical presentation is not the “real” object (there is no real object) but rather the presentation’s content, and it is no argument against this view simply to assert that what a nonveridical presentation is in fact directed 62

toward is a mind-independent object. Indeed, is there not a certain plausibility to the idea that the presentation “Atlantis”, for example, is about a mental image of that city? (It certainly seems odd to say that it is about the “real” Atlantis!) In short, Husserl’s “argument” against the content as ersatz object solution amounts to little more than an unsupported assertion that, like veridical presentations, nonveridical presentations are directed toward real (though nonexistent) things, not mentally immanent images or other mental objects. To summarize: Husserl has given us neither a reason to think that the representationalist account of veridical intentionality is untenable nor a reason to reject a quite appealing representationalist account of nonveridical intentionality. Despite his many arguments, representationalism at this point would appear to be a viable theory.10 6. Does Husserl have an argument against representationalism? I have argued that none of the Husserlian arguments so far considered is sufficient to show that representationalism is an untenable theory of intentionality. Nonetheless, I believe that there are resources in Husserl’s thought that can be used to show the inadequacy of representationalism, and I would like, by way of conclusion, to indicate in broad outline what these resources are. Let us return for a moment to the anti-representationalist argument presented in the Appendix to the fifth Logical Investigation. Recall that Husserl there claims that it is “absurd” to distinguish between the intentional object and the real object of an intentional act. This claim is repeated frequently in the Ideas, where Husserl writes, for example: I perceive the thing, the object of nature — the tree there in the garden. That and nothing else is the real object of the perceiving “intention”. A second immanent tree or an “inner image” of the real tree standing out there before me is in no way given, and to make such hypothetical suppositions only leads to absurdity [Widersinn]. (Husserl 1976, 207–08) 10. There are many other, good reasons to reject the notion that mental acts are directed toward mental objects, not the least of which is the fact that it conflicts with common sense, at least in the veridical case. My point is only that however well Husserl may intuitively have understood the shortcomings of immanent-object theories, some of his best-known arguments against them are very thin. For a résumé of better anti-immanent-object arguments, see Jacquette 1990/91, 180–81.

63

In another passage, Husserl tells us that representationalism conflicts with the very “sense of a thing being given [Sinn vom Dinggegebenen] and thereby with the sense of ‘thing as such’”, adding that since this sense is the “absolute norm of all rational talk about things”, any violation of it is “absurd [widersinnig] in the strictest understanding” (ibid., 111). But even if we stress the idea that the representationalist doubling of objects is absurd because it is “contrary to the sense” of something being given in perception, such claims, I argued in section 2 above, are in effect little more than unsupported assertions, and they do little to undermine representationalism. The problem, to put it in Brentanian terms, is that Husserl’s arguments are carried out from the standpoint of descriptive psychology, whereas representationalism is a theory that is at home in (or, more exactly, provides the underlying theoretical framework for) genetic psychology. Brentano does not hold that the things of the real world that impinge on us causally and thereby produce representations of themselves in us are available for psychological description; we only “reach” these things, he holds, via scientific — that is, genetic — hypotheses. And that means that the representationalist theory cannot be undermined on descriptive grounds alone. If, on the other hand, there were grounds for abjuring the standpoint of genetic psychology and remaining solely within the descriptive standpoint, then it might be possible to rehabilitate Husserl’s false doubling argument. For he is surely right that from the perspective of a first-person description of conscious experience, only one experiential object “is found to be present and possible” (ibid., 207). What is needed, then, is some way to block Brentano’s supplementation of psychological description with genetic explanation — some philosophical grounds for insisting that we take up and remain exclusively within the descriptive standpoint. Does Husserl have such grounds? Indeed he does. For to adopt a resolutely descriptive standpoint is nothing else than to abide by what Husserl calls the “principle of all principles”, that is, the principle that (as he puts it in the Ideas) everything that presents itself to us originarily (so to speak, in its bodily actuality) in “intuition” is simply to be accepted as what it presents itself as being, but also only within the limits in which it presents itself there. (Ibid., 51, emphasis removed)

This principle articulates a philosophical demand that stands at the center of Husserlian phenomenology, viz., that one adhere rigorously to the first-

64

person perspective and philosophize only from that perspective. What is the source of this demand? What justifies it? In the Ideas, Husserl argues that the principle of all principles rests on the need for an “empiricist” radicalism in our cognitive practice, a radicalism that opposes “the powers of tradition and superstition, of crude and refined prejudices of every sort” by acknowledging “the right of autonomous reason as the sole authority on questions of truth” (ibid., 41). More simply put, it is derived from the demand “to judge rationally or scientifically about things”, a demand which requires us, Husserl says, to orient ourselves to the things themselves, that is, to go from words and opinions back to the things themselves, to consult them in their self-givenness and to set aside all prejudices alien to them. (Ibid.)

The principle of all principles, in other words, is the demand to repudiate all theories that take us beyond what is available to us in intuitive givenness, since only what is intuitively given is radically free from prejudice and therefore is an appropriate datum of a genuinely rational or scientific philosophy. For Husserl, then, this most fundamental of phenomenological principles stems from the demand for an epistemologically radical philosophy, a philosophy that is grounded in the most absolute sense. Now, if we adhere strictly to the principle of all principles, then the representationalist’s scientific hypotheses come to take on a different appearance than they otherwise would. For from a purely descriptive standpoint, it becomes clear that what Brentano and other representationalists in fact do is construct a kind of “two-world” theory of perceptual presentation. Such a theory may well have great explanatory power, as is attested by the continuing success of cognitive science. But the representationalist is only able to construct his theory by granting himself hypothetical, third-person access to what, from the first-person perspective, is inaccessible. In § 52 of the Ideas Husserl argues that the representationalist theory takes physical nature to be an “unknown world of thing-like realities-inthemselves”, whereas in fact this “world” has been hypothesized by the representationalist “for the purposes of a causal explanation of our appearances” (ibid., 114f.). In engaging in such hypothesizing, Husserl charges, the representationalist ascribes to this unknown world “a mythical absolute reality”, putting it forth as the true (as over against the experienced) reality. But this hypothetical construction, Husserl points out, is a product of thought; it is “bound to the sphere of thought-syntheses” (ibid.). And this means that from a descriptive or first-person point of view, the representa65

tionalist hypothesis, once we have demythologized it and stripped it of its absolutistic pretensions, can be analyzed as to its origins in thought — that is, analyzed phenomenologically: the kind of thinking pertaining to physics establishes itself on the foundation laid by natural experience … Following the rational motives which the full context of experience presents to it, this thinking is compelled to employ, as required by reason, certain ways of conceiving of things, certain intentional constructions; and it does so for the purpose of theoretically determining sensuously experienced things. Precisely because of this there arises the opposition between the thing as object of the direct sensuous imaginatio and the physical thing as object of the physicist’s intellectio … (Ibid., 113–14.)

Once we accept the principle of all principles, in other words, it can be seen that the representationalist claim that our experiences are in fact images, signs, or some other kind of representatives of an unknowable reality depends upon our first-person experience of things. The “real” objects are precisely the objects of experience, and the notion of a world lying beyond these objects is a mere scientific hypothesis, one with no legitimate claim to absolute validity. Whatever validity it may have can only be understood on the basis of a phenomenological analysis of scientific practice, which itself rests — as do all our practices — on concrete experience. Representationalism is thereby shown to be not a metaphysically absolute account of reality and experience, but a theory constructed out of the materials of phenomenological experience. Only in this way, I submit, can phenomenology argue successfully against the representationalist doubling of objects.11 WORKS CITED BonJour, Laurence. “Foundationalism and the External World”. In Philosophical Perspectives. Vol. 13. Edited by James Tomberlin. London: Blackwell, 1999. — “Epistemological Problems of Perception”. In The Stanford Encyclopedia of Philosophy. Edited by Edward N. Zalta. Fall 2001 Edition. URL = http://plato. stanford.edu/archives/fall2001/entries/perception-episprob/. 11. Earlier versions of this paper were presented at conferences in Cologne and New York. I thank the audiences of both presentations and my commentator in New York, Prof. Richard Cobb-Stevens. I would also like to thank the anonymous reviewer for this journal, who requested clarification on several key points, and Elisa Hurley for her many helpful comments on earlier drafts.

66

Brentano, Franz. Psychologie vom empirischen Standpunkt. Vol. 1. Hamburg: Felix Meiner Verlag, 1924 (orig. 1874). — Psychologie vom empirischen Standpunkt. Vol. 2. Hamburg: Felix Meiner Verlag, 1925. — Wahrheit und Evidenz. Hamburg: Felix Meiner Verlag, 1930. — Deskriptive Psychologie. Hamburg: Felix Meiner Verlag, 1982. Cartwright, Richard. “Negative Existentials”. Journal of Philosophy 57 (1960): 629–39. Dummett, Michael. Origins of Analytical Philosophy. Cambridge: Harvard University Press, 1993. Føllesdal, Dagfin. “Brentano and Husserl on Intentional Objects and Perception”. In Husserl, Intentionality, and Cognitive Science. Edited by H. L. Dreyfus and H. Hall. Cambridge: MIT Press, 1982. Jackson, Frank. Perception: A Representative Theory. Cambridge: Cambridge University Press, 1977. Jacquette, Dale. “The Origins of Gegenstandstheorie: Immanent and Transcendent Intentional Objects in Brentano, Twardowski, and Meinong”. Brentano Studien 3 (1990/91): 177–202. Haldane, John. “Brentano’s Problem”. Grazer philosophische Studien, 35 (1989): 1–32. Hedwig, Klaus. “Über das intentionale Korrelatenpaar”. Brentano Studien 3 (1990/91): 47–61. Höfler, Alois and Meinong, Alexius. Logik. Vienna, 1890. Husserl, Edmund. “Intentionale Gegenstände”. In Aufsätze und Rezensionen (1890–1910). Edited by Bernhard Rang. Vol. 22 of Gesammelte Werke (Husserliana). The Hague: Martinus Nijhoff, 1979. — “Rezension von K. Twardowski”. In Aufsätze und Rezensionen (1890–1910). Edited by Bernhard Rang. Vol. 22 of Gesammelte Werke (Husserliana). The Hague: Martinus Nijhoff, 1979. — Logische Untersuchungen. Edited by Ursula Panzer. Vol. 19, part 1 of Gesammelte Werke (Husserliana). The Hague: Martinus Nijhoff, 1984. — Ideen zu einer reinen Phänomenologie und einer phänomenologischen Philosophie: Erstes Buch. Edited by Karl Schuhmann. Vol. 3, part 1 of Gesammelte Werke (Husserliana). The Hague: Martinus Nijhoff, 1976. McAlister, Linda. “Chisholm and Brentano on Intentionality”. In McAlister, ed., The Philosophy of Brentano. Atlantic Heights: Humanities Press, 1977. Philipse, Herman. “The Concept of Intentionality: Husserl’s Development from the Brentano Period to the Logical Investigations”. Philosophy Research Archives 12 (1987): 293–328. Plato, Theaetetus. (Many editions.)

67

Rang, Bernhard. “Repräsentation und Selbstgegebenheit”. In E. W. Orth, ed., Phänomenologie Heute. Freiburg/München, 1975: 105–37. — “Einleitung des Herausgebers”. Editor’s introduction to Edmund Husserl, Aufsätze und Rezensionen (1890–1910). Vol. 22 of Gesammelte Werke (Husserliana). The Hague: Martinus Nijhoff, 1979. Rorty, Richard. Philosophy and the Mirror of Nature. Princeton: Princeton University Press, 1979. Russell, Bertrand. “On Denoting”. Mind, 1905. Schuhmann, Karl. “Intentionalität und intentionaler Gegenstand beim frühen Husserl”, Phänomenologische Forschungen 24/25 (1991): 46–75. Smith, Barry. Austrian Philosophy: The Legacy of Franz Brentano. Chicago: Open Court, 1994. Twardowski, Kasimir. Zur Lehre vom Inhalt und Gegenstand der Vorstellung. Munich: Philosophia Verlag, 1982 (orig. 1894). Zahavi, Dan. Intentionalität und Konstitution. Copenhagen: Museum Tusculanum Press, 1992.

68

Grazer Philosophische Studien 73 (2006), 69–82.

LOST ON THE WAY FROM FREGE TO CARNAP: HOW THE PHILOSOPHY OF SCIENCE FORGOT THE APPLICABILITY PROBLEM Torsten WILHOLT Bielefeld University Summary This paper offers an explanation of how philosophy of science in the second half of the 20th century came to be so conspicuously silent on the problem of how to explain the applicability of mathematics. It examines the idea of the early logicists that the analyticity of mathematics accounts for its applicability, and how this idea was transformed during Carnap’s efforts to establish a consistent and substantial philosophy of mathematics within the larger framework of Logical Empiricism. I argue that at the end point of this development, philosophical discussion of the applicability problem was terminated although important aspects of the logicists’ original response to the applicability problem had had to be sacrificed along the way.

1. The applicability problem It is a curious fact that the question of why mathematics is applicable in the natural sciences has received so little attention in post-war philosophy of science. Early in the 20th century, the situation was quite different. The applicability problem was at least superficially addressed by many thinkers, and it was frequently regarded as a phenomenon that stands in need of philosophical explanation. As witness, let me quote from a lecture course given by David Hilbert in 1919. (The choice of witness is, of course, not arbitrary, as we shall see before long.) We are confronted with the peculiar fact that matter seems to comply well and truly to the formalism of mathematics. There arises an unforeseen unison of being and thinking, which for the present we have to accept like a miracle. (Hilbert 1992 [1919/20], 69)

It is evident that the applicability problem did not make it into the canon of established questions of professionalized post-war philosophy of science. Temporally, its disappearance coincided with the ascent and heyday of Logical Positivism and Logical Empiricism. This observation alone would lead us nowhere but to another instance of a much used but feeble rule of thumb in folk HOPOS: Whenever anything seems to have gone wrong in philosophy of science, blame it on Logical Empiricism. (For the benefit of the uninitiated, HOPOS is the History of the Philosophy of Science, as in the homonymous international society, cf. http://cas.umkc. edu/scistud/hopos/.) However, in the case of the applicability problem, it is more than just temporal coincidence that points toward a connection of its disappearance with the rise of logical positivism: One of the most widely known tenets of logical positivism, namely the thesis that mathematics is a body of analytic truths, was often taken to neatly explain why mathematics is applicable. If this explanation was widely believed by adherents and sympathizers of Logical Empiricism, that would explain why the applicability problem received so little consideration: It was abandoned because it was considered solved. In this paper, I want to propose a way in which this idea can be tracked from its logicist origins to mature (Carnapian) Logical Empiricism. I will also argue that in the course of the transformations that the basic idea underwent, an important aspect of its original explanatory potential with regard to the applicability problem was lost. At the end-point of this development, the applicability problem was excluded from the realm of relevant and sensible theoretical discussion, while at the same time the original logicist contribution to its solution had had to be sacrificed. But to begin with the beginning, let us first have a look at one relatively early articulation of the idea that the analyticity of mathematics accounts for its applicability. It is taken from a paper of Hans Hahn titled ‘Logik, Mathematik und Naturerkennen’, published in 1933 in the Vienna Circle’s book series Einheitswissenschaft. The title of the paper is significant, because it evidently mirrors that of a paper by David Hilbert, ‘Naturerkennen und Logik’, published in the journal Naturwissenschaften in 1930, where Hilbert repeats the statement from the 1919 lecture course that was quoted above. In a section bearing the headline: ‘Mathematics and Reality’, Hahn describes what he calls “the tautological character of mathematics”:

70

That we can put forward the sentence 2 + 3 = 5 apodictically and with universal validity, that we can say with complete certainty even before we perform an observation that it will not show that 2 + 3 = 7, is founded in the fact that by 2 + 3 we mean the same as by 5 […]. […] The use is again based on the fact that we do not, e.g., see right away that by 24 u 31 we mean the same as by 744; but if we calculate the product 24 u 31, we transform it step by step, in such a way that we are aware during every single transformation that due to the agreements about the use of the occurring signs […] we still mean the same after the transformation as before the transformation, until finally we have become aware that by 24 u 31 we mean the same as by 744. (Hahn 1988 [1933], 158)

A little later he proceeds: And now one should realize how vastly far apart our view is from the old — maybe one may say platonizing — view that the world is constructed according to the laws of logic and mathematics […] and that with our thinking […] we have been given a means to grasp these eternal laws of the world. No! It is no reality that our thinking can grasp, no fact of the world that thinking can bring us lore about; it only refers to the way we talk about the world; it can only transform tautologically what has been said. (Ibid., 160)1

The tautological character of mathematical statements is here adduced as evidence that worries about the relation between mathematics and the world are misguided. The general stance expressed in these paragraphs is of course not just Hahn’s private opinion, neither is it exclusively Viennese. Passages very close to this one can, for example, be found in Ayer’s Language, Truth and Logic (1971 [1936], ch. 4). 2. Logicism and the applicability of mathematics What Hahn calls, in Wittgensteinian terminology, the tautological character of mathematics is of course the heritage of the doctrine of logicism in the philosophy of mathematics. The logicists themselves, Frege and Russell in particular, were seriously interested in the applicability of mathematics. As Frege formulated it in Grundgesetze, it is the applicability of mathematical signs that requires them to have cognitive content (“Gedankeninhalt”): 1. The passages quoted are my translations from the original. An English translation of the complete essay can be found in McGuinness (ed.) 1987, 24–45.

71

Without cognitive content, no application will be possible. Why can one not get an application of a formation of chess-pieces? Obviously because it does not express a thought. […] Why can one get applications of arithmetical equations? Only because they express thoughts. […] Now it is application alone that elevates arithmetic beyond a game to the rank of a science. So applicability necessarily belongs to it. (Frege 1962 [1893/1903], 2:100)

Frege thus makes it clear that the reason to look for cognitive content in mathematical expressions is to attain an explanation why mathematics is useful in a way that a mere game could not be. Russell gets to the point in a down-to-earth manner: [W]e want our numbers to be such as can be used for counting common objects, and this requires that our numbers should have a definite meaning, not merely that they should have certain formal properties. This definite meaning is defined by the logical theory of arithmetic. (Russell 1993 [1919], 10)

As is well known, the logicists supplied the necessary “definite meaning” or “cognitive content” by means of a definition of mathematical concepts in terms of what were taken to be logical concepts. In Frege’s original version, e.g., positive whole numbers are applicable to concepts, such that for every concept F, the number of F ’s is identical with the extension of the concept “equinumerous with F ”, where “equinumerous” is in turn definable in terms of a one-to-one correspondence between extensions of concepts. Assuming that the theory of extensions of concepts is a part of logic, arithmetic is thereby defined in logical terms. Arithmetical truths are logical truths and thus, a fortiori, analytical truths. The version of Russell and Whitehead, though couched in terms of classes and types rather than extensions of concepts, also pursued these same objectives. Did the Logical Empiricists’ absorption of logicistic ideas have anything to do with the concern for applicability? It certainly did for one of the main sponsors of logicism within the movement, namely Rudolf Carnap. In his 1930 paper ‘Die Mathematik als Zweig der Logik’, where he promotes logicistic philosophy of mathematics against its formalistic alternative, we read the following: On the other hand, however, the logicistic judge discovers a gap in the system of formalism as it exists so far. As we noticed already about Frege’s basic idea, logicism directs its attention especially to the fact that mathematics must be applicable to reality. (Carnap 1930, 309)

72

He goes on to explain that in order to account for the applicability of mathematics, the mathematical signs must be equipped with meanings, namely purely logical ones. This defense of logicism thus mirrors directly the one expressed by Frege in his vivid comparison of numbers with formations of chess pieces. Logicism is to be favored over formalism because it explains applicability. 3. The demise of reductive logicism As is well known, the logicist reduction of mathematical to logical terms did not enjoy a happy fate. Frege’s theory of extensions of concepts turned out to be inconsistent, and its successors and candidates for substitutes — theories of types and axiomatic set theories — all displayed the same deficiency: They all contained premises or axioms that could not reasonably be taken to represent purely logical truths. Carnap himself showed, in his Abriss der Logistik (1929, cf. esp. § 9d), how to get around the problematic axiom of reducibility from the Principia Mathematica, but the axiom of infinity and the axiom of choice remained stubborn obstacles to any attempt to view mathematical truths as reducible to logical ones. As we have seen, this reduction was the core of the original logicist explanation of the applicability of mathematics. So how could the view that the applicability problem was no longer a problem survive the demise of reductive logicism? To answer this question, it is advisable to look at the account of that empiricist who paid the most scrupulous attention to the intricacies of such questions, i.e., again, Carnap. Carnap arrived at the position offered in The Logical Syntax of Language that amounts to a relativization of logic, called, by himself, the principle of tolerance. There is not one authoritative logic, but many possible logics. “Everyone is at liberty to build up his own logic, i.e. his own form of language, as he wishes.” (Carnap 1967 [1934], 52) Among these possible forms of language, there are also ones that contain the controversial axioms. Carnap formulates the rules and principles of one such language, language II. The axiom of infinity is made redundant by the syntax of number expressions adopted, and the axiom of choice is explicitly included in the transformation rules of language II in form of a Grundsatz or primitive sentence (primitive sentence PSII 21, ibid., § 30, p. 92). As Alan Richardson has stressed, Carnap’s approach amounts to an abandonment of the idea of a reduction of mathematics to pure logic

73

(Richardson 1994, 76). In fact, as Carnap himself explicitly notes, there is not even a clear distinction between logic and mathematics in his view. Both are equally constitutive of a form of language. The truths of mathematics are therefore not logical truths in a Fregean sense, but they are nevertheless analytic. Much could (and has of course) been said about Carnap’s concept of analyticity; at this point we may rest content with the notion that for Carnap, mathematical truths were true in virtue of our adoption of a form of language, or, as he was later to call it, a linguistic framework (Carnap 1956 [1950]). 4. What happened to the applicability problem? The title of my paper insinuates that in the development of the logicist idea I have just sketched, something of significance for the applicability problem was lost. I will now attempt to make this claim more precise. On one level, the Logical Empiricist version of logicism as developed by Carnap manages to retain an important ingredient of the Fregean idea. In Carnap’s terms, the ingredient in question is the claim that for arithmetical sentences, no question of validity can arise (cf. Carnap 1967 [1934], § 25). They are true in virtue of the language we have adopted. So, the question: “What makes arithmetical sentences true?” receives the answer: “The fact that they are analytic truths”, quite in the manner suggested by Hahn. This is close to Frege’s own answer: “The fact that they are analytic truths (that are ultimately logical truths).” But Frege’s reductive idea had additional explanatory potential when it comes to the phenomenon of the universal applicability of arithmetic. From Frege’s reduction it follows that arithmetic is applicable wherever concepts are applicable (and have finite extension). Presumably, this is a gain for the explanation of applicability. More precisely, for Frege it is not the things falling under the concepts that arithmetic gets applied to, but the concepts themselves: The laws of number, therefore, are not really applicable to external things; they are not laws of nature. They are, however, applicable to judgements holding good of things in the external world: they are laws of the laws of nature. (Frege 1980 [1884], 99)

Frege’s original version of logicism thus also provides an answer to the question: “Why is arithmetic applicable?” The answer is:

74

(A-F)

Arithmetic is applicable because it is a body of useful truths about concepts of finite extension; and we apply concepts whenever we make judgments about the external world; and a great many of them have finite extensions.

This is an answer that goes far beyond the mere claim that arithmetical truths are analytic. Analyticity alone does explain why an application of a sentence can never do any harm; it explains why, regardless of what the world is like, you can always add that sentence to a given body of truths and never thereby produce any falsehoods. But applicability, of course, requires more than doing no harm. Mere analyticity does not explain why the statements classified as analytic are of any use in the overall system of knowledge. The best that the Carnapian would have to offer for an answer to the question “Why is arithmetic applicable?” would be: (A-C) Arithmetic is applicable because language II is a serviceable language to adopt. If this seems a weaker answer than the one given on Frege’s behalf, it is because in this case the answer seems suspiciously to need an explanation itself. To this observation it might be objected that to ask Carnap: “Why is language II a serviceable language to adopt?” would be the same as to ask Frege “Why do we always use concepts when we relate to the external world?” But the difference between these two questions is remarkable, or rather: it would have been remarkable, had the Fregean version managed to live up to its promises. The universal applicability of concepts is a matter that one may plausibly accept as a brute fact, as part of the validity of pure logic. In contrast, the usefulness of language II, a system that explicitly incorporates a good deal of sophisticated mathematical content (like the axiom of choice), is no more acceptable as a brute fact than the applicability of arithmetic itself. The Carnap who wrote, in 1950, ‘Empiricism, Semantics and Ontology’ would probably have scorned the question why language II is a serviceable language to adopt, insisting as he did that questions about the acceptance of a linguistic framework are practical rather than theoretical. Philosophers who ask, e.g.: “Are there numbers?” are not asking a question that could possibly receive any reasonable answer in so far as they intend to

75

ask an external question, i.e. one that is meant to be a question prior to the acceptance of the number framework. He comments: Unfortunately, these philosophers have so far not given a formulation of their question in terms of the common scientific language. Therefore our judgment must be that they have not succeeded in giving to the external question and to the possible answers any cognitive content. Unless and until they supply a clear cognitive interpretation, we are justified in our suspicion that their question is a pseudo-question, that is, one disguised in the form of a theoretical question while in fact it is non-theoretical; in the present case it is the practical problem whether or not to incorporate into the language the new linguistic forms which constitute the framework of numbers. (Carnap 1956 [1950], 209)

It should be noted that the question whether to adopt the number language is not discredited by Carnap. But his judgment that this question is practical rather than theoretical implies that it calls for a practical decision and not for a verdict in terms of truth (cf. Bird 2003). A theoretical account of the reasons why the number language should be adopted is therefore out of place. This means that from the mature Carnapian perspective, the applicability problem was dropped from the philosophical canon, not because it was considered to have received an answer, but because all matters of applicability of frameworks were considered practical questions. A theoretical account of the applicability of mathematics was thereby ruled out as inappropriate. 5. The choice between the reductive and the logicist content of the original idea Once we discuss the development of the idea that the analyticity of mathematics accounts for its applicability, we must of course acknowledge that this process has brought undeniable philosophical progress. After all, the original ideas of Frege and Russell were stricken with a contradiction that doomed at least their version of the logicist project. Carnap showed how a position can be achieved that avoids the pitfalls of reductive logicism while preserving a compelling account of what it is that makes mathematical statements true. One might want to leave it at that and claim that Carnap’s achievement was truly a success story. Maybe Carnap had just shown that there never 76

was anything to explain with regard to applicability, i.e.: Maybe, abstracting form the illusionary pretense to a “pure logic”, Frege’s answer (A-F) was no easier to accept as a brute fact and thus no more explanatory than the Carnapian (A-C). In this last section, I would like to suggest that there is an aspect to the applicability problem and some of the Fregean ideas relating to its solution that was not automatically made obsolete by the demise of reductive logicism. Losing sight of this aspect is really what marks the loss that this paper’s title alludes to. This can best be made plausible if we consider the case of real analysis instead of arithmetic. Frege started on a theory of real numbers that remains fragmentary, because he abandoned the project after having seen the failure of his logicist reduction of arithmetic due to Russell’s paradox. But the basic ideas are well established in the account he published in the second volume of Grundgesetze (Frege 1962 [1893/1903] esp. § 73 and §§ 158–164). As in the case of whole numbers, Frege looked out for the meaning of real numbers that makes them useful for applications. He dismissed the idea proposed by Hermann von Helmholtz (1971 [1887]), that real numbers in scientific applications do not have a meaning by themselves, but form hybrid expressions together with names of units of measurement — the socalled “named numbers” — where only the composite named numbers as a whole serve to refer to magnitudes. (In fact he remarked about Helmholtz’s paper: “Hardly anything has ever felt more unphilosophical to me than this philosophical essay, and the sense of the epistemological question has hardly ever been more misconceived than here.” Frege 1962 [1893/1903], 2:139, footnote 2. Cf. also ibid., 2:85, footnote 2.) Again, the number expressions themselves must have definite meaning. Now what does the numeral mean by itself? Obviously a ratio of magnitudes. […] If now by “number” we understand the meaning of a numeral, then real number is the same as ratio of magnitudes. […] We will not say anymore that a number or a numeral signifies now a length, now a mass, now a luminous intensity; but we will say that a length can be in the same ratio to a length as a mass to a mass or a luminous intensity to a luminous intensity; and this same ratio is the same number and can be signified by the same numeral. (Ibid., 2:85)

According to this Fregean idea, the ratio of the area of the earth’s surface to the area of the northern hemisphere is identical to the real number two. In the sentence asserting that the height of the door is 2 m, the numeral “2” 77

refers to the ratio of the height of the door to the length of the standard meter, and this ratio, again, is the number 2. It might at first sight seem a shortcoming of Frege’s approach that it explains numbers in terms of ratios and magnitudes, which might themselves be considered mathematical concepts. However, one can in fact define extensive magnitudes and their ratios without recourse to the concepts of real analysis, such that the resultant totality of ratios of extensive magnitudes becomes identifiable with the positive real numbers. Bob Hale (2000) has recently developed this Fregean idea in a rigorous manner, so allow me to just sketch the basic ideas. The core of this approach is to regard extensive magnitudes as concepts that can be introduced by abstraction from properties of concrete objects. To make this abstraction possible, it suffices to suppose that there are kinds of properties whose concatenation obeys certain principles (commutativity, associativity, strict trichotomy and the Archimedean axiom), which can all be defined and understood without any prior understanding of real analysis. Ratios can then in turn be abstracted from extensive magnitudes. Here, the idea is that the required identity condition (determining when two pairs of magnitudes exemplify the same ratio) can be formulated in terms of identities and inequalities of equimultiples of those magnitudes. In this way, real analysis can be built up from an understanding of concatenations of different magnitudes (e.g., regarding the mass of objects A and B, considered together, as a function of the mass of object A and object B) and multiple concatenations of the same magnitude.2 In contrast to this, the Helmholtzian approach does not imply that the number two in itself has anything to do with either masses, lengths and 2. This identity condition and the Archimedean axiom make use of multiple concatenations of magnitudes, and therefore seem to presuppose arithmetic. While this seems to have been Frege’s own view (see e.g. 1962 [1893/1903], 2:160 ff.), it might be interesting to note that the conditions require no more than an understanding of expressions of the form “n u a”, where n is a numeral and a an extensive magnitude. Presupposing no more than an ability to cope with numerals and their successor function V, one can define recursively: 1 u a =df a and V(n) u a =df (n u a) † a. (where “†” represents the concatenation function on the respective magnitude kind). The identity condition for ratios could be formulated as follows: The ratio of magnitude a to magnitude b = the ratio of magnitude c to magnitude d if and only if for all pairs of numerals n, m: [(m u a = n u b iff m u c = n u d) and (m u a < n u b iff m u c < n u d ) and (m u a > n u b iff m u c > n u d)]. The order relation < is definable as a < b {df x: a † x = b. The feasibility of this definition is ensured by the aforementioned condition of strict trichotomy, which demands that for every pair of magnitudes of a given extensive magnitude kind, exactly one of the three following conditions holds: Either a = b, or there is a magnitude c, such that a † c = b, or there is a magnitude c, such that b † c = a. This implies that there are neither vanishing (or “zero”) nor inverse (or “negative”) extensive magnitudes in our present sense.

78

the like, or their ratios. Numbers, or rather numerals, are only employed to represent magnitudes. They are only put into this service by means of the use we make of them in our practices of measurement. This theory of the application of real analysis was later elaborated by Norman Campbell (1957 [1920])) and also largely adopted by the Logical Empiricists, e.g. by Carl Hempel (1952), by Ernest Nagel (1961), as well as by Carnap (1966). Again, this theory of real numbers as mere representational devices is not reductive, while Frege’s approach strives to identify each real number with something that secures its applicability to extra-mathematical reality. Frege’s fragmentary theory of real numbers thus suggests the following answer to the question: “Why is real analysis applicable?” (R-F)

Real analysis is applicable because it is a body of truths about ratios of extensive magnitudes.

Going slightly beyond what Frege himself explicitly says, one might extend this explanation: (R-F+) Real analysis is applicable because it is a body of truths about ratios of extensive magnitudes, and ratios of extensive magnitudes happen to abound in the world we inhabit, where they play decisive roles in all kinds of events and regularities. Have we come full-circle, back to the “platonizing view” scolded by Hahn, that the world is constructed according to the laws of logic and mathematics? The difference is that (R-F+) incorporates the idea that the truths of real analysis can be abstracted from certain properties of concrete objects. The abstraction presupposes that these properties obey certain principles, and therefore real analysis will only provide us with immediately useful knowledge insofar as there are in fact some properties that do obey these principles and can therefore be considered extensive magnitudes. It is therefore only in a very partial and restricted sense that real analysis gives us a means to “grasp the eternal laws of the world”, as Hahn’s bogey platonist would claim. Neo-Fregeans will be keen to explain that whatever grasp of reality real analysis has, it has solely in virtue of its successful employment of abstraction principles. Alternatively, albeit in a decidedly less Fregean spirit, one might also refer to the considerable empirical import that is incorporated in such conceptions as multiple concatenations of extensive magnitudes. But this point cannot and need not be

79

debated here, since we are chiefly interested in the fate of the applicability problem. As regards the explanatory force of (R-F+) with respect to the applicability problem, it is not instrumental whether the truths about ratios of extensive magnitudes are in any sense logical truths. All that matters is that the usefulness of the concepts of extensive magnitudes and their ratios for the description of physical reality is immediately explainable, in a way that the usefulness of numbers is not. (R-F+) thus indicates why real numbers are useful for at least some physical applications. It thereby certainly suggests an explanation of the applicability of real analysis that is not included in the representational approach of Helmholtz, Campbell and latter day Logical Empiricism. Their account only details how real numbers can be used as representational instruments. The case of real analysis thus shows that the reductive element in reductive logicism makes an important contribution to one aspect of the applicability problem, namely the question why mathematics is useful for many applications. What is essential is that Frege identifies mathematical concepts with something that clarifies the connection of analysis with its realm of application. The basis of this reduction is in effect provided by the conceptual resources that are required to understand multiple concatenations of extensive magnitudes. Now is this reductive basis itself “purely logical” or even “non-mathematical” in any rigorous sense? For the applicability problem, this question is not the decisive one. What matters is that (RF+) provides a conceptual connection between real numbers and masses, lengths, energies and other extensive magnitudes, and therefore provides a starting-point for an explanation of its usefulness for the description and explanation of natural phenomena. Conceiving of the number three as a certain ratio of extensive magnitudes is helpful in understanding why this mathematical concept figures at all in the description of, say, the design of a bridge — and this helpfulness is quite independent of the question whether or not the concepts of extensive magnitudes and their ratios can in turn be considered “logical”.3 When Carnap and the Logical Empiricists chose to preserve the logicist tenets of reductive logicism, sacrificing its reductive elements, they made a decision that brought with it gains and losses. While it secured the logicist’s 3. In my judgment, Frege’s conception of real numbers can be incorporated within a systematic explanation of the applicability of real analysis that is independent of logicism or neologicism. Cf. Wilholt 2004, esp. ch. 5.

80

answers to questions about the general character of mathematical truths, it forfeited the potential to account for the usefulness of mathematics in the sciences. This explanatory problem, which had been a major concern for the early logicists, was relegated to the realm of practical questions and thus excluded from the sphere of sensible theoretical discussion. It is most interesting that this decision was not reversed when shortly afterwards Logical Empiricism began to fall on hard times. After all, it was primarily the reliance on the concept of analyticity that was getting the received view into trouble. Nevertheless, the problem of applicability did not re-emerge in post-positivist philosophy of science until fairly recently. (Cf. Hellman 1989, Steiner 1998, Pincock 2004, Wilholt 2004.) Why it took such a long time to rediscover the problem is a different question. However, the preceding reflections explain how it first came to fall into philosophical oblivion. Its relegation from theoretical discussion was the price for the development of a sustained version of Carnapian Logical Empiricism.

REFERENCES Ayer, Alfred Jules (1971) [1936] Language, Truth and Logic, London: Penguin. Bird, Graham (2003) ‘Carnap’s Internal and External Questions’, in: Language, Truth and Logic, ed. by Thomas Bonk, Dordrecht: Kluwer, 97–131. Campbell, Norman R. (1957) [1920] Foundations of Science, [unaltered reissue of Physics: The Elements], New York: Dover 1957. Carnap, Rudolf (1929) Abriss der Logistik mit besonderer Berücksichtigung der Relationstheorie und ihrer Anwendungen, Wien: Springer. — (1930) ‘Die Mathematik als Zweig der Logik’, Blätter für Deutsche Philosophie 4, 298–310. — (1967) [1934] The Logical Syntax of Language, transl. Amethe Smeaton, London: Routledge & Kegan Paul. — (1956) [1950] ‘Empiricism, Semantics and Ontology’, in: Meaning and Necessity, enl. ed., Chicago & London: University of Chicago Press, 205–221. — (1966) Philosophical Foundations of Physics, ed. by Martin Gardner, New York: Basic Books. Frege, Gottlob (1980) [1884] The Foundations of Arithmetic, transl. J. L. Austin, Oxford: Blackwell. — (1962) [1893/1903] Grundgesetze der Arithmetik, 2 vols., reprint in one vol., Hildesheim: Olms.

81

Hellman, Geoffrey (1989) Mathematics Without Numbers, Oxford: Oxford University Press. Hale, Bob (2000) ‘Reals by Abstraction’, Philosophia Mathematica (3) 8, 100–123. Hahn, Hans (1988) [1933] ‘Logik, Mathematik und Naturerkennen’, in: Hans Hahn: Empirismus, Logik, Mathematik, ed. by Brian McGuinness, Frankfurt a.M.: Suhrkamp, 141–172. Helmholtz, Hermann von (1971) [1887] ‘Zählen und Messen, erkenntnistheoretisch betrachtet’, in: Hermann von Helmholtz, Philosophische Vorträge und Aufsätze, ed. by Herbert Hörz und Siegfried Wollgast, Berlin: Akademie Verlag, 301–355. Hempel, Carl G. (1952) Fundamentals of Concept Formation in Empirical Science, Chicago: University of Chicago Press. Hilbert, David (1992) [1919/20] Natur und mathematisches Erkennen: Vorlesungen, gehalten 1919–1920 in Göttingen, ed. by D.E. Rowe, Basel: Birkhäuser. McGuinness, Brian (ed.) (1987) Unified Science, Dordrecht: Reidel. Nagel, Ernest (1961) The Structure of Science: Problems in the Logic of Scientific Explanation, New York etc.: Harcourt, Brace & World. Pincock, Christopher (2004) ‘A New Perspective on the Problem of Applying Mathematics’, Philosophia Mathematica (3) 12, 135–161. Richardson, Alan (1994) ‘The Limits of Tolerance: Carnap’s Logico-Philosophical Project in Logical Syntax of Language’, Proceedings of the Aristotelian Society, Supplementary Volume 68, 67–82. Russell, Bertrand (1993) [1919] Introduction to Mathematical Philosophy, reprint, London & New York: Routledge. Steiner, Mark (1998) The Applicability of Mathematics as a Philosophical Problem, Cambridge (Massachusetts): Harvard University Press. Wilholt, Torsten (2004) Zahl und Wirklichkeit: Eine philosophische Untersuchung über die Anwendbarkeit der Mathematik, Paderborn: Mentis.

82

Grazer Philosophische Studien 73 (2006), 83–95.

JANIK ON HERTZ AND THE EARLY WITTGENSTEIN 1 John PRESTON The University of Reading Summary Various claims have been made about the influence of Heinrich Hertz’s Principles of Mechanics on Wittgenstein’s work. I consider some such recent claims, made by Allan Janik, to the effect that Hertz exercised a very strong influence on Wittgenstein, early and late. I suggest they are ill-founded, in virtue of misinterpretations either of Hertz, or of Wittgenstein, or of both. I try to set the record straight on issues such as the three criteria Hertz suggests for evaluating scientific ‘representations’ [Darstellungen] or ‘images’ [Bilder], his conception of philosophy, the nature of Hertz’s project and its relation to philosophy, the extent to which he agrees and disagrees with Ernst Mach, and his influence on the Tractatus.

The literature on Heinrich Hertz’s influence on Ludwig Wittgenstein’s Tractatus Logico-Philosophicus goes back some way. Not all the main commentators discuss or even notice that influence, although it has been particularly emphasised by James Griffin (1964), by Allan Janik and Stephen Toulmin (1973), and by Leonard Goddard and Brenda Judge (1982). My concern here is with Allan Janik’s recent treatment of Hertz and Wittgenstein, which claims that Hertz’s influence was stronger, deeper, and more abiding than previously acknowledged.2 I take issue here mainly with Janik’s interpretation of Hertz. I try to set the record straight on issues such as the three criteria Hertz suggests for evaluating scientific ‘representations’ [Darstellungen] or ‘images’ [Bilder], his conception of philosophy, the nature of Hertz’s project and its relation to philosophy, the extent to 1. I am grateful to the Arts and Humanities Research Board, an award from whose Research Leave Scheme enabled me to complete work on this article, and to an anonymous referee for this journal, for very helpful comments on the first version. 2. Janik’s article has drawn praise from Brian McGuinness, who describes it as giving ‘a good account of the influence of Hertz’ (McGuinness 2001, p. 169, note 22).

which he agrees and disagrees with Ernst Mach, and his influence on the Tractatus.3 Wittgenstein’s scientific background Philosophers who conceive of their discipline as continuous with science sometimes have the impression that Wittgenstein was an irrationalist or obscurantist. Janik seeks to defend him against such accusations by arguing that Wittgenstein was, throughout his career, influenced by Hertz. He feels that ‘[b]ringing out the links between Hertz’s technique of presenting alternative representations of mechanics to clarify its conceptual problems and Wittgenstein’s mature method for dissolving philosophical problems will … be a way of dismissing the charges of irrationalism and obscurantism that have been levelled against him’ (Janik 1994/5, p. 22; 2001, p. 149; 2002, p. 80). This overall strategy is itself problematic. For one thing, although Janik is quite right to contest the portrayal of Wittgenstein as an irrationalist, the particular strategy he deploys is not one that is necessary to contest that portrayal.4 For another, Janik’s strategy puts a lot of weight on the mere fact that Hertz was a scientist, neglecting the fact that the aspects of his book which most influenced Wittgenstein were properly philosophical, not scientific. The same can be said of Janik’s references to Wittgenstein’s ‘early scientific background’ (Janik 1994/5, p. 21; 2001, p. 148; 2002, p. 80), or his ‘background in physics’ (Janik 2001, p. 169), apparently meaning by these phrases not his training as an engineer, but his acquaintance with Hertz’s work. Such references, along with talk of the ‘scientific origins’ of Wittgenstein’s philosophy (Janik 1994/5, p. 22, 2001, p. 149; 2002, p. 80), his ‘philosophical odyssey from science’ (Janik 1994/5, p. 47; 2001, p. 168; 2002, p. 92), and the ‘scientific conception of philosophy’ (Janik 2001, p. 169) which influenced him should not impress anyone except those who think that merely being influenced by an important scientist makes a person’s philosophical views scientifically respectable.

3. I deal with Hertz’s influence on Wittgenstein’s later work, and with some of Janik’s claims about that influence, in (Preston forthcoming a). 4. See, for example, Glock 2001, where it is argued that the irrationalist-sounding elements in Wittgenstein’s remarks on philosophy can and should be glossed in a way that brings them closer to a more rational conception.

84

Appropriateness as rhetorical effectiveness? Hertz, as is well-known, lays down three criteria in terms of which he thinks physicists’ representations or images should be assessed: logical permissibility [logische Zulässigkeit], correctness [Richtigkeit], and appropriateness [Zweckmässigkeit]. The most problematic aspect of Janik’s interpretation is his reading of the latter, which Hertz himself divides into ‘distinctness’ and ‘simplicity’: [T]wo permissible and correct images [Bilder] of the same external objects may yet differ in respect of appropriateness. Of two images of the same object that is the more appropriate which pictures more of the essential relations of the object, — the one which we may call the more distinct. Of two images of equal distinctness the more appropriate is the one which contains, in addition to the essential characteristics, the smaller number of superfluous or empty relations, — the simpler of the two. (p. 2)5

According to Janik, Hertz agrees with Mach that the aim of physical theory is the simplest representation of observed phenomena, but goes on to ask the further question: simple for whom? Since there are several importantly different answers to this question, Hertz’s ‘simplicity’ and ‘appropriateness’ are audience-relative in different ways, and Janik claims this relativity is generated because different audiences are attuned to different rhetorics. Hertz, he says, regards scientific rhetoric as ‘useful and necessary in the development of physical theory’ (Janik 1994/5, p. 26; 2001, p. 152; 2002, p. 82), and ‘proceeds from the view that even within science it is necessary to construct different representations of the same data depending on whom you want to talk to’ (ibid.): a representation that is suitable for theorists is hardly suitable, say, for engineers or for chemists working with the same subject, let alone introductory students. Thus Hertz differs from Mach at the very outset by emphasizing how it is that the normal development of science requires a plurality of representations. (Janik 1994/5, p. 27; 2001, p. 152; 2002, p. 82)

Janik mentions in passing several kinds of evidence for the sort of audience-relativity he is concerned with. A full list would include at least the following. First, Hertz recognises (in his Preface) that although his book 5. Page references not attributed to other works are to (Hertz 1894), in its English translation of 1956. Because the pages of Hertz’s Preface are unnumbered in the English edition, references to the Preface do not have page numbers here.

85

‘may serve as a systematic text-book of this science [mechanics]’, it is not one for beginners. He develops this idea in his Introduction thus: Our representation of mechanics bears towards the customary one somewhat the same relation that the systematic grammar of a language bears to a grammar devised for the purpose of enabling learners to become acquainted as quickly as possible with what they will require in daily life. The requirements of the two are very different, and they must differ widely in their arrangement if each is to be properly adapted to its purpose. (p. 40)

Second, he clearly states that appropriateness, unlike logical permissibility and correctness, is contestable: ‘we cannot decide without ambiguity whether an image is appropriate or not; as to this differences of opinion may arise’ (p. 3). Third, he comments that the first image or representation of mechanics he considers, the Newtonian, ‘closely follows the course of historical development and the sequence of discoveries’ (p. 4). And, fourth, he distinguishes between images in their mature form, within ‘a perfect science’ (p. 9), and images as they are developed under the pressure of circumstances: Mature knowledge regards logical clearness as of prime importance: only logically clear images does it test as to correctness; only correct images does it compare as to appropriateness. By pressure of circumstances the process is often reversed. Images are found to be suitable for a certain purpose; are next tested as to their correctness; and only in the last place purged of implied contradictions. (p. 10)

Whether these features mean that either Hertz’s ‘appropriateness’ or his ‘simplicity’ are anything like ‘rhetorical appositeness’ (Janik 1994/5, p. 27; 2001, p. 153; 2002, p. 82), ‘rhetorical adequacy’ (Janik 1994/5, p. 28; 2001, p. 153; 2002, p. 83), or ‘rhetorical effectiveness’ (Janik 1994/5, p. 28; 2002, p. 154; 2002, p. 83) depends on what we mean by ‘rhetoric’. Janik tells us that by ‘rhetorically apposite’ he means ‘constructed with a view to the communication situation in which the scientist finds himself ’ (Janik 1994/5, p. 27; 2001, p. 153; 2002, p. 82), i.e. what I have been calling ‘audience-relativity’. But these features do not induce any kind of audience-relativity of ‘images’ or ‘representations’, since the images and representations Hertz considers are not presentations of the same data for different audiences. First, even though Hertz correctly considers his own treatment of the principles of mechanics to be beyond beginners, this does not mean that

86

any of the three images or representations he is concerned with can be identified as a representation for beginners, for engineers, or for theorists alone. Some of them may lend themselves more than others to (e.g.) introductory treatments, but they would each be capable of development in different versions, for different audiences. Audience-relativity applies to these ‘versions’, but not to the images or representations of which they are versions. One can see this quite clearly if one considers that at the very end of his Introduction Hertz envisages testing his own image against the Newtonian one, establishing a kind of crucial experiment to assess which is the more correct. Crucial experiments, however, are not a matter of comparing one rhetorical take on things with another. Second, there is no suggestion in Hertz that ‘appropriateness’ is contestable because different representation-users appreciate or are susceptible to different kinds of rhetoric, or because they are in different communicative situations. Indeed, Hertz’s definition of appropriateness (given above) makes it clear that it is precisely the kind of feature that Janik would contrast with rhetorical features, viz., an architectural feature (Janik 1994/5, p. 28; 2001, p. 153; 2002, p. 83). Third, one need not take Hertz’s comment on the fact that the Newtonian representation follows the course of history as a complaint about it (and even if one does, one does not have to think this is a necessary feature of any version of that representation). And fourthly, Hertz’s contrast between mature knowledge and images as they are developed under the pressure of circumstances is a comment on the difference between the order in which an image is assessed in a completed science, and the order in which it happens to have been developed. It does not mean that different ‘images’ or ‘representations’, in Hertz’s sense, are generated. What is worse, Janik simply ignores most of what Hertz says about appropriateness, distinctness and simplicity. If ‘appropriateness’ was only rhetorical appositeness, why would Hertz have taken the trouble he takes over the objection to certain Newtonian forces? None of his complaints about such forces is that they are rhetorically inapposite or ineffective. His most famous is that they are idling side-wheels [leehrgehende Nebenräder];6 they perform no function. Why would he bother constructing from scratch a new representation in which no such idle wheels occur, rather than merely tinkering around with Newtonian rhetoric? Finally, but decisively, 6. Hertz, pp. 11–12. Unfortunately the English translation is poor at this point, rendering Hertz’s phrase as ‘sleeping partners’.

87

consider the following passage, which immediately precedes the grammar analogy Janik quotes: [I]t will be well to point out again that we have only spoken of appropriateness in a special sense — in the sense of a mind which endeavours to embrace objectively the whole of our physical knowledge without considering the accidental position of man in nature, and to set forth this knowledge in a simple manner. The appropriateness of which we have spoken has no reference to practical applications or the needs of mankind. In respect of these latter it is scarcely possible that the usual representation of mechanics, which has been devised expressly for them, can ever be replaced by a more appropriate system. (p. 40)

Not only does this show that Hertz thinks Newtonian mechanics unsurpassable in its practical applicability, it utterly disavows the sort of rhetoric-induced relativity that Janik supposes Hertz to be promoting. It also shows that Janik is just wrong to say that Hertz’s ‘third criterion for the acceptability of a model is its usefulness in a given situation’ (Janik 1994/5, p. 27; 2001, p. 153; 2002, p. 82). When it comes to ‘appropriateness’ as regards Wittgenstein, Janik claims that philosophical clarity in the Tractatus is something that is neither an empirical nor a formal matter, but first and foremost a matter of obtaining the right perspective on the relation between the empirical and the formal, i.e., something that neither empirical nor logical propositions can say, but is in fact a matter of the application of propositions (5.557). In short, Wittgenstein understood clarity as something essentially attached to showing the limits of language appropriately. (Janik 1994/5, p. 37; 2001, p. 161; 2002, p. 88)

I have no quarrel with anything but the last sentence here. From what has already been said, though, one can gather that any invocation of Hertz’s ‘appropriateness’ in this context is so gestural as to be spurious. Neither ‘zweckmässig’ nor ‘Zweckmässigkeit’ occur in the Tractatus, not even when Wittgenstein talks of ‘unnecessary elements in a symbolism’ (5.47321) — which is certainly a related feature. He does want to draw the limits of language, but there’s no suggestion that there are multiple ways of doing so, some ‘appropriate’ and others not. What Hertz means by that term is something specific to the comparison of scientific representations and something which, by his own account, in ‘mature knowledge’ can only be evaluated after such representations have been deemed both ‘logically

88

permissible’ and ‘correct’ (i.e. empirically adequate). Because ‘appropriateness’ is a criterion which applies to representations (images, pictures), were Wittgenstein to apply it to ways of showing the limits of language (which cannot themselves be pictures, since they are not propositions, let alone empirical propositions, but examples of philosophical activity (4.112)), that would evince either deliberate misreading or simple incomprehension of Hertz’s views on his part. The immanence of philosophy? According to Janik, the crucial point about the concept of philosophy that Hertz developed and bequeathed to Wittgenstein is ‘the insistence on the immanent character of the philosophical enterprise: if philosophical problems arise in physics, then they must be handled in physics itself rather than in some theory about physics. Physics must take care of itself, as it were’ (Janik 1994/5, p. 23; 2001, p. 149; 2002, p. 80). To say that Hertz conceives of philosophical problems as immanent is strange. If it means that Hertz thinks a philosophical problem in physics can be solved not by doing philosophy, or not only by doing philosophy, but (also) by doing physics, that seems right. Hertz proposes to solve a conceptual problem about force by putting forward his own, very different, representation of mechanics, in which force no longer appears as a basic concept. But then this is a poor candidate for having influenced Wittgenstein, since for him philosophy and science are sharply different kinds of activities. The more significant problem is, however, that if Janik’s attribution to Hertz of an immanent method were correct one could not possibly go on to say anything like this: Whether Hertz succeeds or fails in his efforts to axiomatize classical mechanics is a question that need not concern us here, for it is his strategy as a philosopher of science that is so important for Wittgenstein. (Janik 1994/5, p. 31; 2001, p. 156; 2002, p. 84)7

On the contrary, if Hertz’s work is ‘immanent’ in the way Janik suggests, it really does matter whether his own system of mechanics is a success or 7. In fact, Janik’s description here of Hertz as trying to ‘axiomatize classical mechanics’ is misguided, since what Janik himself (1994/5, p. 29; 2001, p. 155; 2002, p. 84) calls ‘classical mechanics’ is the Newtonian representation, and Hertz is not even trying to axiomatise that.

89

not! His dissolution of the conceptual problem with force only succeeds if his own mechanics is an improvement on Newton’s. Otherwise he has not shown that there is any system of mechanics which is both acceptable (by the three criteria he himself has laid down) and avoids the problem which can beset the concept of force. This suggests a way in which Wittgenstein’s conception of philosophy differs from Hertz’s. For Wittgenstein, philosophy is always a second-order activity, a reflection on our forms of representation. But Hertz’s philosophical project depends on the construction and perfection of a new scientific representation; he makes the success of his philosophical project depend upon his physics.8 What does Wittgenstein mean when he says that logic must take care of itself? In the Notebooks he explicitly tells us that by saying this he means, at least in part, that ‘[i]t must in a certain sense be impossible for us to go wrong in logic’ (Wittgenstein 1979, p. 2. Cf. Tractatus 5.473-5-4731). Every possible sentence [Satz] is well-formed, every possible sign is capable of signifying, and every possible thought is logical: the notions of sentence and well-formedness are, as it were, two sides of the same coin, and the same is true of the notions of sign and signification, and those of thought and logic. The dictum’s context, especially in the Notebooks, makes clear another aspect of its meaning, which is that logic does not require ‘help’ from ontology or metaphysics, it does not need any ‘theory of things, properties, etc.’. These features don’t have much to do with logic being ‘immanent’ in Janik’s sense. Wittgenstein may have ‘got’ the idea that logic must take care of itself from Hertz, but not because Hertz had that idea. Hertz’s agreement with Mach? Janik is quite right to stress that Hertz should not be seen as a positivist, in the tradition of Mach. He usefully brings out several contrasts between the two. Ironically, though, Janik’s Hertz is still a little too close to Mach, in two respects. First, it is only half-true to say that Hertz is ‘in full agreement with Mach concerning the conundrums that the Newtonian notion of force brings with it’ (Janik 1994/5, p. 30; 2001, p. 155; 2002, p. 84). If this just means that Hertz and Mach agree that there are problems with that notion 8. In fact, Jesper Lützen, in his exemplary study of Hertz’s mechanics, shows pretty conclusively that Hertz’s ‘image theory’ was an afterthought. See Lützen 2005, and my review of it (Preston forthcoming b).

90

of force, it is unobjectionable. But if it means more than this, such as that they agree in their criticisms of that notion, it is false. Of Hertz’s objection to Newtonian mechanics on the score of logical permissibility, Mach says witheringly (but not without justification): ‘Hertz’s criticism that the Galileo-Newtonian system of mechanics, particularly the notion of force, lacks clearness appears to us justified only in the case of logically defective expositions, such as Hertz doubtless had in mind from his student days (Mach 1897, p. 319). Hertz’s other objection, as we have seen, is not that most of the forces Newtonian physics would postulate are unobservable, but that they have no function. Mach does not seem fully to have understood this, but he does indicate his disagreement by observing that ‘To characterize forces as being frequently “empty-running wheels”, as being frequently not demonstrable to the senses, can scarcely be permissible’ (ibid.). Second, and more importantly, it is not so obvious that Hertz thinks of himself as ‘dealing with’ or ‘handling’ ‘metaphysical problems in science’ (Janik 1994/5, p. 22; 2001, p. 149; 2002, p. 80), trying to ‘eliminate metaphysical problems from science’ (Janik 1994/5, p. 26; 2001, p. 151; 2002, p. 81), or that he has a ‘hermeneutic program for the elimination of metaphysics within mechanics’ (Janik 1994/5, p. 44; 2001, p. 166; 2002, p. 91). This is still too close to the positivist interpretation. Hertz talks of metaphysics only once, when discussing the logical permissibility and appropriateness of the energeticist system. It can be objected to this system, he points out, that Hamilton’s principle (or whatever other integral principle is selected) attributes intentions to inanimate nature. The usual answer to this objection, which physics nowadays keeps ready, Hertz says, is that these considerations are based upon metaphysical assumptions; that physics has renounced these, and no longer recognises it as a duty to meet the demands of metaphysics. It no longer attaches weight to the reasons which used to be urged from the metaphysical side in favour of principles which indicate design in nature, and thus it cannot lend ear to objections of a metaphysical character against these same principles. (p. 23)

This ‘usual answer’, as Mach makes amply clear in his comments on Hertz’s book, is Mach’s answer.9 Hertz, though, not only fails to endorse it, he objects to it: 9. Mach says: ‘as to [Hertz’s] criticism against the employment of minimum principles, that it involves the assumption of purpose and presupposes tendencies directed toward the

91

If we had to decide upon such a matter we should not think it unfair to place ourselves rather on the side of the attack rather than of the defence. A doubt which makes an impression on our mind cannot be removed by calling it metaphysical; every thoughtful mind as such has needs which scientific men are accustomed to denote as metaphysical. (p. 23)

Further, Hertz explicitly says one cannot eliminate unobservables from mechanics: If we try to understand the motions of bodies around us, and to refer them to simple and clear rules, paying attention only to what can be directly observed, our attempt will in general fail. We soon become aware that the totality of things visible and tangible do not form a universe conformable to law, in which the same results always follows from the same conditions. We become convinced that the manifold of the actual universe must be greater than the manifold of the universe which is directly revealed to us by our senses. If we wish to obtain an image of the universe which shall be well-rounded, complete, and conformable to law, we have to presuppose, behind the things which we see, other, invisible things — to imagine confederates concealed beyond the limits of our senses. (p. 25)

Mach could not have said these things, and the role of the concept of unobservable entities in Hertz’s work is more subtle than any assimilation of his views to those of Mach would suggest. The priority of conceptual clarification? Janik recognises, albeit grudgingly, that much of Hertz’s work in The Principles of Mechanics is the construction of an axiomatised system of mechanics, and that his ‘emphasis upon purging our models of inconsistencies has a lot to do with logical analysis (i.e., the mathematical component in modelling)’ (Janik 1994/5, p. 28; 2001, p. 154; 2002, p. 83). Nevertheless he claims that Hertz’s main emphasis is not on ensuring that our representations are logically permissible, but on conceptual clarification: [I]t is all too easy to be misled by [Hertz’s] very real concern with logical clarification into thinking that it was the main element in his program; future, the present work shows in another passage quite distinctly that the simple import of minimum principles is contained in an entirely different property from that of purpose’ (Mach 1897, p. 319).

92

whereas he, in fact, wants to place the main stress on our capacity to achieve conceptual clarification in physical theory. (Janik 1994/5, pp. 28–9; 2001, p. 154; 2002, p. 83)

This is something we Wittgensteinians, in particular, might like to believe about Hertz, but it is not true. When attempting to assess the merits of his own representation of mechanics, Hertz explicitly compares the weight of the three criteria with which he has proposed to assess representations. What he says is that the first of these, logical permissibility, matters far more to him than either or both of the other features: This merit of the representation [viz., its logical permissibility] I consider to be of the greatest importance, indeed of unique importance. Whether the image is more appropriate than another; whether it is capable of including all future experience; even whether it only embraces all present experience, all this I regard almost as nothing compared with the question whether it is in itself conclusive, pure and free from contradiction. (p. 33, emphasis added)

Logical permissibility, then, is Hertz’s principal concern. But ought it to be contrasted, as Janik contrasts it, with clarity? What Hertz says makes it certain that it ought not. His mentions of clarity almost always occur within his discussions of logical permissibility. In his first discussion of the logical permissibility of the Newtonian representation, for example, he remarks on ‘how easy it is to attach to the fundamental laws considerations which are quite in accordance with the usual modes of expression in mechanics, and which yet are an undoubted hindrance to clear thinking’ (p. 5, emphasis added). His objection to the Newtonian representation on the score of logical permissibility is to its ‘logical obscurity’ (p. 6, emphasis added). And he thinks the dignity and importance of mechanics demand ‘not simply that we should readily take for granted its logical clearness, but that we should endeavour to show it by a representation so perfect that there should no longer be any possibility of doubting it’ (p. 9, emphasis added). It’s pretty clear, too, that Hertz thinks of logical clarification as at least partly a matter of axiomatic presentation. ‘We are convinced’, he says at the very end of his discussion of the Newtonian representation’s logical permissibility, that the existing defects are only defects in form; and that all indistinctness and uncertainty can be avoided by suitable arrangement of definitions and notations, and by due care in the mode of expression. (p. 9)

93

Finally, referring back to his discussion of logical permissibility at a later point, he remarks (as we saw earlier) that ‘Mature knowledge regards logical clearness as of prime importance: only logically clear images does it test as to correctness; only correct images does it compare as to appropriateness’ (p. 10). For Hertz, then, clarity is worth striving for not in addition to (or instead of ) logical permissibility, but because the two are associated (if not identified). Clarity is not, however, associated with appropriateness (and thus not with simplicity). Conclusion: The aim of philosophy Some of the main parallels Janik draws between Hertz and Wittgenstein, then, are unconvincing, and do not sustain strong claims about the former’s influence on the latter. Nonetheless, nothing I have said casts doubt on the general idea that Hertz influenced Wittgenstein or, in particular, on the very sensible suggestion, to which Janik is committed, that Wittgenstein’s conception of the aim of philosophy was influenced by Hertz’s. Janik (1994/5, p. 33; 2001, p. 158; 2002, p. 86) rightly points out that in the very passage Wittgenstein considered as motto for the Philosophical Investigations Hertz speaks of the mind of the physicist ceasing to be gequält by the contradictions surrounding concepts like force and electricity (p. 8). Wittgenstein’s deeply-felt spiritual attitude to philosophy means he would surely have been attracted by the idea that philosophical questions might cease not merely (as the standard English translation of Hertz has it) to vex us, but, more poignantly, to torment us. The aim, for both Hertz and Wittgenstein, is complete conceptual clarity, the complete disappearance of philosophical torments.

REFERENCES Glock, H-J. (2001) ‘Wittgenstein and Reason’, in J. C.Klagge (ed.), Wittgenstein: Biography and Philosophy. Cambridge: Cambridge University Press, pp. 195– 220. Goddard, L. & Judge, B. (1982), ‘The Metaphysics of Wittgenstein’s Tractatus’. Australasian Journal of Philosophy Monograph, no.1, pp. 1–72.

94

Griffin, J. (1964) Wittgenstein’s Logical Atomism. Oxford: Oxford University Press. Hertz, H. R. (1894) Gesammelte Werke, Band III: Die Prinzipien der Mechanik in neuem Zusammenhange dargestellt, ed. P. Lenard, intr. H. L. von Helmholtz. Leipzig: J. A. Barth; translated as The Principles of Mechanics, presented in a new form, trans. D. E. Jones & J. T. Walley, intr. R. S. Cohen. New York: Dover Publications, 1956. Janik, A. S. (1994/5) ‘How did Hertz Influence Wittgenstein’s Philosophical Development?’, Grazer Philosophische Studien, 49, pp. 19–47. — (2001) ‘Saying and Showing: Hertz and Wittgenstein’, in his Wittgenstein’s Vienna Revisited. New York: Transaction Publishers, pp. 147–69. — (2002) ‘Hertz, Wittgenstein and Hermeneutics’, in B. E. Babich (ed.), Hermeneutic Philosophy of Science, Van Gogh’s Eyes, and God: Essays in Honor of Patrick A. Heelan, S.J. Dordrecht: Kluwer, pp. 79–95. Janik, A. S. & Toulmin, S.E. (1973) Wittgenstein’s Vienna. New York: Simon & Schuster. Lützen, J. (2005) Mechanistic Images in Geometric Form: Heinrich Hertz’s Principles of Mechanics. Oxford: Oxford University Press. McGuinness, B. F. (2001) Approaches to Wittgenstein: Collected Papers. London: Routledge. Mach, E. (1897) Die Mechanik in ihrer Entwickelung historisch-kritish dargestellt, 3rd edition. Leipzig: F. A.Brockhaus; translated as The Science of Mechanics, 6th American edition, trans. T. J. McCormack. LaSalle, IL: Open Court, 1960. Preston, J. M., forthcoming a ‘Hertz, Wittgenstein and Philosophical Method’, forthcoming in Philosophical Investigations. — forthcoming b Essay review of Lützen 2005, forthcoming in Studies in History and Philosophy of Science. Wittgenstein, L. (1961) Tractatus Logico-Philosophicus, trans D. F. Pears & B. F. McGuinness. London: Routledge & Kegan Paul. — (1979) Notebooks 1914–1916, 2nd edition ed. G. H. von Wright & G. E. M. Anscombe, transl. G. E. M.Anscombe. Oxford: Blackwell.

95

Grazer Philosophische Studien 73 (2006), 97–111.

REDEFINITION AND ALSTON’S ‘ILLOCUTIONARY ACTS’ Friedrich Christoph DOERGE University of Tübingen Summary The original definition of a technical term, the paper argues, should not be altered without a good reason. This notion is applied to the conception of illocutionary acts suggested by Alston, which markedly differs from the conception originally introduced by John L. Austin. Alston appears to agree with the argument; at least, he does attempt to justify his re-definition. The paper argues, however, that the reasons he gives fail.

I. During the past three and a half decades, one of the most prominent notions in the philosophy of language has been that of an ‘illocutionary act’. What is an illocutionary act? According to John L. Austin (1962), it is1 an act that essentially requires being recognised by an audience (this effect on the audience is called “uptake”), and that involves the production of conventional effects as, for example, duties and obligations.2 Promising has been taken to be a prototypical example of an illocutionary act type. It is a classical example of an illocutionary act type according to Austin because a promise requires uptake by the promisee, and it contains the conventional effect of the actor’s being obliged to keep the promise. According to Schiffer (1972, 91f., 103), however, Austin errs; in Schiffer’s account, an illocutionary act is simply an act of meaning something by what one does. Promising, it follows, is not an illocutionary act, for to make a promise is more than only to mean something; it may however, 1. The three following outlines of illocutionary act conceptions are, indeed, very rough. 2. For a detailed defense of this conception of Austin’s account see Doerge (2006, Part 1).

include one, given the actor means something in the course of the performance of her promise. According to Bach and Harnish (1979), Austin’s and Schiffer’s opinions are both not quite correct. According to them, there are two rather diverse kinds of illocutionary acts. To perform a communicative illocutionary act, on the one hand, is to express an attitude such as a belief or a desire, where to express the attitude means to “R-intend” a hearer to take one’s utterance as reason to think that one has that attitude. Such an act is successful only if the hearer identifies the attitude expressed in the way the speaker intends her to identify it (Bach and Harnish 1979, xvf., 15) — however, no conventions need to be involved here. On the other hand, there are also conventional illocutionary acts, acts of satisfying a convention (counts-as rule); here, in turn, it is not necessary that any audience recognises that the act is performed (Bach and Harnish 1979, 110). Bach and Harnish accommodate promising by making a distinction between “undertaking an obligation” and “creating” it (1979, 125). According to them, promising is an illocutionary act type of the first class just because it does not involve the creation of an obligation, but only its undertaking (ibid.). Other authors3 have various other conceptions of the illocutionary act; and, indeed, barely two of all those conceptions amount to the same thing. A sensible discourse about illocutionary acts, it would then seem, should be in terms of ‘illocutionary act according to Austin’, ‘illocutionary act according to Bach and Harnish, class 1 (or class 2)’, ‘illocutionary act according to Schiffer’, and so on. What does this mean for our initial question, “What is an illocutionary act?”? The best answer appears to be “it depends”: it depends on which of these conceptions we adopt. Strictly speaking, it appears, the question makes a false presupposition when it suggests that being an illocutionary act is a unanimous affair. In fact there are lots of diverse ways in which an act can be supposed to be illocutionary, depending on various different conceptions of what “illocutionary act” is supposed to refer to. One may object that there is (or must be) some kind of acts that are properly called illocutionary acts, in contrast to others which can only mistakenly be called illocutionary acts. I would agree if this amounted to saying that illocutionary acts are just those acts that satisfy the original 3. Cf., e.g., Searle (1969) and (1979), Andersson (1975), Hornsby (1994), Kemmerling (2001).

98

conditions Austin required for them. For “illocutionary act” is a technical term; technical terms get their meaning by stipulative definition, and once the term is defined it should be used according to this definition — all other usage is just incorrect. Unfortunately, this argument depends on a certain view of how we should treat our scholarly terminology which is, at least implicitly, disputed. Some feel that it is acceptable to define one’s technical terminology arbitrarily, and that there is no problem, for instance, in using the predicate “true” for ascribing to a proposition that it is useful to believe it — at least as long as one, as they may require, makes clear that this is how one uses the word. In the final consequence, this view does not strike me as plausible; but let us grant that it is not obviously absurd. In section 3 of this paper I will argue that, still, there is reason not to re-define technical terms in philosophical discourse. If so, then someone who does re-define a technical term for philosophical discourse can at least be required to have, and give, a reason why she does not just apply the original sense. I shall apply my argument to the conception of illocutionary acts that has recently been suggested by William P. Alston in Illocutionary Acts and Sentence Meaning (2000). Since this tremendous book provides a profound picture of sentence meaning, and since Alston is to be counted as one of the most prominent theorists of illocutionary acts,4 his conception of an illocutionary act will not fail to have a considerable appeal. As I am going to show, however, it is wholly different from Austin’s; and if my basic view of how we should treat technical terms is true, then Alston’s conception is in need of a justification for re-defining the term “illocutionary act”. Alston appears to agree, for he does provide arguments against the conception Austin introduced. I shall argue, however, that these arguments fail. II. Let me start with a short exposition of Austin’s conception of illocutionary acts. Austin’s investigation had started from the supposed contrast between two different kinds of sentences, “constatives” and “performatives”. The 4. During the past decades he published many important contributions in terms of “illocutionary acts” including, e.g., Alston (1964), (1977), (1987), (1991), and (1994).

99

characteristic of the “performative” were supposed to be that it is issued in the course of the performance of a special kind of action — we may call it “AUSTIN-act”5 —, and that it is, as Austin said, “happy” or “unhappy”, rather than true or false. After failing to find a clear formulation as to what the difference between performatives and constatives ultimately is, Austin undertakes what he calls a “fresh start” on the problem: he sets out to consider in more detail the acts we perform when we are using performatives and constatives, hoping that, as he says, “some clarification and definition here may help us out of our tangle” (Austin 1962, 91). According to Austin, there are three particular kinds of actions: (1) the “locutionary act” (the act of ‘saying’ something), the illocutionary act (which is nothing other than the ‘AUSTIN-act’, the act Austin was concerned with in connection with performatives6), and the “perlocutionary act”. The crucial exposition of Austin’s conception of illocutionary acts is stated in separating them from perlocutionary acts (Austin 1962, 116f.). The authority of this exposition is decisively reconfirmed in two places. Firstly, it is repeated in the summary of the relevant lecture (Austin 1962, 121); secondly, it is put into practice when Austin later examines wether stating is an illocutionary act type (Austin 1962, 139). According to this conception, there are (at least) two features for an act’s being an illocutionary act: (1) In order to perform the act one must secure uptake in an audience that the act is performed (Austin 1962, 22, 36, 116f., 121, 139). (2) The act is conventional in that it is an act “done as conforming to a convention” (Austin 1962, 14, 26, 105) and that it entails the production of certain conventional effects (Austin 1962, 14, 26, 102, 117, 121f., 139). The natural picture of such a kind of act is that there is a convention which demands the satisfaction of certain conditions for the performance of the act to take place (this performance comes about by virtue of this convention), among which conditions is the requirement that the actor secure uptake in an audience. 5. Cf. Doerge (2006, §§ 1.1.3, 1.2, 1.3). 6. This is argued for in Doerge (2006, § 2.2).

100

Typical examples, according to Austin, are promising7 and betting8. According to the picture, they are to be analysed as follows. There is a convention, requiring for the occurrence of a promise that the actor secures uptake in the audience of the fact that she promises; the convention may require further conditions as, for instance, that the person promising must not be completely insane at the time of the attempted promise. If all those conditions required by the convention are satisfied, then the promise comes about, which entails the conventional effect of the speaker’s being committed to the promised thing. Likewise, there is a convention for betting, requiring for the success of a bet that all participants involved are aware of the bet. Further conditions may be necessary; if, for example, my intended partner refuses to bet then no bet will come into existence. If all the required conditions are satisfied then the bet takes place, which entails the conventional effect of the winner’s being entitled to take the stake. For the present purpose we also have briefly to consider Austin’s notion of ‘locutionary acts’. This is meant to capture nothing more than the “act of ‘saying something’” in “the full normal sense” (Austin 1962, 94). The act of saying, according to Austin, splits into three parts, or aspects. Firstly, to say something is to “utter certain noises” (Austin 1962, 92, 95); Austin dubs this aspect the “phonetic act”. Secondly, to say something is to utter noises “of certain types, belonging to and as belonging to a certain vocabulary, conforming to and as conforming to a certain grammar” (Austin 1962, 95); this partial act, or aspect, is called the “phatic act”. Thirdly, these grammatical noises have, and are uttered with, a meaning; as Austin puts it, saying something entails “using those vocables with a moreor-less definite sense and reference” (Austin 1962, 95). This last aspect, of uttering words with their (linguistic) meaning, is called the “rhetic act”. Since the publication of Austin’s How to Do Things With Words until today, a fair number of more or less elaborate accounts in terms of illocutionary acts have been published. Some of these accounts witness the intent of the authors to account for the above two conditions, as, for example, Strawson (1964), Searle (1969), Andersson (1975), or Bach and Harnish (1979). However, other authors leave (more or less deliberately) Austin’s criteria for illocutionary acts behind them and propose other definitions 7. See, e.g., Austin (1962, 9ff., 22f., 32ff., 40ff., 50, 54, 59, 69ff., 117, 120, 136f., 151). 8. See, e.g., Austin (1962, 5ff., 13ff., 18, 32, 36f., 40, 58, 63ff., 134, 142).

101

of what an “illocutionary act” is.9 Alston belongs to this latter group. Let me set out the very basics of his conception of illocutionary acts. III. Alston suggests three ways of capturing his conception of illocutionary acts. His first definition goes as follows. (1) An act is an illocutionary act iff it is the uttering of a sentence or “sentence surrogate” with a certain “content” (Alston 2000, 2, 26, 33). The “content” is explicated as “anything that [the utterer] seeks to communicate, anything a hearer (H) must grasp in order to understand what the speaker is saying” (Alston 2000, 15). Some caveats are necessary. Firstly, “what the speaker is saying” is here not meant to refer to “what sentence is uttered” (2000, 13), it is to be understood in a sense that “does not involve direct quotation” (2000, 14); apparently, it aims at something very similar to, although not identical with (cf. 2000, 271), the content of meaning in the sense of Grice (1957, 1969), or Schiffer (1972). Secondly, although the content is something a hearer must grasp in order to understand what the speaker is ‘saying’, it is not necessary for an act to be illocutionary in Alston’s sense that any such hearer (over and above the speaker herself ) is present (2000, 44f., 94); and even if a hearer is present, it is not necessary that she in fact understands what the speaker is ‘saying’; that is, the securing of uptake is not necessary (2000, 24, 67, 135). Finally, a “sentence surrogate” can be a “non-linguistic device” (2000, 28); in principle, “any perceptible device” can be such a “sentence surrogate”, as long as it is tied to its being usable for the illocutionary act in question “by convention or rule” (2000, 29). Let us turn to Alston’s second definition. (2) An act is an illocutionary act iff it is the uttering of a sentence [or sentence surrogate] as subject to a so-called “I-rule” (“illocutionary rule”). (Alston 2000, 272) 9. E.g., Schiffer (1972), Hornsby (1994) and Kemmerling (2001).

102

An “I-rule” is a rule “laying down necessary and sufficient conditions for sentence utterance, where the social rationale of the rule is the facilitation of communication” (2000, 272). The third definition goes thus: (3) An act is an illocutionary act iff it is an act of “taking responsibility” for — briefly, the “R-ing of ” — “certain conditions in performing some lower-level act” (Alston 2000, 269), where the lower-level act is performed “with the intention of getting the adressee to realize that [the utterer] is R-ing” (ibid., 271). “U R’d that p in uttering S”, Alston stipulates, iff “in uttering S, U subjects his uttering to a rule that, in application to this case, implies that it is permissible for U to utter S only if p” (2000, 69). The next thing to note is that Alston’s conception of illocutionary acts accounts for neither of the two criteria Austin demands.10 Firstly, as we have seen, Alston views the requirement for the securing of uptake as not necessary for an act’s being an illocutionary act. Secondly, Alston does not require conventional effects for illocutionary acts. As he (2000, 84) himself emphasises, to ‘R’ certain conditions is not to produce conventional effects in the relevant sense. Alston (2000, 92ff.) in general rejects the view that illocutionary acts entail the production of conventional effects; he instead decides to view as the illocutionary act ‘component’ — in the case of acts that do include the production of conventional effects — only the ‘purporting to perform’ the act. To show only one strange consequence of this decision: the act-type of promising, which is commonly taken to be a prototypical instance of an illocutionary act type, is not in fact an illocutionary act type according to Alston; instead, the illocutionary act is here restricted to the “purporting to” perform the act of promising (2000, 96f.). IV. I should now like to establish a requirement concerning the treatment of the words used in philosophical discourse, particularly, a requirement 10. Alston (2000, 14) admits that he specifies illocutionary acts “very differently”, compared with Austin’s account.

103

concerning the treatment of technical terms.11 If a philosopher introduces a new technical term, ascribing to it a meaning by stipulation, then the term has thereby been provided with this meaning as its meaning for the philosophical discourse. Imagine that after this another philosopher defines the same term in a different way. This will, among other things, have the consequence that the term is now ambiguous in the philosophical discourse. However, I think that it is also fairly obvious that in philosophical discourse ambiguity should be avoided. Given this, and given that the redefinition of a technical term causes ambiguity, it follows that we should, all things being equal, not re-define technical terms. To be sure, I do not want to deny that there can be conditions under which we in fact should, all things considered, re-define a certain technical term. What I claim, however, is that in such a case a reason to this effect must be present; in order to justify the re-definition one must provide an overriding reason to the effect that the re-definition is recommendable, or at least acceptable, in this particular case. “Illocutionary act” obviously is a technical term; it was introduced12 by Austin with reference to the two necessary features mentioned above, ‘conventionality’ and ‘uptake’. As we have seen, Alston introduces a new definition, which does not adopt the features Austin had originally stated. Why does Alston feel free to define a new conception of illocutionary acts? His overall argument goes as follows. 11. I think that my argument can be applied to any scholarly term whatsoever, as well as natural language expressions. In both cases, however, several special problems would have to be taken into consideration; whereas for my purpose it suffices to state the argument as restricted to technical terms. 12. One might want to raise doubts on the application of my general argument about redefinition to the case of “illocutionary acts” in one of the following ways: What Austin did when he introduced the expression “illocutionary act” should not really be called a “definition” because (1) he did not intend his characterisation to be absolutely final, or (2) he did not intend his characterisation to provide both necessary and sufficient conditions, rather than only necessary ones, or (3) his characterisation remains vague in certain ways, or (4) his characterisation does not go far enough to provide a “complete analysis”, or (5) his characterisation does not satisfy one or the other requirement that has been stated for “formal definition”. — When I use the term “definition”, then I do not thereby intend to pick out something that is particularly intended to be immune to reasonable revision, or something that is more than only a partial definition, or an exact, rather than a vague definition, or something that provides a detailed, rather than only a rough analysis, or particularly a “formal definition”. I mean the word “definition” in the rather broad sense that, I think, in fact is usually meant when the word is used; and the argument about the re-definition of technical terms is indeed meant to apply to each and all kinds of “definition” in this broad sense.

104

After a rather condensed exposition of Austin’s account, he says that “[i]n the material so far cited all Austin has done along this line is to (a) provide a number of examples and (b) employ the unexplained term ‘force’” (2000, 24). He then asks the crucial question whether Austin “goes beyond that” (ibid.). He continues by examining several candidates for essential characteristics of illocutionary acts according to Austin; the result, according to Alston, is that all of those candidates are to be rejected. He concludes that “Austin has done little more, on the question of the nature of [illocutionary acts], than present us with an intuitive list of illocutionary verbs” (2000, 25). His justification for re-defining the concept of illocutionary acts, that means, consists in the denial of the existence of a definition, or at least partial definition, by Austin.13 So far the rough overall outline. When Alston argues that Austin fails to define illocutionary acts then this contradicts the assumption, which I established above, that Austin demands (1) the ‘securing of uptake’ and (2) ‘conventional consequences’ as essential features of illocutionary acts; for to demand these two features for illocutionary acts is to make at least a partial definition of the notion of illocutionary acts. Therefore, I shall presently examine Alston’s reasons for the rejection of these two features as essential for illocutionary acts. Before doing so, however, I should briefly mention that Alston considers two further candidates for conditions of an act’s being an illocutionary act mentioned by Austin. Firstly, Austin speaks about the fact that “many illocutionary acts invite by convention a response or sequel” as, for instance, that “an order invites the response of obedience and a promise that of fulfilment” (Austin 1962, 117). Secondly, Austin insinuates a connection between the doctrine of illocutionary acts and his initial doctrine of performative sentences when he insists that illocutionary acts “could be made explicit by the performative formula” (Austin 1962, 103). I agree with Alston that these two features should not be adopted as defining traits of illocutionary acts; Austin himself (1962, 139) unambiguously rejects the first one as an essential feature of illocutionary acts; and it does not appear to me that he really presents the second one as a criterion at all. Since I agree with Alston that both features are not to be valued as criteria for illocutionary-actness, they need not be considered here in any further detail.

13. Cf. Alston (2000, 18f.).

105

V. Alston does observe that “Austin discusses the audience’s understanding ‘of the meaning and the force of the locution’ [i.e., the occurrence of uptake]” as a defining trait of illocutionary acts (Alston 2000, 24). According to him, however, Austin “wavers between saying that the illocutionary act cannot be ‘happily, successfully performed’ without this ‘uptake’, and suggesting that this is a necessary condition of the mere performance of the act”. Alston objects that, “clearly”, “the latter is not the case” (ibid.). For, he argues, the act types of telling and asking do not require uptake: Whether I told you that the dean is coming to dinner or asked you to bring me a towel does not hang on whether you heard or understood me. If you didn’t, my communicative purpose has been frustrated. But it doesn’t follow that I didn’t tell or ask you. (Alston 2000, 24)

As it stands, Alston’s argument goes as follows. (1) Austin’s exposition leaves it open whether uptake is necessary for an act to be an illocutionary act, or whether it is only a condition of an illocutionary act’s being happy and successful. (2) The illocutionary act types of telling and asking do not require uptake. (3) Hence, Austin is to be interpreted as not essentially demanding uptake for illocutionary acts. This argument does not work. Let me first say that I do not feel inclined to agree with (2); I rather feel that if someone does not succeed in getting across that she is telling something or asking someone for something then she did not really tell or ask. Maybe I am wrong; but at least it follows that we are here confronted with conflicting intuitions. That is, the truth of (2) does not seem to be so obvious as Alston suggests. Secondly, premise (1) is not true. Admittedly, Alston is right when he notes that Austin sometimes uses rather weak formulations when uptake is at issue. However, in the end Austin unambiguously fixes that uptake is necessary for illocutionary acts: I cannot be said to have warned an audience unless it hears what I say and takes what I say in a certain sense. An effect must be achieved on the audience if the illocutionary act is to be carried out. […] So the performance of

106

an illocutionary act involves the securing of uptake. (italics mine, except in the last case; Austin 1962, 116f.)

Thirdly, given this, premise (2) may appear to be begging the question against the argument that Austin in fact requires uptake; for according to Austin’s exposition it holds that if telling and asking do not require uptake then they cannot be illocutionary acts. Fourthly, in order to arrive at the conclusion that Austin did not stipulate the securing of uptake as a necessary condition, as Alston in fact intends, it is not helpful to assume something about what illocutionary acts ‘really are’; the argument must be about how Austin took things to be. It is easy to see how to re-state the argument according to this latter objection. For Austin himself indeed counts telling (1962, 162) and asking (1962, 117) among the illocutionary act types (although they are by no means among his favourite examples). So a restatement can be put as follows. (4) Austin takes telling and asking to be illocutionary act types. (5) Austin takes telling and asking not to require uptake. (6) Hence, Austin is to be interpreted as not essentially demanding uptake for illocutionary acts. The main problem with this argument lies in (5). To start with, there is — to my best knowledge — no indication in Austin’s writings to the effect that he really does take telling and asking not to require uptake. Secondly, in the absence of such an indication, and given the fact that Austin after all defines illocutionary acts as requiring uptake, we have good reason to ascribe to him the opposite view. Thirdly, there is independent reason that he would have demanded uptake for these act types. As I argued above, we appear to be confronted with conflicting intuitions concerning how admissive the concepts of telling and asking are. Now, as has often been observed,14 Austin shows a rather strong tendency towards taking a formally demanding view on our everyday activities. Hence we may assume that he would, if being pressed, have chosen the more formal view on telling and asking as indeed requiring the securing of uptake. Given all this, it is not implausible that (5) is false. And if it is plausible that Austin did assume, or would have assumed, that telling and asking 14. Cf., e.g., Kemmerling (1986, 133), Bach and Harnish (1979, 55, 305 n.).

107

do require uptake, and given the fact that Austin after all did demand uptake for illocutionary acts, it is hard to see how this condition might be rejected. Notice, finally, that even if it turned out that telling and asking in fact do not (always) require uptake, this would still not mean that the demand is to be dropped after all. The much less dramatic consequence would be that telling and asking, contrary to Austin’s view, are not (or not in all cases) illocutionary acts; at least, neither telling nor asking are among Austin’s favourite examples of illocutionary acts. VI. Let me turn to Alston’s doubts concerning Austin’s second condition, that illocutionary acts are essentially ‘conventional’. Alston in fact does not deny that, for example, “it is plausible to say that what makes a certain utterance a case of promising to read your paper is something on the order of a social convention or rule” (2000, 25). “But”, he objects, “it is equally clear that this does not mark a distinction between illocutionary and rhetic acts. If I say ‘It’s going to charge’ with a certain sense and reference, this is due, at least in part, to the relevant linguistic rules and conventions” (ibid.), as well. “Thus”, he concludes, “Austin fails to show that conventionality will even differentiate illocutionary acts from other types of acts in the scheme [that Austin provides for speech acts]. Much less does he show how to give an account of their nature in these terms” (ibid.). The argument can be put as follows. (1) If the demand for feature M does not differentiate illocutionary acts from all other kinds of acts, then there is no appropriate account of the nature of illocutionary acts in terms of M. (2) The demand for conventionality does not differentiate illocutionary acts from rhetic acts. (3) Hence, there is no appropriate account of the nature of illocutionary acts in terms of conventionality. This argument fails, for (1) is not valid. If a certain feature is necessary, but not sufficient, for an act’s being an illocutionary act, then there is no reason to demand that this feature, in order to figure in an appropriate

108

definition, must differentiate illocutionary acts from all other kinds of acts; this could be demanded only from an allegedly sufficient condition. In fact, Austin suggests conventionality only as a necessary feature of illocutionary acts, but not as a sufficient one. He further demands (at least) the requirement of uptake. Now it is clear that Austin’s locutionary act (the act of saying something) does not require uptake; I can say “It’s going to charge” with a certain sense and reference even if no one recognises that I do so. So even if the reference to conventionality would not distinguish illocutionary acts from locutionary acts, and thereby from rhetic acts — that is, even if premise (2) was true — , the requirement for uptake would suffice to distinguish them. Furthermore, premise (2) does not seem to be true either. Alston supports it by arguing as follows: If we are to distinguish illocution from locution by its conventionality, we have to find a sort of conventionality that is involved in the former but not in the latter. (Alston 2000, 25)

Of course, this is true; however, in fact there is a sufficiently clear difference between linguistic conventions and those conventions that constitute acts like promising and marrying. Actually, Alston (2000, 91, Fn. 7, 94) himself repeatedly relies on the assumption that there is a clear difference between them. The difference might preliminarily be outlined by saying that linguistic conventions will supposedly have (or be reducible to15) the form “x means y”, whereas illocutionary conventions have a form like “To do M in context C is to perform act y”.16 These two kinds of conventions are clearly distinct from each other, if only because linguistic conventions assign a linguistic meaning to a type of sign, whereas illocutionary conventions do not. VII. Let me sum up the findings of this paper. A comparison, although very rough as it was, of Austin’s conception of illocutionary acts with the one Alston defines in Illocutionary Acts and Sentence Meaning has shown that 15. Cf. Searle (1969, 33ff.). 16. Cf. Bach and Harnish (1979, 109) for (roughly) such a construal of illocutionary conventions.

109

Alston rejects both of the necessary conditions Austin originally declared to be essential for illocutionary acts. I suggested an argument to the effect that there is reason not to re-define technical terms just as one likes, that is, to the effect that such a re-definition requires having some reason to the effect that the re-definition is justified, or even required. Alston appears to agree with my conclusion; for he in fact attempts to show that Austin did not provide us with a definition, or at least partial definition, of illocutionary acts. As I have argued, however, his arguments against ‘uptake’ and ‘conventionality’ — those two features to which Austin sticks in the end as necessary for illocutionary acts — fail.

REFERENCES Alston, W. P. (1964) “Linguistic Acts”, American Philosophical Quarterly, 13, 107–124. — (1977) “Sentence Meaning and Illocutionary Act Potential”, Philosophical Exchange, 2, 17–35. — (1987) “Matching Illocutionary Act Types”, in: J. J. Thomson (Ed.), On Being and Saying. Cambridge: MIT Press, 151–163. — (1991) “Searle on Illocutionary Acts”, in: E. Lepore & R. van Gulick (Eds.), John Searle and His Critics. Oxford: Blackwell, 57–80. — (1994) “Illocutionary Acts and Linguistic Meaning”, in: S. L. Tsohatzidis (Ed.), Foundations of Speech Act Theory. London: Routledge, 29–49. — (2000) Illocutionary Acts and Sentence Meaning. Ithaca: Cornell University Press). Andersson, J. S. (1975) How to Define “Performative”? Stockholm: Libertryck. Austin, J. L. (1962) How To Do Things With Words. 2nd edition, Oxford: Oxford University Press 1975. (First published 1962.) Bach, K. & R. M. Harnish (1979) Linguistic Communication and Speech Acts. Cambridge: MIT Press. Doerge, F. C. (2006) Illocutionary Acts — Austin’s Account and What Searle Made Out of It. (Tuebingen: University of Tuebingen.) Grice, H. P. (1957) “Meaning”, The Philosophical Review, 66, 377–388. — (1969) “Utterer’s Meaning and Intentions”, The Philosophical Review, 78, 147–177.

110

Hornsby, J. (1994) “Illocution and its Significance”, in: S. L. Tsohatzidis (Ed.), Foundations of Speech Act Theory. London: Routledge, 187–207. Kemmerling, A. (1986) “Utterer’s Meaning Revisited”, in: R. E. Grandy & R. Warner (Eds.), Philosophical Grounds of Rationality. Oxford: Oxford University Press, 131–155. — (2001) “Gricy Actions”, in: G. Cosenza (Ed.), Paul Grice’s Heritage, Bologna: Brepols, 73–99. Schiffer, S. R. (1972) Meaning. Oxford: Clarendon. Searle, J. R. (1969) Speech Acts. Cambridge: Cambridge University Press. — (1979) Expression and Meaning. Cambridge: Cambridge University Press. Strawson, P. F. (1964) “Intention and Convention in Speech Acts”, The Philosophical Review, 73, 439–460.

111

Grazer Philosophische Studien 73 (2006), 113–131.

NOT TOO PROUD TO BEG (THE QUESTION): WHY INFERENTIALISM CANNOT ACCOUNT FOR THE A PRIORI Michael VEBER East Carolina University Summary The inferentialist account of the a priori says that basic logical beliefs can be justified by way of rule circular inference. I argue that this account of the a priori fails to skirt the charge of begging the question, that the reasons offered in support of it are weak and that it makes justifying logical beliefs too easy. I also argue that recent modifications to inferentialism spell doom for it as a general theory of a priori justification.

We believe that Modus Ponens (MPP) is valid.1 If this belief is justified it is justified either inferentially or non-inferentially. The latter road is taken by a more or less traditional view of the a priori. Fundamental logical, mathematical, conceptual and even philosophical propositions, are thought to be justified on the basis of some non-empirical source of justification that goes by the name of “rational insight” (BonJour, 1998) or “intuition” (Bealer, 2000). When it comes to certain sorts of proposition (MPP is valid being as good a candidate as any), cognizers can, upon understanding them, intellectually “see” that they are true. Many philosophers have been deeply disappointed by the traditional view of the a priori. The reason usually offered is that there is no good account of how an act of rational insight or intuition (whatever that is) could serve to justify belief. It’s a big mystery. This sort of worry leads some to the conclusion that there is no such thing as a priori knowledge or justification at all.2 But interest in the a priori has grown in recent years 1. The ‘we’ here excludes some philosophers such as Lycan and McGee who have argued that MPP is invalid. But even they take it that some inferences are valid. The comments here apply mutatis mutandis to whatever inference form(s) they happen to accept. 2. For example, Devitt, (2005).

and now there are many alternatives to the traditional view. One interesting alternative is inferentialism. According to this sort of view, there is no need to appeal to any faculty of rational insight or intuition to explain how our a priori beliefs are justified because a priori justification is inferential. Inferentialism attempts to account for all of our knowledge of things like logic and mathematics without postulating ‘inutition’ or ‘rational insight’ as a special non-empirical source of justification and to maintain that our knowledge in these areas is nonetheless a priori. In this paper, I will argue that this project does not succeed. The most pressing problem for inferentialism is that the justification it attempts to provide for fundamental logical beliefs involves a fallacy of begging the question. I will begin by explaining what it is to beg the question and why it is a problem for inferentialism. The inferentialist account of the a priori can be independently motivated by criticisms of an alternative epistemological view. In the second section, I will explain what this view is and how certain criticisms of it can be made to support the inferentialist account of the a priori. Following that, I will explain why the sort of justification inferentialism offers for basic principles of logic needs to be restricted and how a plausible restriction can be grounded. In section four, I will argue that the inferentialist’s attempt to avoid the charge of begging the question fails. In section five, I will argue that none of the reasons designed to motivate the sort of epistemological externalism required by the inferentialist account of the a priori are compelling and that, even with the above mentioned restriction, inferentialism still makes justifying rules of inference too easy. The penultimate section discusses some modifications that have been made to inferentialism’s underlying conceptual role semantics. I argue that, while these modifications would help avoid some of the problems with inferentialism’s theory of concepts, they only serve to make its theory of the a priori much worse. I conclude by pointing out some very surprising remarks from inferentialism’s most prominent proponent. 1. Begging the question When we ask about the epistemic status of paradigmatically a priori beliefs such as the belief that MPP is valid, it is easy to see how the inferentialist runs into problems. If my belief that MPP is valid is justified on the basis of an inference from other beliefs, this inference must be truth conducive.

114

But the inference is itself either an instance of MPP or not. If it is not an instance of MPP, then the problem is merely postponed. We can simply ask about the truth-conduciveness of this other inference. If my belief that MPP is valid is arrived at using MPP, then we can see that the inference is (at least in some sense) circular. The inference attempts to justify the very rule it employs. It is rule circular. Showing that rule circularity is at least sometimes epistemically unproblematic is one of the central difficulties facing inferentialism. Opponents of inferentialism will charge that a rule circular argument fails because it commits the fallacy of begging the question. Before we can assess whether this charge is legitimate, we must ask exactly what it amounts to. One might try to define a question begging argument as one in which the conclusion appears as a premise. This definition, however, cannot even capture the textbook cases. Consider the following: 1. Everything in the Holy Text is true. 2. The Holy Text says that God exists. 3. Therefore, God exists. In the typical case, although perhaps not in every case, this argument begs the question not because the conclusion appears as a premise (for that is not even true here) but because belief in a premise epistemically depends on belief in the conclusion.3 Most who believe (1) base this belief ultimately on a belief that God exists. 4 An argument is an ordered pair of a set of propositions (the premises) and a proposition (the conclusion).5 Arguments have their logical and semantic properties intrinsically but their epistemic properties are derivative of the epistemic states of the subjects who use them in reasoning or demonstration. To say that an argument fails to justify its conclusion is to say that it is being used in a way such that a certain subject cannot be warranted in believing the conclusion based on his belief in the premises. In general, an argument serves to justify a subject in believing its conclusion only if the subject is already justified in believing the premises of 3. Providing an account of epistemic dependence is outside the scope of this paper. For a discussion of this issue see Korcz (1997). 4. A subject might accept (1) because he has taken himself to have established it independently, perhaps by induction. This would not result in a question begging argument (but of course it might be bad for some other reason). 5. This definition of ‘argument’ can be found in Sinnot-Armstong (1999).

115

that argument. If a subject bases his belief in a premise on a prior belief in the conclusion then the subject is justified in believing that premise only if he is already justified in believing the conclusion. Therefore, a question begging argument does no real work towards providing the subject with justified belief in the conclusion. If he had that, he had it already. The argument is, at best, a rhetorical exercise. The use of an argument begs the question (and thus we can loosely say that the argument begs the question) when belief in a premise is epistemically based on belief in the conclusion.6 The success of inferentialism as an account of the epistemology of basic logical beliefs turns on whether there are rule circular arguments that are not question begging in this sense. The distinction between rule circular arguments and question begging arguments is drawn by appeal to two general philosophical ideas: (a) a limited sort of epistemological externalism and (b) conceptual role semantics. These two ideas and how they fit into the inferentialist account of the a priori will be discussed in the next two sections. 2. The internalist challenge One theory about the nature of inferential justification, “Simple Inferential Internalism” (SII), would pose serious problems for any attempt to rule circularly justify basic principles of inference. SII is the view that for S to be justified in believing that p on the basis of an inference from some other set of propositions, S must be able to know “by reflection alone” that these other propositions provide good reasons for believing that p.7 In this case, if S is to know that MPP is valid on the basis of a modus ponens inference from other propositions, S must know (or at least be in a position to know) in advance that MPP is valid. Therefore, if SII is true, then a rule circular argument can contribute nothing to the justification we have for believing that the rule in question is valid. The legitimacy of the argument would require that we already have warrant for belief in the conclusion. This is not the same thing as saying that SII entails that all rule circular arguments are question begging. But it is to say that if SII is true, then all rule circular arguments fail to provide anything new in the 6. A consequence of this sort of account is that the same argument might be question begging in one usage context but not in another. But as the case described in footnote 4 shows, this is exactly as things should be. 7. Boghossian (2004), p. 229.

116

way of warrant for belief in their conclusions. This would mean that rule circular arguments are epistemically uninteresting for the same reason that question begging arguments are. Since the inferentialist account of the a priori requires that rule circular arguments provide substantive epistemic justification, proponents of inferentialism must find reason to reject SII. There are two major criticisms of SII. The first is that ordinary folk, on this view, can have no inferentially justified beliefs. SII entails skepticism. Granny wonders what’s on TV tonight. So she opens the TV guide and reads that Matlock starts at 8. She then forms the belief that Matlock will begin at 8. We can model Granny’s reasoning this way: 4. If the TV guide says Matlock starts at 8, then Matlock starts at 8. 5. The TV guide says Matlock starts at 8. 6. Therefore, Matlock starts at 8. We assume that Granny has good inductive support for 4 and good observational evidence for 5. This, we hope, allows Granny to be justified in believing 6. Is Granny able to know “by reflection alone” that (4) and (5) constitute good reasons for believing (6)? Where deductive inferences are concerned, it is fair to assume that the standard the inference must meet to be a good one is deductive validity. And to say that an argument is valid is to say that it has a logical form to which there is no counterexample. Just being able to identify the logical form of an argument requires theoretical knowledge which Granny doesn’t have right now, and it is not clear that she could obtain it if she just sat down and thought about it. Philosophers such as BonJour (2001), however, have contended that naïve subjects can very easily know that inferences such as the one above are good ones. This has the surprising consequence that one can know that his deductive inferences are good ones without knowing that they are valid. There is another problem for the view that seems even more devastating. Some argue that no one could ever in principle know that all of his inferences are good ones prior to making them.8 The argument for this claim borrows from Lewis Carroll’s (1895) famous note ‘What the Tortoise Said to Achilles’. SII requires that Granny know that (4) and (5) constitute good reasons for (6). Granny’s requisite knowledge therefore must consist in knowledge of something like the following: 8. Boghossian (2001).

117

7. Necessarily: p o ((p oq) o q) Suppose Granny knows that (7) is true. Does this suffice for Granny to know that (4) and (5) constitute good reasons for (6)? Boghossian, the most prominent proponent of the inferentialist view of the a priori, says that if we demand that Granny’s warrant for (6) is dependent on her knowing something like (7) then “we commit ourselves to an unstoppable regress”. (Boghossian, 2001, 639) If (7) is part of Granny’s justification, her reasoning is not merely an inference from (4) and (5) to (6). Rather, it takes the following logical form: 8. 9. 10. 11. 12.

p o ((p o q) o q) p (p o q) o q poq Therefore q

But again, if this is actually Granny’s argument, we must ask whether Granny knows that the inferences in this argument are good ones. And again, we can grant Granny some further logical knowledge and reconstruct Granny’s reasoning with a third argument. But this will only raise the same question all over again. The regress is up and running. The conclusion is that SII requires too much, not just of Granny, but of anyone; even the logically sophisticated. Our warrant for belief in the conclusions of our inferences cannot depend on an antecedently justified belief in the truth-conduciveness of those inferences. “At some point is must be possible simply to move between thoughts in a way that generates justified belief ”. (Boghossian, 2001, 639) This Carrollian argument, if sound, does three things. First, it demonstrates the falsity of SII. It also undermines the first paragraph of this section’s criticism of rule circular justifications for basic logical beliefs because that criticism presupposed a commitment to SII. In addition, the Carrollian regress argument independently motivates us to accept a certain sort of epistemologistical externalism. Subjects can reason in accordance with at least some rules of inference without knowing or even being able to know in advance that such rules are valid. Once we acknowledge the existence of “blind but blameless reasoning” we can start to see how an inferentialist can skirt the charge of begging the question. Since begging the question requires that one believe in advance

118

that his conclusion is true, the Carrollian regress argument opens the door to there being a rule circular argument for the validity of MPP that is not question-begging. But open doors often lead to vermin. 3. Meaning-constituting rules If we allow that rules may be used to justify themselves, then won’t we have to allow that ridiculous inference rules such as tonk-introduction and tonk-elimination are justified?9 The validity of the rules governing ‘tonk’ can be “proven” as long as we avail ourselves of those rules. If inferentialism is to have even a chance at succeeding, there must be some non-ad hoc way to avoid these consequences. Just as the inferentialist can independently motivate his rejection of SII by appeal to the Carrollian regress argument, this problem can be blocked on independent grounds by appeal to some apparatus from the philosophy of language. Conceptual Role Semantics (CRS) is the idea that the meaning of a word or concept is determined by the role that it plays in a certain set of inferences.10 The inferences that determine the meaning of a word or concept are the “meaning-constituting inferences” for that word or concept. For example, a proponent of CRS might say that ‘bachelor’ means what it does in virtue of the fact that ‘x is a bachelor’ and ‘x is an unmarried man’ are interderivable. In addition to providing an account of the nature of concepts, an account of concept possession also drops out of CRS. It is in virtue of understanding ‘x is a bachelor’ that thinkers who understand the terms are disposed to infer ‘x is an unmarried’ or ‘x is a man’. And since these inferences are meaning-constituting, the inferences can successfully transmit warrant from premises to conclusion even if the thinker in question has no beliefs concerning the nature of this inference. In the same way, if CRS is true, then the logical constant ‘if, then’ gets its meaning from a proper subset of the inferences that it participates in. If MPP is a meaning-constituting inference for ‘if, then’, then there is a 9. ‘Tonk’ is Prior’s (1960) logical pseudo-constant which is stipulated to have the same introduction rule as disjunction and the same elimination rule as conjunction. Endorsement of ‘tonk’ has the unhappy consequence that if anything is true then everything is true. 10. Boghossian (2000) prefers to construe it in terms of words in the language of thought. Whether one understands the basic idea as applying to concepts or words is not important for the purposes of this paper.

119

good explanation for the fact that Granny can employ MPP and arrive at justified beliefs blindly. It is simply in virtue of the fact that she possesses the concept ‘if, then’ that she is able to justifiably believe (6) on the basis of (4) and (5) (provided of course that she has independent grounds for (4) and (5)). Contra SII, prior justified belief in the validity of MPP is not necessary. These considerations motivate adoption of the following principle. (E) S is entitled to infer according to M, independently of having supplied an explicit justification for M if and only if M is a genuinely meaning-constituting rule for S.11 To supply an ‘explicit justification’ for a deductive rule of inference M is to construct a cogent argument whose conclusion is that M is valid. It may seem that E contradicts SII only if SII requires that inferences be justified explicitly. But according to SII, subjects must be in a position to justifiably believe that their inferences are good ones prior to carrying those inferences out. E, on the other hand, does not have this requirement. E allows that a subject might become justified in believing that his inference is a good one only after making that inference. This, as we shall see, is why the acceptance of E and the rejection of SII are crucial to the inferentialist account of our knowledge of basic logical principles. E provides a way for the inferentialist to screen off counterexamples to his idea that rules may be used to justify claims of their own validity. In the case of ‘tonk’ “it is readily shown, by attempting to construct a truth table for ‘tonk’, that its introduction and elimination rules do not determine a meaning for it; there is no proposition expressed by sentences of the form ‘A tonk B’”. (Boghossian, 2000, 251) If ‘tonk’ is meaningless, its introduction and elimination rules cannot be meaning-constituting and therefore they cannot be employed in rule circular arguments. Appeal to Granny cases and the Carrollian regress argument are designed to broaden the set of rules one is entitled to rely on and thereby open the door for a rule circular justification of our belief that MPP is valid that is not question begging. CRS and E serve to restrict what sorts of rules we are entitled to blindly rely on and thus restrict what sorts of rules can be 11. Boghossian, (2000). The biconditional (E) is not explicitly stated by Boghossian in this paper but it is entailed by two conditionals that he endorses.

120

rule circularly justified. In the next section, two particular sorts of rule circular arguments for the validity of MPP will be examined. 4. Two rule circular arguments and the problem of begging the question again Before we sign on to the project, we should ask what exactly a rule circular argument for the validity of MPP would look like. One might say that we can easily establish the validity of MPP by proving the truth of its corresponding material conditional (MPPC): ‘((P Š Q) & P) Š Q’ in standard propositional logic (PL). Although some of what Boghossian says suggests this route, it cannot be the whole story. First, the project is to justify the belief that MPP is valid where this is understood in some extrasystematic sense that is applicable to everyday reasoners such as Granny. Inferentialists nowhere suggest that their project rides on the truth of the controversial thesis that ‘Š’ is semantically equivalent to ‘if, then’. Therefore, a proof that MPP is valid for ‘Š’ must not be what we are after. An argument which serves to justify the belief that MPP is valid, at the very least, needs to have the claim that MPP is valid as its conclusion. A proof of MPPC in PL, whose conclusion is a material conditional, clearly does not meet this requirement. But this is not to say that such a proof is completely irrelevant for it does establish: 13. MPPC is a theorem of PL. To get from here to the claim that MPP is valid we need the additional premise: 14. If MPPC is a theorem of PL then MPP is valid. From these two we can rule circularly infer that MPP is valid. To know whether an inference from 13 and 14 to the claim that MPP is valid is question begging we need to know our grounds for accepting these premises. If our acceptance of either 13 or 14 turns out to be grounded in prior belief that MPP is valid then the inference is question begging. Our basis for 13 is clear enough. We can point to the proof in PL. Our belief in 14, however, must rest ultimately on our belief that PL is a sound

121

system. And since MPP is a rule of PL, it would appear that we cannot be justified in accepting 14 without being antecedently justified in believing that MPP is valid. If this is true, then an argument for the validity of MPP based on an inference from 13 and 14 would fail for the same reason that the argument from the Holy Text does, it begs the question.12 Of course, a proof of the soundness of PL is readily available. But there is a good deal of dispute about the philosophical significance of metalogical proofs.13 The most frequent and predictable complaint is that the reasoning used in the metalanguage is of the same sort as that used in the object language and thus the metalogical proofs beg the question. But from what has been said so far, it is not at all clear that this charge sticks. The inferentialist, armed with principles such as E and Granny cases, will insist that one can use a rule without assuming in advance that the rule is valid. The worry mentioned earlier is more pressing. We are after a justification of our belief that MPP is valid understood in an extrasystematic sense. What is needed then is an argument to establish that MPP is valid for the ordinary ‘if, then’. It is for this reason that inferentialists also offer an informal argument. 14 This sort of argument, if successful, would allay the worries of those skeptical about the achievements of metalogic because it would provide a demonstration of the legitimacy of the reasoning that goes on in the metalanguage (provided of course that the metalanguage is English). 15. If ‘if, then’ means what it does, then MPP is valid. 16. ‘If, then’ means what it does. 17. Therefore, MPP is valid. This argument is obviously rule-circular. To determine whether this argument is question begging, however, we must determine what our bases for believing the premises are supposed to be. If our basis for believing one of the premises is a belief in 17 then the argument is also question begging. 12. It is important to note that this criticism does not assume any commitment to SII. The point is only that the formal proof by itself is not an argument for the validity of MPP. Other principles are needed and our acceptance of these principles is clearly based on a prior belief in the validity of MPP. 13. For starters, see Haack (1982). 14. Boghossian (1996, 376).

122

16 is another way of saying that ‘if, then’ does not have the same status as ‘tonk’. It is meaningful. Exactly how we are supposed to know something like this is a good question and may present problems15, but I won’t quarrel with it here. The problem with this argument concerns our basis for the first premise. 15 is equivalent to the claim that MPP is meaning-constituting for ‘if, then’. In other words, ‘if, then’ picks out whatever it is that makes MPP inferences valid. But what is our basis for believing that? Part of the problem here stems from familiar worries over CRS. The standard objection to CRS is that it fails to provide any clear distinction between meaningconstituting inferences and all the rest. (Fodor and Lepore, 1991) But even if one can solve this problem and explain what generally distinguishes meaning-constituting from non-meaning-constituting inferences, we still need to know what reason there is for believing that MPP falls on one side of this distinction rather than the other. Suppose we know MPP to be meaning-constituting non-inferentially. Upon understanding ‘if, then’ we can look at MPP inferences and just see that they are meaning-consituting. However plausible this view might seem to some, it is not available to the inferentialist. The idea that we can intellectually “just see” certain things upon understanding them is the very sort of traditional view that inferentialists (and others) casually dismiss as being too obscure to be philosophically illuminating. (Boghossian, 2000, 231) If skepticism about the existence or justificatory power of “intuitions” or “rational insights” is warranted, then it applies to our beliefs about semantic content just as well as it does to our beliefs concerning the truths of logic themselves. If inferentialism is true, it must be that our belief that MPP is a meaning-constituting inference, insofar as it is a priori, is justified inferentially, by way of some other set of beliefs. Therefore, the argument from 15 and 16 to 17 is non-question begging only if our belief that MPP is meaning-constituting (i.e., 15) is not based on our belief that MPP is valid. This follows from the earlier analysis of what it is to beg the question. If 15. One problem concerns whether the justification for this is supposed to be empirical or a priori. If it is the former then, it would seem, the justification for the belief that MPP is valid is also partly empirical. This may not be objectionable in itself but it would be a problem for Boghossian since his aim is to show that this justification is a priori. This sort of criticism is developed some in Laurence and Margolis (2001). On the other hand, if the basis for believing 16 is a priori then Boghossian’s account requires that it be based on some other beliefs. We are then left wondering what those could be.

123

it turns out that part of our basis for believing that 15 is true is a belief that MPP is valid, then our basis for a premise of the argument for 17 is 17 itself and the argument is fallacious. The challenge for inferentialism is to ground our acceptance of 15 in something other than our acceptance of 17. As I will argue, this challenge has not been met. There are several candidate bases for the belief that MPP is a meaning-constituting inference for ‘if, then’ but they turn out to either be too weak or ultimately grounded in the belief that MPP is valid. We might base the belief that MPP is meaning-constituting on the belief that anyone who understands ‘if, then’ is disposed to reason in accordance with it. But there are philosophers such as Lycan (1993) and McGee (1985) who explicitly claim that MPP is invalid. Whether they are correct is not the issue. The point is rather that these philosophers understand ‘if, then’ if anyone does and yet they do not appear to meet CRS’s condition for doing so. Following Williamson (2003), the example can be parlayed into a general objection to CRS. For any particular rule of inference involving a concept C, there could be someone who rejects that rule for sophisticated theoretical reasons but still understands C. The proponent of CRS may be able to slip out of trouble here depending on how he cashes out what it means to be “disposed” to infer in accordance with a certain rule of inference. Philosophers often do not practice what they preach and it may be that even those who espouse disdain for MPP or other inferential rules still happily employ them outside the seminar room. A more pressing problem for any who hope to defend the claim that MPP is meaning-constituting for ‘if, then’ with the claim that all who understand ‘if, then’ are disposed to reason in accordance with MPP, is that we must again ask what the basis for this second claim is. And, once again, I see nothing short of a belief that MPP is valid that could possibly serve justify a claim like this. The disposition to infer according to MPP will most likely be taken as evidence of understanding ‘if, then’ and absence of that disposition as evidence of semantic ignorance. If that is the case, then our belief that all who understand ‘if, then’ are disposed to reason in accord with MPP is grounded in our belief that MPP is valid and not vice versa. It would then follow that our acceptance of 15 is ultimately grounded in prior acceptance of 17 and the argument begs the question. Furthermore, it is difficult to see why the fact that people are disposed to reason a certain way would provide an adequate basis for believing that way of reasoning to be constitutive of the meanings of the concepts involved. Most would happily infer from ‘Joe is a bachelor’ that ‘Joe is a

124

guy who likes to party’ but this, I take it, is merely a contingent aspect of bachelorhood. Worse yet, people are often disposed to reason invalidly and therefore these dispositions don’t support the idea that MPP is constitutive of the meaning of ‘if, then’ any more then they do the idea that the fallacy of ignoring the base rates is constitutive of the meaning of ‘probability’. A proponent of a rule circular argument for the validity of MPP can skirt the charge of begging the question only if he can show that the argument does not contain a premise which epistemically depends on the conclusion. But inferentialists have failed to show how the crucial premise, that MPP is meaning-constituting for ‘if, then’ (i.e., 15 above), could be justified without appeal to the belief that MPP is valid. In fact, Boghossian comes very close to admitting this when he says that a rule circular argument can at best be “non-suasive”. (2000, 252) Anyone who doubts the validity of MPP will also be rational in doubting that MPP is meaning-constituting and there is no independent argument that can rationally compel him to change his mind on either. To say that anyone who doubts the validity of MPP will also doubt that MPP is meaning-constituting is not to say that the belief that MPP is meaning-constituting is based on the belief that MPP is valid. But it is difficult to see why the former would be true unless the latter were also. 5. Granny, Carroll and some other problems The fact that the inferentialist has failed to provide a non-question begging argument for the validity of MPP does not entail that one could never be provided. In one place, Boghossian (2001, 229) says that he doesn’t wish to actually present and defend a particular rule circular justification for our belief that MPP is valid but only to defend the weaker claim that our basic logical beliefs could be justified in this way. But neither Granny nor Carrollian considerations tell decisively in favor of this softer thesis. Granny cases at most show only that it is possible to employ a rule of inference and arrive at justified belief without being antecedently justified in believing that the inference is valid for some propositions. Since begging the question is not a formal fallacy of reasoning but depends on the content of the propositions believed and their epistemic bases, it is not at all clear that Granny cases generalize. A very good candidate for an exception to Granny-motivated externalism would be a case where the reasoner is employing the very same

125

rule she is attempting to justify. Second, the intuitive appeal of Granny type cases also stems from the fact that the conceptually unsophisticated are typically not aware of the kinds of inferences they are employing and we do not expect them to be. But the person who attempts to deduce the proposition that MPP is valid will certainly not be logically uninitiated and he will (or should be) well aware of the inference he employs. This is another reason to think that Granny cases are not analogous and may be of no relevance here at all. The Carrollian regress argument also fails to support the inferentialist in an even more dramatic way. SII stipulates a necessary condition on inferential knowledge. Subjects are required to be in a position to know that the inferences they employ are good ones. The regress argument gets up and running on the false assumption that this requisite knowledge must be stated as a premise in an argument somewhere. If the mind/brain identity theory is true then a necessary condition on Granny having any beliefs at all (and thus any justified beliefs) is that she has a brain. But the mind/brain identity theory clearly does not require that “I have a brain” be a premise in an argument for each belief that Granny has. In the same way, SII requires that S be antecedently in a position to know that his reasons for p are good ones but it does not require this second-order epistemological assessment to be included as a premise in the inference to p. The Carrollian regress argument confuses reasons for believing that p (i.e., things from which p is inferred) with a necessary condition for justified belief in p. The regress argument fails to support the sort of epistemological externalism that is designed to motivate adoption of the crucial principle E. Another problem concerns inferentialism’s attempt to restrict what rules can be rule circularly justified. The initial worry for those who think that circularity of any sort is epistemically acceptable was that if we allow it, then we can justify anything we wish and thus the distinction between justified and unjustified beliefs collapses. The inferentialist attempts to prevent this collapse by focusing on rule circular arguments and by distinguishing meaning-constituting from non-meaning-constituting inferences. We may be convinced that the latter distinction rules out the introduction and elimination rules for ‘tonk’. It is important to note, however, that Boghossian’s own rejection of ‘tonk’ rests primarily on the claim that ‘tonk’ is inconsistent. Even opponents of CRS can make that point. Thus the appeal to “meaning-constituting rules” doesn’t provide any new reason to abjure obviously absurd inferential rules.

126

Consider less obviously absurd but still very controversial rules: the KK-rule (From ‘S knows that p’ infer ‘S knows that S knows that p’), the KJ-rule (From ‘S knows that p’ infer ‘S is justified in believing that p’) and the S5 rule (from ‘◊ p’ infer ‘ p’). These can all be rule circularly defended.16 And, given that there is no clear procedure for determining when an inference is meaning-constituting, one can very easily assert that these rules are so without committing himself to any obvious contradiction. Thus the inferentialist approach does little toward providing us with a way to distinguish good rules from bad ones because it all turns on whether the rule in question is meaning-constituting. We have, as yet, no way of determining that. Even if worries over patently invalid rules such as those that govern ‘tonk’ are allayed, inferentialism still makes the justification of many inferential rules too easy. 6. A change of heart Remarks from Boghossian’s more recent (2003) suggest a way around this objection. There, he gives up on unrestricted commitment to E and he rejects his earlier view that only a valid inference can be meaning-constituting. He claims that there are concepts such as the pejorative ‘boche’ that, while not incoherent in the way ‘tonk’ is, nonetheless derive their meanings from invalid rules.17 Furthermore, he argues that there may be concepts whose introduction and elimination rules are not invalid but, nonetheless, one should not be entitled to rely on them simply in virtue of understanding the concept. An example of the last sort can be developed this way. Following Ramsey, Carnap and Lewis, we might represent neutrino theory, T(neutrino), as the conjunction of the Ramsey sentence (S) (x) Tx and the Carnap sentence

16. On the assumption that a proof of the corresponding conditional constitutes an argument for the validity of the rule, the arguments should be obvious. 17. Boghossian suggests that the introduction rule for ‘boche’ is “From ‘x is German’ infer ‘x is boche’” and the elimination rule is “From ‘x is boche’ infer ‘x is cruel’”. Since it is false that all Germans are cruel, at least one of these rules is invalid.

127

(C) (x) Tx o T(neutrino).18 We could then say that possession of the concept neutrino requires only acceptance of C. But now suppose someone introduces the concept neutrino+ and says that possession of this concept requires acceptance of the conjunction S & C. This person is, in effect, conditionalizing on the truth of neutrino theory. Anyone who employs this concept must thereby commit himself to the existence of neutrinos. But even if neutrino theory is perfectly true, we should not say that one is entitled to believe in the existence of neutrinos simply in virtue of understanding a concept. Boghossian’s way of handling such cases is to restrict E to only those concepts that cannot be stated in conditional form as in C. The Carnap sentence gives us a way to state possession conditions for the concept neutrino without thereby committing ourselves to the existence of neutrinos. And when such a route is available, we should take it because otherwise we “foreclose on the possible falsity of some particular set of claims about the world” (246). After all, neutrinos might not exist and we ought not make an a priori commitment to their existence if such a commitment can be avoided. The possession conditions for our basic logical concepts such as the conditional, however, cannot be given in such a non-committal way. This is because “if the conditional is one of your primitive logical constants you couldn’t conditionalize on the existence of an appropriate truth function for it, for you would need it in order to conditionalize on anything”. (247) Something similar to what was said about the possession conditions for the stipulated concept neutrino+ could be said of the controversial rules mentioned earlier. For example, someone who thinks that the S5 rule is meaning-constituting for the modal concepts, we might say, is conditionalizing on the truth of “S5 Theory”. Our understanding of possibility and necessity need not commit us to this theory. The S5 rule, it could be said, forecloses on certain possibilities in way that is avoidable. This modification seems to resolve some problems in conceptual role semantics but it has dire consequences for the inferentialist account of the a priori. Recall that the motivation for that view was on the one hand, to grant that radical empiricists are correct to reject traditional rationalism’s reliance on a priori “intuitions” but, on the other hand, to maintain that 18. For the unfamiliar, S essentially says that there is something which does what neutrino theory says that neutrinos do and C says that if S is true then neutrinos are what does what neutrino theory says that neutrinos do.

128

there is such a thing as a priori knowledge and justification. The latter is supposed to be explained by appeal to the resources of conceptual role semantics and rule circular inference. Now, if the blind entitlement to rely on an inferential rule is to be restricted to rules that cannot be conditionalized on, it follows that rule circular argurments will be acceptable only for those rules. It follows from this that the only things we know a priori are the most basic principles of logic, in fact, it seems to follow that the only things we know a priori are that modus ponens and conditional proof are valid. Even if the KJ rule, the KK rule, the S5 rule are valid, we cannot know this a priori. For the same reason, we can’t know a priori that double negation is valid (since this rule can also be stated conditionally). We can’t even know a priori that all bachelors are unmarried or that all oculists are eyedoctors. Anyone who thinks that the inference from ‘x is a bachelor’ to ‘x is unmarried’ can be seen as conditionalizing on the truth of “Bachelor Theory” and we need not join him. Whatever this modification does for the plausibility of CRS, it spells doom for the inferentialist view of the a priori. 7. A concluding ad hominem One of the main motivations behind the traditional view of the a priori has always been the claim that one cannot account for our knowledge of basic truths of logic and mathematics without appealing to some non-empirical source of justification such as rational insight or intuition. While inferentialism is designed to acknowledge the existence of a priori knowledge without commitment to such faculties, some of Boghossian’s own comments surprisingly suggest that his view cannot fulfill its promise. Boghossian considers a pseudo-rule (R) which says: from any proposition p, infer all snow is white. This rule could be used to prove its corresponding conditional. Boghossian hopes to dismiss any attempt to rule circularly justify this rule by appeal to the condition that we can rule circularly justify only those rules that are meaning-constituting. To show that this condition is not met he says, “it is obviously not part of the meaning of ‘all’ that ‘all snow is white’ can be inferred from any proposition”. (2000, 251) But to call some claim about meaning “obvious” in a way that is clearly not perceptual and to assume that this carries any epistemic weight is just to endorse something that is, if not equivalent to old fashioned rationalism, no better than it. In the same essay, Boghossian

129

also attempts to defend some of the principles that he uses to support his view of the a priori by calling them “quite intuitive” (234). This appears only three pages after he derides the traditional view’s appeal to a faculty of intuition for being obscurantist. Of course, none of this shows that the rationalist is right to think that appeal to something like intellectual insight or intuition is epistemologically indispensable. But it is suggestive. To summarize, a rule circular argument will be non-question begging only if we can be justified in believing the premises of this argument without already being justified in believing the conclusion. Two types of rule circular arguments have been proposed and they have shown to be question begging in the usual sense. While Granny cases seem to support the claim that one can sometimes reason in accordance with a rule without being justified in believing anything about that rule, they do not provide adequate support for the claim one could justify the belief that MPP is valid using MPP; nor does the Carollian regress argument. Furthermore, the inferentialist appeal to meaning-constituting inferences is unnecessary for dismissing obviously absurd rules and insufficient to curtail the problem of making justification of inferential rules too easy. Inferentialists have provided no cogent rule circular argument for any of our fundamental logical beliefs nor have they given us good reason to believe that such an argument waits to be discovered. Recent modifications to the underlying semantics of inferentialism only serve to destroy it as a general account of the a priori. And in the end, some inferentialists seem committed to a more or less traditional view of the a priori anyway.19

REFERENCES Bealer, George (2000) “A Theory of the A priori”. Pacific Philosophical Quarterly. V. 81. pp. 1–29. Boghossian, Paul (1996) “Analyticity Reconsidered”. Nous v. 30 (3) pp. 360– 391. — (2000) “Knowledge of Logic”. In New Essays on the A Priori. eds. Boghossian, P. and Peacocke, C. Oxford: Clarendon Press. 19. Versions of this paper were presented at the Fifth European Congress of Analytic Philosophy in Lisbon, Portugal and at St. Cloud State University in St. Cloud, Minnesota. Special thanks to Albert Casullo, Max Deutsch, Ed Erwin, Jeremy Morris, Michael Shaffer, and Ümit Yalçin for their comments on earlier drafts.

130

Boghossian, Paul (2001) “Inference and Insight”. Philosophy and Phenomenological Research. v. 63 (3) pp. 633–640. — (2004) “Blind Reasoning”. Proceedings of the Aristotelian Society. Supp 77, pp. 225–248. BonJour, Laurence (1998) In Defense of Pure Reason. New York: Cambridge University Press. — (2001) “Replies”. Philosophy and Phenomenological Research. V. 63 (3), pp. 673–698. Carroll, Lewis (1895) “What the Tortoise Said to Achilles”. Mind. v. 4 (14), pp. 278–280. Devitt, Michael (2005) “There is no A Priori” in Steup and Sosa (eds), Contemporary Debates in Epistemology. New York: Blackwell Publishing Co. Haack, Susan (1982) “Dummett’s Justification of Deduction”. Mind. v. 91, pp. 216–239. Korcz, Keith Allen (1997) “Recent Work on the Basing Relation”. American Philosophical Quarterly. v. 34(2), pp. 171–191. Laurence, Steven and Margolis, Eric (2001) “Boghossian on Analyticity”. Analysis. v. 61(4), pp. 293–302. Lycan, W. (1993) “MPP, RIP”. in J. Tomberlin (ed.), Philosophical Perspectives, Vol. 7: Language and Logic. Atascadero: Ridgeview Publishing. McGee, Vann (1985) “A Counterexample to Modus Ponens”. Journal of Philosophy, 82, 462–471. Peacocke, Christopher (1993) “Proof and Truth” in Haldane and Wright (eds.) Reality, Representation and Projection . New York: Oxford University Press. Prior, A.N. (1960) “The Runabout Inference Ticket”. Analysis, 21, 38–39. Sinnott-Armstong, Walter (1999) “Begging the Question”. The Australasian Journal of Philosophy. v. 77 (2), pp. 174–191. Williamson, Timothy (2003) “Blind Reasoning: Understanding and Inference”. Proceedings of the Aristotelian Society. Supp 77, pp. 249–293.

131

Grazer Philosophische Studien 73 (2006), 133–144.

CAN THERE BE EPISTEMIC REASONS FOR ACTION?1 Anthony Robert BOOTH Queen’s University, Belfast Summary In this paper I consider whether there can be such things as epistemic reasons for action. I consider three arguments to the contrary and argue that none are successful, being either somewhat question-begging or too strong by ruling out what most epistemologists think is a necessary feature of epistemic justification, namely the epistemic basing relation. I end by suggesting a “non-cognitivist” model of epistemic reasons that makes room for there being epistemic reasons for action and suggest that this model may support moral realism.

I. Can there be epistemic reasons for action? Might the following actions, for instance, be justified from the epistemic point of view: taking a trip to the library to borrow the latest book on, say, the philosophy of mind; performing a control experiment; asking an eminent professor their opinion on a matter on which they have expertise; conducting a conference or research seminar in a particular way? In this paper, I argue that there can, indeed, be epistemic reasons for action. Let me begin by saying what I take an epistemic reason to be. How are we to distinguish between different sorts of reason? My answer is this: to say that someone has a justificationary (normative) reason to \ is to positively evaluate their doing \ relative to some standard or norm. So we might say, then, that the way to classify, distinguish or sort out different sorts of reason is via appeal to the standard or norm in question. If the legitimate goal of enquiry is “truth” (or, more precisely, believing truths and avoiding falsehoods) then when we 1. Many thanks are due to an anonymous referee at Grazer Philosophische Studien for some helpful comments on an earlier draft of this piece as well as to Robin Hendry, Chris Hookway, Darrell Rowbottom, and Alan Weir, for their thoughts on the issues raised here in conversation.

are making epistemic appraisals, we are judging with regard to the attaining of that goal (which is thereby the relevant standard of appraisal). To say that S has an epistemic reason to \, then, is to appraise S’s \-ing by that standard — i.e. we ask: does \-ing help in the attainment of believing truths and avoiding falsehoods? For S to have epistemic reason to \, means that \-ing will result in S having more true beliefs and less false ones than not \-ing. If this view about what it is to have an epistemic reason is correct, then there can be such things as epistemic reasons for action, since actions can clearly help in the attainment of the goal of believing truths and avoiding falsehoods. It is not difficult to see how taking a trip to the library, for example, will often result in S’s having more true beliefs and less false ones. The positive part of this paper will be to give this view some plausibility by arguing against its opponents. The argument is: if we can pick out epistemic reasons simply via appeal to the standards of epistemic appraisal (i.e. the furthering of the truth-goal), then it is possible to have epistemic reasons for action. Now, S’s having a reason for \-ing can also mean (at least) a couple of things: it can mean that S’s \-ing can be adequately explained, it can also mean that S’s \-ing is justifiable (if not fully justified). For most of this paper, I will be concerned with whether S can coherently be said to have an epistemic reason for \-ing, where that reason is potentially capable of giving S justification for \-ing from the epistemic point of view (which I take to be from the aim of believing truths and avoiding falsehoods). However, the argument will depend, to a large extent, on whether S’s \-ing can be explained by an epistemic reason, in other words, on whether an epistemic reason can motivate S into action. I focus on the “normative” aspect of “reasons” in an effort to respond to perhaps the most obvious objection that actions, such as those cited above, which seem to include an epistemic component, are best explained by practical reason. S’s going to the library, for example, is best explained by S’s desire to know about the latest theories in the philosophy of mind — given that S wants x and \-ing will make it the case that S has x, then S has practical reason for \-ing. It seems to me that what motivates this claim is the intuition that, as a matter of fact, what motivates individuals into taking trips to libraries, performing control experiments, asking the opinion of experts etc. is their desire to know about whatever it is that performing these actions will make it the case (or help make it the case) they will know about. However, even if this were the case as a matter of fact, it would not follow that some actions (motivated by non-epistemic reasons) could not

134

be justified by different epistemic reasons. What motivates S into \-ing might not always be what gives (or doesn’t give) justification for S’s \-ing. So we may still coherently talk about epistemic reasons for action in the normative sense, even if the counter-intuition is correct. Nevertheless, as I have already mentioned, it would be non-sensical (I’ll argue) to speak of epistemic reasons for action if no epistemic reason could motivate S into action. I think that the reasons for which one might claim that no epistemic reason can motivate action emerge (wrongly) out of the considerations that deny (also wrongly) that there can be epistemic reasons for action in a normative sense. So I begin with examining three considerations as to why we can’t have epistemic reasons for action (in the normative sense) and argue that those considerations are either somewhat question-begging or prove too much (i.e. in ruling out epistemic reasons for action, they also rule out what most epistemologists think is a necessary feature of epistemic justification — namely, the epistemic basing relation). I then return to consider the question of whether an epistemic reason can ever motivate action. II. The first consideration, probably the weakest, that I’ll consider concerns whether my way (that sanctions epistemic reasons for action) is the correct way to classify different sorts of reasons, or, at least, that in particular epistemic reasons cannot solely be identified via appeal to the standards of epistemic appraisal. According to the opponent, perhaps, we also need to appeal to the objects of appraisal, namely, in the case of epistemic reasons, they might argue, belief. This last move has a surface negative bearing on the question of whether there can be epistemic reasons for action; if it is correct that there is a strong link between epistemic appraisals and appraisals involving belief, i.e. that belief and only belief can be appraised from the epistemic perspective, then actions are excluded from the arena of epistemic appraisal. What motivates the move, however? It may perhaps be motivated by the idea that, unlike action, we do not have voluntary control over belief and that coming to believe that p only ever involves considering whether p is true. For instance, Buckareff has is that:

135

Epistemic reasons are reasons for believing the truth of a proposition; they are not reasons for performing an action … When a belief is formed on the basis of epistemic reasons, the acquisition of the belief is a wholly passive affair that is quite unlike an action. If epistemic reasons provided the means of controlling attitudes, then they would share the same direction of fit and function as practical reasons. We would entertain them with the goal of actively changing our minds, not with the goal of having our minds conform to the way the world is. (Buckareff 2006, p. 310)

To paraphrase, the argument seems to run like this: (i)

Unlike action, believing is not something that can be done at will.

(ii)

If there are such things as reasons for belief, then such reasons would have to be epistemic (since non-epistemic reasons require control).

(iii) If epistemic reasons were ever reasons for action, they would function in the same way that non-epistemic reasons (such as practical reasons) do. (iv) But if epistemic reasons are such that they do not require control, then epistemic reasons do not function in the same way that nonepistemic reasons do. (v)

So there cannot be epistemic reasons for action (only beliefs can be up for epistemic scrutiny).

For fairly simple reasons the argument does not, however, do the work intended here. First of all, though we may lack the ability to believe at will directly we may be able to do so indirectly. Further, there does not seem to be any convincing reasons why we cannot use the vocabulary of action with respect to belief when we are talking about this sort of indirect control2. Secondly, the truth of doxastic involuntarism (the thesis that we cannot believe at will) does not by itself establish that there cannot be 2. Scott-Kakures 1994 thinks there is, but his argument is successfully refuted by Dana Radcliffe 1996. I won’t go into the details of this exchange since there are stronger reasons to think that this consideration fails in this context.

136

non-epistemic reasons for belief (that coming to believe that p only ever involves considering whether p is true), since it is entirely consistent with doxastic involuntarism that we always involuntarily come to believe for practical or moral reasons for example, or that we sometimes involuntarily believe for non-epistemic reasons even though that is not usually the case (and even if beliefs can only, and should only, be evaluated from an epistemic perspective, it does not follow that actions cannot be so evaluated). Thirdly, even if it the case that, as a matter of fact, we do always come to believe only for epistemic reasons, it does not follow that we should. Finally, premise (iii) is problematic since there does not seem to be any principled way of answering the question of why we should think that if there were epistemic reasons for action, such reasons would “share the same direction of fit and function as practical reasons” (i.e. what is wrong in entertaining reasons for action with the goal of “having our minds conform to the way the world is”?) In the absence of such an answer, we should conclude that epistemic reasons do not apply solely to belief. III. The next consideration (related to the former) that tells against there being epistemic reasons for action involves asking how the epistemic aim of believing truths and avoiding falsehoods is to be attained. My opponent’s answer is that the “truth goal” can only ever be attained by synchronic means, since “diachronic questions are moral or prudential questions rather than epistemic questions” (Feldman 2000, p. 689) and the truth goal is a synchronic aim: “the goal of now having beliefs that are true and now not having beliefs that are false” (David 2001, p. 161). What motivates these claims is the thought that we cannot have a justified belief achieving the truth goal via causal means, since that would clash with the intuition that a subject may or may not be justified in believing that p independently of the consequences that believing that p may produce. Suppose, for instance, that a student comes to believe they are poor at philosophy on the grounds that they have never received a good mark on any of their philosophy assignments (despite a great deal of effort on their part) and have never said anything that their peers or tutors thought bore any philosophical insight whatsoever. So they come to believe that they are a poor philosopher on seemingly good (though obviously not infallible) grounds. However, suppose that their coming to believe this causes them

137

to drop out of the academic enterprise in favour of a career at, let’s say, McDonald’s. Suppose, further, that despite offering the student a happier time, pursuing the McDonald’s career causes the student to satisfy the truth goal less well than had they toiled on with a career in academia. If the epistemic aim can be achieved via causal means, then we’d have to say that the student is unjustified in believing themselves to be poor student of philosophy despite the overwhelming evidence in favour of it. This seems counter-intuitive. I think there are two problems with this sort of counter-consideration, however. The first is that it is problematic to consider the truth goal as a synchronic aim. Take the following reductio argument (Fumerton 1990, Maitzen 1995, Sartwell 1992, David 2001) that runs as follows: consider that if the truth goal is a synchronic aim, then it can only be attained by constitutive means (as opposed to causal means). However, if this is the case then we get into a situation where every true belief is justified and every false belief unjustified: “the reason is, roughly, that with a synchronic goal only constitutive means count, and a constituent of the goal must always be a better constitutive means than a non-constituent” (David, p. 161). If a subject had an intuitively unjustified but true belief that p then according to the truth-goal, believing that p must be a good thing constitutively, for it will serve the goal better than not believing that p. Conversely, an intuitively justified but false belief that p must be a constitutively bad thing relative to the truth-goal since it will serve the synchronic goal worse than not believing that p. So relative to the synchronic truth-goal, all true beliefs are justified and all false beliefs unjustified. As Sartwell (1992) points out, this makes any talk of epistemic justification pretty vacuous. So taking this strategy to deny there being epistemic reasons for action, I suggest, is too strong since it rules out any non-vacuous talk of epistemic reasons (for actions or otherwise). The third problem, which I consider the strongest, is that the reasons for thinking that the truth-goal is a synchronic aim are question-begging. Recall that the motivation for thinking that the truth-goal is a synchronic aim is to block the idea that someone could be justified in holding a false belief on very inadequate grounds because holding that belief would cause the bearer to subsequently hold lots of true ones. Now, I think it is important to notice that it is irrelevant here whether or not the belief is false, since the consideration is meant to leave open the possibility that one can have false but justified beliefs, and true but unjustified ones. Consequently, what is important here is that the bearer of the belief is holding

138

it on inadequate grounds, but would be “justified” in holding the belief they do if it would cause them to have more true, and less false, beliefs than they would if they did not hold it. What is wrong with having a belief held on inadequate grounds? The intuition we are being urged to evoke is that a belief based on such grounds would always be unjustified (because it would be unlikely to be true). Thus, that someone could not be justified (i.e. not fulfil the truth goal in the correct way) if what was justifying their belief was its causal relation to other possible true beliefs. But it seems that we have smuggled in from the start the way of talking about justification that aligns itself to the truth goal in the synchronic way. So it looks like the complaint runs: the truth-goal must be a synchronic aim because if it were a diachronic aim, it would not be satisfied by criteria determined by the truth-goal being a synchronic aim. IV. I move now to discuss a final, and in my opinion the most interesting, argument against there being epistemic reasons for action. Paul Moser claims that for a given subject, an epistemic reason is essentially just an indication that the proposition believed is true. He makes a distinction between a belief state and the propositional object of that state and argues that because belief states cannot be either true or false and since epistemic reasons are, for a believer, an indication of the truth of their belief, then it follows that epistemic reasons apply not to belief states but to their propositional objects. On the contrary, non-epistemic reasons, he continues, are not indications of the truth of a proposition but indications that a belief has a particular property. For example, prudential reasons for a believer “indicate that one’s belief state is prudentially valuable for someone, but do not thereby indicate that the believed proposition is true … It might be prudent for me to believe, for example, that I shall recover from an illness even though there is no indication whatsoever that I shall recover” (Moser 1989, pp. 48–49). Moser contrasts epistemic reasons with explanatory reasons for having a belief. An explanatory reason is “a reason that explains why a belief state has been formed or maintained” (Moser 1989, p. 49). As Conee puts it: “what gets justified by the answer to the question — How do you know? — is the known proposition not believing it” (Conee 2004, p. 252). According to Moser, it is clear that an explanation as to why a belief is being held is not an indication of the truth of its propositional

139

object. Further, an epistemic reason for a believed proposition is not always the explanatory reason for having the believed state: “conceivably what explains why a belief state has been formed or maintained is one thing, and what indicates that the proposition thereby believed is true is quite another” (Moser 1989, p. 49). According to Moser, for a subject to have epistemic reasons for believing that p is for the subject to have an indication that p is true. Under Moser’s view about what an epistemic reason is, there cannot be epistemic reasons for action, since actions do not bear truth-conditions3. There are, however, two problems with this sort of consideration. The first involves asking the question: are epistemic reasons the only sorts of reasons that are indications of something’s truth? Cognitivists in ethics claim that moral judgements do have propositional content, and so, for example, my belief that it is wrong to lie, if moral, is true — but I do not thereby believe that it is wrong to lie for epistemic reasons. So moral reasons, for the cognitivists, are also an indication of the truth of something. This, however, is a naïve criticism since what the moral cognitivists are really saying is that there are epistemic reasons for moral truths4. Nevertheless, the view about the nature of epistemic reasons under consideration suffers from a problem similar to that which moral cognitivism is accused of suffering from, i.e. that if moral reasons are (really) just epistemic reasons, then how do they have normative authority and/or motivational force upon agents? According to substantive realism, then, ethics is really a theoretical or epistemological subject. When we ask ethical questions, or practical normative questions more generally, there is something about the world we are trying to find out. The world contains a realm of inherently normative truths whose existence we have noted, and the business of ethics, or of practical philosophy more generally, is to investigate them further, to learn about them in a more systematic way. But isn’t ethics supposed to be practical subject, a guide to action? (Korsgaard 1996, p. 44)

So the strategy non-cognitivists take is to make a strong internal connection between moral judgements and motivation and then claim that their theory explains that connection better than cognitivist ethics does. But cognitivists can then try to explain how some sort of internal connection is possible from within their theory and so deny the inference to the best 3. Moser, however, does not make this last inference explicit. 4. Thanks to Chris Hookway for pointing this out to me.

140

explanation argument for non-cognitivism. However, if as Simon Blackburn points out, the Humean idea that belief cannot (on its own) motivate action is true, then a moral belief which functions as a reason cannot, on its own, motivate action. So the non-cognitivist is, after all, best placed to take on board the intuitions behind motivational internalism. Of course, this last argument presupposes a particularly strong version of internalism. What I want to point out here, however, is that under the view about epistemic reasons currently under consideration, there is no place for even the weakest sort of internalism. In other words, it rules out any connection whatsoever between having a reason to believe that p and being motivated to believe that p. Now, why should we think there ought to be a connection between having a reason to believe that p and being motivated to believe that p? Well, something like it would follow from epistemic deontologism (the view that epistemic justification should be cashed out in terms of there being certain epistemic duties, obligations, requirements the failure to comply with makes us blameworthy). So, if we are going to be subject to such epistemic obligations then, under the principle that ought implies can, epistemic reasons need to have the capacity to motivate us. But the view about epistemic reasons we are considering here rules this out in principle. If we can never be motivated to belief by an epistemic reason, then it seems nonsensical to attribute epistemic obligations to us. So under epistemic deontologism, epistemic reasons must apply to the belief states (and so are not just an indication of a proposition’s truth) even though the propositional objects of belief can also be constituents of the epistemic reason. So if the modus operandi, if you like, of an epistemic reason is not its being an indication of a proposition’s truth, then there is nothing that has been said so far that prohibits the possibility of there being epistemic reasons for action. Interesting as it may, or may not, be, discovering that epistemic deontologism must sanction the possibility of there being epistemic reasons for action, it does not license the claim that there definitely can be epistemic reasons for action, since there are lots of epistemologists who strongly deny the truth of epistemic deontologism (e.g. Alston 1989, 2004 thinks that epistemic justification is a merely evaluative affair and does not carry any deontic notions whatsoever). The second problem with this third type of consideration involves invoking the epistemic basing relation. This relation is normally held to be a necessary, though not sufficient, condition for epistemic justification; it states, roughly, that for a reason R to justify S’s belief that p, R must be

141

the same reason for which the belief that p is held. The debate concerning the epistemic basing relation does not centre around whether or not the relation is necessary for epistemic justification, but how we are to construe the relation, i.e. whether it should be construed in a causal, doxastic, or counterfactual way. Some of the ways it can be construed (i.e. the causal theory) are compatible with epistemic externalism and some are not. So, Moser (1989), for example, tries to give a causal account of the epistemic basing relation. However, not even a causal account of the epistemic basing relation is compatible with Moser’s view about the nature of epistemic reasons. This is because if an epistemic reason is simply an indication of a proposition’s truth then it is, in principle, impossible for a subject to hold a proposition for an epistemic reason. So even under the causal theory of the epistemic basing relation: that S’s seeing a cat in front of him causes him to believe there is a cat in front of him, cannot hold true under Moser’s view about what an epistemic reason is, since seeing a cat in front of him (the indication that the proposition believed in is true) cannot explain why he sees a cat. Since causal explanations are obviously explanatory, Moser seems to have ruled out even a causal theory of the epistemic basing relation. The reason that (causally) explains the formation of the belief is not the same reason that justifies the belief. Asking whether seeing a cat in front of S caused him to believe there was a cat in front of him is not going to answer the question of whether there is a cat in front of S. Believing a proposition because of an indication that it is true is not what gives the indication of truth to the proposition. So, the reason that “justifies” S’s belief that p cannot be the same reason for which the belief that p is held (since it is not an indication of the belief ’s truth and thus not an epistemic reason for Moser), and this is clearly in breach of the requirement of the epistemic basing relation. Now, if epistemic reasons must sometimes be able to motivate belief, then shouldn’t epistemic reasons also be able to motivate action for there to be such things as epistemic reasons for action? According to the putative opponent, only epistemic reasons can motivate belief and he might argue for this via appeal to the truth of doxastic involuntarism. Does this mean that epistemic reasons are motivationally inert with regard to action, however? Certainly not. As I have already mentioned, even if it were the case that belief can only be motivated by epistemic reasons5, it does not 5. And this claim is dubious enough, as I’ve also already mentioned, since it is consistent with doxastic involuntarism that we are always (involuntarily) motivated for non-epistemic reasons. Further is an involuntary motivation the sort that is needed for epistemic reasons to

142

follow that epistemic reasons can only motivate belief. So, in the absence of any strong argument to the contrary, we must concede it possible that epistemic reasons can at times motivate action, even though it might be the case that seemingly “epistemic actions” are most often de facto best explained by practical rationality. V. The alternative view I’m proposing might be seen, if you like, as a kind of epistemological non-cognitivism. Epistemic reasons (for belief, or otherwise), are not just indicators that a particular proposition is true. To put it in a similar way that Simon Blackburn (Blackburn 1998) does about moral judgement: although some propositional content may be constitutive of an epistemic judgement, that content won’t be its modus operandi, i.e. indications that the content of a belief is true will persist in the absence of motivation (loosely construed) while epistemic judgements (if they have anything to do with epistemic justification) necessitate motive (loosely construed). Further, if to have epistemic reason to \, simply means that \-ing will result in S having more true beliefs and less false ones than not \-ing, then it is at least possible for us to have epistemic reasons for action. As I see it, our role as enquirers might warrant (might even oblige) certain courses of action (such as reading certain books, or conducting experiments in particular ways). Further, perhaps engaging in those practises is constitutive of what it is to be an enquirer. Or at least, ruling them out of the epistemic sphere means ignoring a great deal of what we consider to be doing qua enquirers. I’d like to close by tentatively suggesting that the thesis has the following result: epistemological non-cognitivism supports moral realism when the worry about moral realism is that ethics is threatened to turn into epistemology. One may only wonder, perhaps, how the realist treatment of moral judgements can motivate if one is presupposing a cognitivism about epistemic judgements. Perhaps being an enquirer about certain truths warrants (and perhaps obliges in a way that necessarily also motivates) certain courses of action.

have normative authority? This strikes me as a difficult issue, but I won’t pursue it here since nothing I want to argue for turns on it.

143

REFERENCES Alston, W. P. 1989: “The Deontological Conception of Epistemic Justification”, in Alston Epistemic Justification: Essays in the Theory of Knowledge (Ithaca, NY: Cornell University Press). — 2005: Beyond “Justification”: Dimensions of Epistemic Evaluation (Ithaca, NY: Cornell University Press). Blackburn, S. 1998: Ruling Passions (Oxford: Oxford University Press). Bukareff, A. 2006: “Hobartian Voluntarism and Epistemic Deontologism”, Disputatio 21 pp. 303 – 319. Conee, E. 2004: “The Truth Connection”, in Conee and Feldman Evidentialism: Essays in Epistemology (Oxford: Oxford University Press). David, M. 2001: “Truth as the Epistemic Goal” in Steup (ed.) Knowledge, Truth and Duty: Essays on Epistemic Justification, Responsibility and Virtue (Oxford: Oxford University Press). Feldman, R. 2000: “The Ethics of Belief ”, Philosophy and Phenomenological Research 60 pp. 667 – 695. Feldman R. and Conee, E. 2004: “Internalism Defended” in Conee and Feldman Evidentialism: Essays in Epistemology (Oxford: Oxford University Press). Fumerton, R. 2001: “Epistemic Justification and Normativity”, in Steup (ed.) Knowledge, Truth and Duty: Essays on Epistemic Justification (Oxford: Oxford University Press). Korsgaard, C. 1996: The Sources of Normativity (Cambridge: Cambridge University Press). Maitzen, S. 1995: “Our Errant Epistemic Aim” Philosophy and Phenomenological Research 55, pp. 869–876. Moser, P. 1989: Knowledge and Evidence (Cambridge: Cambridge University Press). Radcliffe, D. 1997: “Scott-Kakures on Believing at Will”, Philosophy and Phenomenological Research 57 pp. 145–151. Sartwell, C. 1992: “Why Knowledge is Merely True Belief ”, The Journal of Philosophy 89 pp. 167–180. Scott-Kakures, D. 1993: “On Belief and the Captivity of the Will”, Philosophy and Phenomenological Research 53 pp. 77–103.

144

Grazer Philosophische Studien 73 (2006), 145–162.

WHY CONTENT RELATIVISM DOES NOT IMPLY FACT RELATIVISM Achim LOHMAR University of Köln Summary There is widespread belief among realists that it is consistent to think of the world as a totality of absolute facts and of our representation of the world as perspectival. As a pluralist Michael Lynch has challenged this view by arguing that relativism about representational content entails relativism about facts. Lynch’s ‘T-argument’ is presented and discussed in detail. It is argued not only that the ‘T-argument’ fails and that content relativism and fact absolutism are compatible, but also that content absolutism entails fact absolutism. These two points add up to a serious challenge of pluralists in committing them to the view that there are many actual worlds.

When William James espoused F. C. S. Schiller’s humanism he made human beings the arbiter of what there is. James’s humanist position combines Kantianism with the protagorean slogan that man is the measure of all things. From Kant he took over the idea that there is no content to the idea of objects or facts as independent of the conceptual powers invoked to explain the possibility of knowledge. From Protagoras as portrayed in Plato’s well-known dialogue he took over the idea that there is no unique way to represent the world correctly, that is, no conceptual fabric common to all representations for there to be knowledge. Linking these ideas together, the humanist advocated a metaphysical picture according to which there is no unique way the world is. The world, as it were, tolerates many different descriptions up to the point of conflicting ones that cannot be unified into a single comprehensive story. James writes: „We conceive a given reality in this way or in that, to suit our purpose, and the reality passively submits to the conception” (1996: 121). The relativistic Kantianism of the humanist position was revived in the philosophy of Hilary Putnam when he turned to ‘internal realism’ in

Reason, Truth, and History. One of the arguments Putnam puts forward to shake off metaphysical realism is the argument from equivalent true descriptions of the world which cannot be part of a comprehensive theory that makes up the one true story of the world. This argument is not only rooted in Jamesian humanism by way of loose tradition but arguably an argument that elaborates on the core of the humanist position. James was confident that the rationalist picture of reality as an ‘eternally complete edition of the universe’ is completely untenable once we take serious the perspectival and fluctuating character of human concepualizations of what there is. Putnam is equally confident that the realist picture of the world as the ‘totality of mind-independent objects’ has to vanish as soon as it is confronted with the fact that there are competing ways to describe reality none of which can on rational grounds be taken as privileged. Those who are inclined to realism, however, rarely feel that there is any force to this kind of arguments. Appeals to the perspectival character of descriptions or to different conceptual schemes fail to trigger in them the antirealist intuitions of pluralists. Roger Trigg, for one, maintains that realism is not committed to the view that there is only one conceptual scheme (Trigg, 1980: 94). And Thomas Nagel actually invokes the idea of a conceptual scheme that is not even translatable into ours as a genuine reason for realism (Nagel, 1986: 93–99). Realists are therefore ready to accept Putnams statement that we cannot describe the world without describing it (Putnam 1992: 123). But they will not accept that this plain and obvious fact has any interesting implications as to the dependency of facts on ways of describing the world. Facts, after all, are those things that make our true statements true and which false statements fail to fit. To account for the falsity of some statements, in particular, we cannot but take the existence of facts as not depending on the concepts we happen to use. That there is conceptual change does not rule out that changes are for better or worse as regards the truth about some subject matter. That there are indefinitely many possibilities to classify or to order a given set of objects does not rule out that some are arbitrary whereas others are not. And if there is any real substance to the view that conceptual schemes are ‘perspectives’ or ‘stances’ adopted by the user of a concept, then the perspectival character of any conceptual engagement does not by itself rule out realism because talk of perspectives is compatible with there being perspectives getting things right or wrong and even compatible with there being one privileged perspective.

146

But things seem not to be that easy for realism. Michael Lynch has argued that realism is committed to the view that thoughts or propositions are by their very nature not perspectival but independent from any conceptual scheme. If his overall argument is sound, then once we admit that what we say and think has determined truth conditions only relative to a conceptual scheme we are committed to the view of metaphysical pluralism “that there are no absolute facts”(Lynch, 1997: 1). The first section of the present paper is devoted to present this argument of Lynch that he himself calls “T-argument”. In the next two following sections I will explain why the T-argument does not work, and argue that it cannot be amended but is bound to fail. The result of the examination so far is that content relativism does not imply fact relativism and is consistent with realism. It should be expected, therefore, that likewise content absolutism has no implications as to the question of whether there are representation independent facts. But this is not the case. In the fourth section an argument will be given which shows that if there are non-relativized true propositions then there are absolute or non-relativized facts. Hence, although content relativism has no interesting metaphysical implications, content absolutism does have. In the fifth and last section I eventually will give a diagnosis of the failure of Lynch’s overall strategy to establish metaphysical pluralism and show that pluralists have to replace their view that there are many ways the world is by the view that there are many worlds. That is, instead of presenting their position in terms of the world tolerating competing descriptions, pluralists have to embrace the rather bizarre ontological view that there are many actual worlds.1 1. A master argument for establishing metaphysical pluralism? The argument Lynch puts forward to show that absolutism about facts cannot be combined with relativism about propositional content has the merit of being simple and straightforward. It uses the Truth-Schema (TS) The proposition that p is true IFF p. and the Fact-Schema

1. A view that is presented and argued for by Blais (1997).

147

(FS) It is a fact that p IFF p. which taken together exhibt the conceptual link obtaining between the concept of truth and the concept of fact: (TFS) The proposition that p is true IFF it is a fact that p. Now, any instance I of TFS together with a premiss stating the left hand side of I obviously leads to a conclusion stating the right hand side of I. Take for example the proposition that grass is green. Then we can deduce that it is a fact that grass is green: (1) The propositon that grass is green is true IFF it is a fact that grass is green. (2) The proposition that grass is green is true. Therefore: (3) It is a fact that grass is green. Let us now introduce the interpretation of propositional content given by content relativists. According to content relativism every proposition is relative to a conceptual scheme C. Regarding our arbitrarily chosen example, content relativism states “that the proposition that grass is green is the proposition that grass is green relative to C”(Lynch, 1997: 25). Along the same lines as above we can now deduce that it is a fact that grass is green relative to C: (1*) The proposition that grass is green relative to C is true IFF it is a fact that grass is green relative to C. (2*) The proposition that grass is green relative to C is true. Therefore: (3*) It is a fact that grass is green relative to C. Because the example was arbitrarily chosen, Lynch takes this argument — the T-argument — to have shown that content relativism implies fact relativism. Therefore, if one admits the relativity of propositional content then one is committed to the view that every fact is relative to some conceptual scheme. In other words, it is inconsistent to combine a pluralist view regarding the content of our thoughts and statements with a realist conception of facts. Either give up fact absolutism or accept absolutism about content.

148

If Lynch’s argument were sound, then metaphysical pluralism could be established by showing that content relativism is the true story about thought and language. And this, of course, would reshape the whole territory of the realist/antirealist debate by pushing it into a direction where everything hinges solely on the study of concepts, of conceptual schemes, and such phenomena as conceptual change. The well-known maxim of Michael Devitt not to put the semantic cart before the horse of metaphysics then becomes totally void as it would rest on a misleading presupposition. But Lynch’s argument, to my view, does not do the job it is supposed to do. This may come as a surprise. Since neither TS nor FS can sensibly be called into question, TFS is safe, and if this is so one might wonder whether there is any room left for discussion. The point I want to stress, however, has nothing to do with the validity of the T-argument presented above. It is as valid as the first argument is. Nor do I want to call into question its second premiss. This would be totally uninteresting because (2*) is introduced just for the sake of argument. The point I want to make is rather that an apt formulation of content relativism (CR) and fact relativism (FR) forecloses the possibility to use TS, FS, and TFS for the pluralist case. 2. How not to formulate content relativism In the very passage in which Lynch introduces his argument he presents content relativism by saying, “that the proposition that grass is green is the proposition that grass is green relative to C”(Lynch, 1997: 25, italic mine). If we read this as a statement of identity we get a result that is not very friendly to Lynch’s own project. Let us assume that these propositions are identical. Then, by using TFS, we get the result that a fact making a relativized proposition true is not itself a relativized fact. (4) The proposition that grass is green = the proposition that grass is green relative to C. (5) The proposition that grass is green is true IFF it is a fact that grass is green. Therefore: (6) The proposition that grass is green relative to C is true IFF it is a fact that grass is green.

149

The conclusion of this argument makes it obvious that the content of CR cannot be stated by means of (4). There are at least three reasons for this. The first one is straightforward: if a non-relativized fact can make a relativized proposition true FR cannot be an implication of CR. The second reason is that pluralism becomes indistinguishable from absolutism. Starting from (4) we can easily deduce that the fact that grass is green is the fact that grass is green relative to C. The third reason is that (6) is obviously false. The obtaining of grass’ being green cannot make it true that grass is green relative to C. This is of course not the end of the story. Lynch is not committed to state the idea of CR in the above-mentioned manner. And I even suspect that he does not really want to say what the quoted passage suggests. The basic idea of CR is clearly this one: There is no proposition independent of any conceptual scheme. If we take this basic idea as a starting point, Lynch’s position is presented more accurately as saying that there is no such thing as the proposition that grass is green. CR is then to be viewed as a thesis about what propositions there really are. CR eliminates apparently innocent candidates from the realm of genuine propositions capable of being true or false. Now, if there is no such thing as the proposition that grass is green, the proposition semantically expressed by the sentence “Grass is green” obviously cannot be the proposition that grass is green. CR therefore has to state something like the following proposition: (7) The sentence “Grass is green” expresses the proposition that grass is green relative to C and is true IFF grass is green relative to C. This explication of CR’s basic idea comes closest to what Lynch himself has to say about propositional content being dependent on conceptual schemes. Sentences, he maintains, may mean different things or express different propositions depending on the context in which they are uttered. Because there is no such thing as the one proposition expressed by “Grass is green” in every context of utterance the phrase “that grass is green” cannot be used to pick out or to specify the proposition expressed by “Grass is green” on every occasion in which this sentence is uttered. “That grass is green”, as it were, gives us only a fragment of the proposition expressed by an utterance of “Grass is green”. Because what is said (what proposition is expressed) by a sentence is only determined in a context of utterance sentences by themselves are like sketches: one

150

cannot say what they are sketches of until they are filled in with further details: And just as exactly what a sketch is a sketch of cannot be determined except in reference to a more detailed filling in of that sketch, so what is said on a particular occasion can only be understood in reference to a particular scheme of concepts. It is in this sense that each proposition can be understood as having an index to a scheme. (Lynch, 1997: 72)

If we now take (7) as a proper explication of the idea of CR then there should be no obstacle anymore for showing that CR implies FR. If we take (7) together with a premiss (8) saying that the sentence (an utterance of the sentence) “Grass is green” is true we get the conclusion (9) that grass is green relative to C. And then, using FS as a further step (10) in the argument, we can easily deduce (11) “It is a fact that grass is green relative to C”. That is, we get the desired conclusion which is in fact identical to the conclusion (3*) of Lynch’s T-argument. To be sure, the present argument circumvents the problems arising from the first proposal of how to state the content or basic idea of CR. If we take a closer look, however, it soon becomes obvious that even this argument cannot do the job of showing that CR implies FR. Why is that? On the present interpretation, CR says that there is no such thing as the proposition that grass is green. From this it follows (i) it is not possible for any sentence to express that grass is green, (ii) that the adjective “green” figuring in a sentence like “Grass is green” does not pick out the property of being green, (iii) that there is no such property as being green. Lynch himself subscribes to this view when he states: Fact relativism […] amounts to saying that there is no such way the world is [i. e. no such fact] as roses being red simpliciter, but only roses being red relative to C. (Lynch, 1999: 581)

So, the general metaphysical implication of CR interpreted along the present lines is a thesis as to what properties there really are. Following Lynch’s own presentation we have to say that “green” picks out the property of being green relative to C. For otherwise the sentence “grass is green” could not possibly be true because it is (according to FR) impossible for grass or anything else to be green simpliciter. The truth of (an utterance of ) the sentence “Grass is green” therefore exclusively depends on the question of whether grass exemplifies the property of being green relative to C. That is, if grass exemplifies it, then the sentence is true, if it does not exemplify it,

151

then the sentence is false. To put it differently: If it is a fact that grass has the property of being green relative to C, it cannot be the case that grass fails to exemplify that very property. Either it has it or it has it not — end of the story. Even the argument under present consideration therefore cannot show that CR implies FR. This is because the putative fact, that grass is green relative to C, is as little a relative fact as the fact that grass is green. And so for the corresponding propositions. The proposition that grass is green relative to C is as little a relative proposition as the proposition that grass is green. Even in the frame of this quite artificial construction of the truth conditions of “Grass is green” nothing at all hinges upon any conceptual scheme for “grass is green” to take the value ‘true’ or ‘false’.2 The truth value of “grass is green” does not depend on and does not vary with conceptual schemes only because in the description of its truth conditions reference is made to a conceptual scheme.3 So, the only thing which Lynch’s argument does show is that if the truth conditions of a sentence S are not to be specified by means of p but by means of q it is the fact that q that makes S true. But this is surely no philosophical news. And therefore the argument fails to show that CR implies FR. The explanation of the failure of the argument resides in its improper statement of the contents of CR and FR. 3. How correctly to formulate content- and fact relativism How should CR and FR be explained? What is the real content of CR and FR? Let us cling to the straightforward proposal. CR says that there 2. Cappellen and Lepore (2005: 132–136) make a similar point in arguing against radical semantic contextualism. If the contextualist does not rest content with his purely negative claim that the truth value of every sentence depends on the context in which it is uttered, and goes on to offer a positive account of the semantic value of sentences he is at once forced into specifying general truth conditions. Take for example the sentence “The leaves are green”. According to contextualism it cannot be the fact that the leaves are green which makes this sentence true. But what is it that can make this sentence true if not that fact? Contextualism says that for the sentence “The leaves are green” to be true the leaves must count as green on the occasion of an utterance of this sentence. But saying something like this just amounts to specifying general, i.e. context-independent truth conditions: “The leaves are green” is true IFF on an occasion of utterance of “The leaves are green” the leaves count as green. 3. Of course, for the possibility of anything to exemplify the property of being green relative to C there must be a conceptual scheme C. But this is not to say that whether something exemplifies that property depends on a conceptual scheme.

152

are no propositions independent of any conceptual scheme. FR says that there are not facts independent of any conceptual scheme. So we can state the basic ideas of CR and FR as follows: (CR) For every proposition P there is at least one conceptual scheme C such that P is relative to C. (FR) For every fact F there is at least one conceptual scheme C such that F is relative to C. These propositions surely capture what the pluralist really wants and is committed to say. And they capture also what Lynch on several occasions does say. As a pluralist Lynch obviously rejects an absolutist picture of the contents of our beliefs and assertions just as much as he opposes a realist picture of reality. He opposes content absolutism according to which there are propositions that “would have been true if there had never been such a language as English, if sapient beings had never evolved on earth, and if life had never existed anywhere in the universe.” (Van Inwagen, 1998: 95) And he opposes even a moderately realist picture of reality according to which “most of the entities and facts with which we have cognitive and practical dealings […] exist independently of any particular conceptual scheme.”(Alston, 2001: 58) Against this background we can now spell out how CR und FR work on the level of particular propositions and facts. CR is obviously not to be viewed as a thesis about what propositions there really are. FR is likewise not to be viewed as a thesis about what facts there really are. The former rather states a condition of existence for every proposition, the latter a condition of existence for every fact. That is, CR does not imply that there is no such thing as the proposition that grass is green. And FR does not imply that there is no such thing as the fact that grass is green. What CR does imply is that the proposition that grass is green is relative to C. What FR does imply is that the fact that grass is green is relative to C. To see the crucial difference between this explanation and the accounts of FR discussed above compare the following two sentences: (i) It is a fact that grass is green relative to C. (Wrong explication of FR) (ii) The fact that grass is green is relative to C. (Correct explication of FR)

153

(i) states a fact. The that-clause specifies the content of the stated fact, and the relativization to a conceptual scheme C figures as part of the content of the fact stated. (ii), in contrast, does not state a fact. It says something about a fact.4 The fact is picked out by the singular term “the fact that grass is green” and of this fact (ii) says that it is relative to C. In (ii) the relativization to a conceptual scheme C does not figure as part of the content of the fact. The underlying problem of Lynch’s own exposition of FR thus becomes obvious. He uses the phrase “relative to C” as an element within the description of the content of facts. But he surely must not do this because FR does not say anything about the content of facts but only something about facts whatever their content happens to be. Therefore, showing that the fact that grass is green is relative to C cannot be managed on the basis of a contextualist semantics according to which the proposition expressed by “grass is green” is the proposition that grass is green relative to C. What we have said about FR is also true for CR. CR does not say that the content of a proposition always includes reference to a conceptual scheme. It says something about propositions whatever their content happens to be. This implies that we cannot phrase the idea of CR by specifying the content of a proposition expressed by a sentence: (iii) The proposition expressed by “Grass is green” is the proposition that grass is green relative to C. (Wrong explication of CR) An apt application of CR consits in a sentence whose subject term picks out a certain proposition and whose predicate term consists in the phrase “is relative to C”: (iv) The proposition expressed by “Grass is green” is relative to C. (Correct explication of CR). Now we are in a position to decide whether CR really implies FR. It should be clear by now that Lynch’s own T-argument does not show that there is a logical connection between CR and FR. There are two obvious reasons for this. The first reason is that the conclusion of Lynch’s T-argument cannot be viewed as an instance of FR. Lynch’s T-argument leads to the uninteresting 4. Of course, I do not set my heart on what it is to state a fact. So, if you prefer, the point of difference between (i) and (ii) can conveniently also be spelled out by saying that whereas (i) states a fact about grass (ii) does not state a fact about grass but a fact about a fact about grass.

154

conclusion that it is a fact that grass is green relative to C. But it does not arrive at the point which alone is of interest here, to wit, the conclusion that the fact that grass is green is relative to C. The second reason why the T-argument as presented by Lynch does not show the alleged connection between CR and FR is that its premisses are not in accordance with the basic idea of CR. Thus, whatever follows from the truth of “Grass is green relative to C” is not to be interpreted as an implication of CR. Can FR be arrived at via a correct explication of CR? Let us try. The proposition expressed by the sentence “Grass is green” is the proposition that grass is green. Now, let us assume that this very proposition is relative to C. Because the string “The proposition that grass is green is relative to C is true” is ungrammatical and senseless, we need semantic ascent in order to construe an intelligible instance of TFS.5 That is, we have to construe a second order proposition to which we can ascribe the truth predicate: “The proposition that the proposition that grass is green is relative to C is true”. Then we will get the following argument: (12) The proposition that the proposition that grass is green is relative to C is true IFF it is a fact that the proposition that grass is green is relative to C. (13) The proposition that the proposition that grass is green is relative to C is true. Therefore: (14) It is a fact that the proposition that grass is green is relative to C. (14) is obviously not an instance of FR. It simply does not say of a fact F that it is relative to a conceptual scheme C. In order to arrive at (an instance of ) FR the argument had to arrive at: (15) The fact that grass is green is relative to C. But such a conclusion, as will be obvious now, cannot be reached by means of CR in combination with TFS. Every instance of TFS necessarily exhibits the sameness of the content of a proposition and a fact. CR and 5. As far as I can see it is a mistake of Lynch (1997: 144/45) to take iterations of “is relative to C” to be as transparent as iterated applications of the truth predicate. A truth evaluation of any sentence of the form “‘P is true’ is true” can be managed by evaluating “P” itself. To truthevaluate “‘P is relative to C’ is relative to C”, by contrast, can obviously not be managed by an evaluation of “P” and likewise not by means of evaluating “P is relative to C”.

155

FR, however, state existence conditions for propositions and facts whatever their content may be. From this it follows that neither CR nor FR imply any constraint as to the content of propositions or facts. And it follows too that neither CR nor FR are reflected in the contents of propositions and facts respectively. Therefore, it is impossible for the pluralist to exploit the conceptual connection between the concept of a true proposition and the concept of a fact to make his case.6 Summing up we have to say that there is no purely conceptual pathway leading from propositions being relative to conceptual schemes to a metaphysics of facts according to which no fact enjoys the status of an independent existent. Truths about propositions do not automatically entail facts about facts. 4. Symmetry between the absolutists and relativists case? The result of our previous discussion was that CR does not imply FR. The combination of CR and fact absolutism (FA) is therefore consistent. That is, a realist is not committed to content absolutism (CA). But not to be committed to CA does not mean that CA does not imply FA. It is only to say that FA does not imply CA. So the question whether CA leads to FA is left open until now. This is a dialectically very interesting point. If the question is to be answered in the positive, then the realist can but need not make his case against pluralism by arguing for CA whereas the pluralist cannot make his case by arguing for CR. That is, realism would be in a much stronger position. To decide the question let us get clear about how to formulate CA and FA. Let us first call to mind what CR and FR really say: (CR) For every proposition P there is at least one conceptual scheme C such that P is relative to C. (FR) For every fact F there is at least one conceptual scheme C such that F is relative to C. How are CA and FA to be construed against the background of these two theses? Some might say that it is in virtue of their very nature that 6. Alston (1996: 78) seems to stress a similar point when he takes antirealist metaphysics as having no implications for what it is for a proposition to be true, i.e. for the concept of a true proposition.

156

propositions and facts do not depend on any conceptual scheme. This way of thinking will lead us to the position that there is no proposition and no fact that depend on any conceptual scheme. Understood this way the absolutist position about contents and facts is indeed incompatible with the relativist position, but no contradiction obtains between absolutism and relativism. That is, understood this way absolutism and relativism cannot both be true but both of them can be false. But a realist is surely not committed to that view. So let us assume that a realist wants to oppose the pluralist view by way of subscribing to a position that must be true if pluralism is false. Then CA and FA are to be stated as follows: (CA) Not (for every proposition P there is a conceptual scheme C such that P is relative to C). (FA) Not (for every fact F there is a conceptual scheme C such that F is relative to C). Stated this way CA amounts to asserting that there are propositions that are not relative to any conceptual scheme, and FA amounts to asserting that there are facts that are not relative to any conceptual scheme. Note, that neither CA nor FA say anything positively about propositions or facts. Both make a purely negative claim. They deny the relativity of all propositions and facts, that is, they deny that conceptual schemes have to exist for every proposition and every facts to exist. The realist, therefore, is not committed to express his view by saying of some facts that they are absolute. To express the independent existence of some facts it is sufficient to say that they are not relative. And this enables us to show that if propositions are not relative then they are made true by non-relativized facts. Let us assume that the proposition P expressed by “Grass is green” is not relative. That is, the proposition P expressed by this sentence exists even if there were no conceptual scheme. From this it follows that P is either true or false even if there were no conceptual scheme. Now, P is true IFF it is a fact that P. But if P can be true regardless of the existence of any conceptual scheme, then what makes P true if P is true must obtain regardless of the existence of any conceptual scheme. Hence, if P is not relative then if P is true the fact that P is not a relative fact. CA implies FA. Let “Pp” stand for “the proposition that p”. The argument showing that CA implies FA can then be presented as follows:

157

(1) (2) (3) (4) (5) (6) (7) (8)

Pp is not relative to C. [Premiss] Pp exists even if C does not exist. [From 1] If Pp is true then Pp is true even if C does not exist. [From 2] Pp is true. [Premiss] Pp is true even if C does not exist. [From 3,4] Pp is true IFF it is a fact that p. [TFS] It is a fact that p even if C does not exist. [From 5,6] The fact that p is not relative to C. [From 7]

I imagine that some readers may wonder whether this line of reasoning could not equally be used to establish that CR implies FR. Let us try. (1*) (2*) (3*) (4*) (5*) (6*)

Pp is relative to C. Pp exists only if C exists. If Pp is true then Pp is true only if C exists. Pp is true. Pp is true only if C exists. Pp is true IFF it is a fact that p.

But here we get into troubles. (5*) states that the existence of C is a necessary condition for Pp to be true. (6*) states that it is both necessary and sufficient for Pp to be true that it is a fact that p. Thus the combination of these two statements is inconsistent. There can be no additional necessary condition for Pp to be true if the fact that p is necessary and sufficient for Pp to be true. How to explain the problem? The problem stems from the fact that (5*) makes the existence of C part of the truth condition of Pp. But this is surely wrong and also not in accordance with the very claim of CR. CR, as we have seen, has no implications as to the truth conditions of propositions but states an existence condition for propositions. That is, CR does not say that Pp would be false if C does not exist. CR says that Pp would not exist if C did not exist. The original fault of this argument therefore lies in the transition from (2*) to (3*). (2*) implies that for there to be something capable of being true or false C must exist. But it does not imply that for something capable of being true or false to be true it must be a fact that C exists. With CA this problem does not arise because CA says that at least some propositions do not depend for their existence on the existence of a conceptual scheme. The transition from (2) to (3) is therefore without

158

problems. If there is a proposition Pp capable of being true or false independently of C then Pp is either true or false independently of C: If Pp is true Pp is true even if C does not exist, if Pp is false Pp is false even if C does not exist. If this reasoning is sound then what we have shown is that CA implies FA although CR does not imply FR. It is to be expected that this result should have consequences not only for the question how metaphysical pluralism can be established but also for the content of metaphysical pluralism itself. 5. Diagnosis and conclusion Our foregoing discussion has shown that content relativity by itself does not rule out, and is consistent with, realism’s basic claim that there are representation independent objects or facts. Let there be as much plasticity to concepts as you can imagine — you cannot show by this that the world itself is plastic7 as long as you have not shown that the character of reality is a function of concepts or a function of experience structured by concepts or something similar. Thus, the first and foremost task the pluralist is confronted with is to offer an argument for FR. Because CR does not imply FR such an argument cannot rely on considerations as to the existence conditions of propositional content. Nevertheless, pluralists have to ensure that CR is the right story about propositions because FA, which is incompatible with metaphysical pluralism, is entailed by CA. It comes as no surprise, then, that the failure of any attempt to establish FR on the basis of CR is evinced by the very examples Lynch offers to convince us of pluralism. One example is of two anthropologists keeping track of some as yet unknown and strange activities a group of islanders engages in. Here the conflict arises as to the correct categorization of the observed activities. Are they to be interpreted as a game or as a religious ritual? One of the anthropologists claims that it is a game, the other claims that the activities are to be interpreted as a religious ritual. Another of Lynch’s examples is indebted to Putnam. Three marbles are put into a bag and then it is asked how many objects there are in the bag. The commonsensical chap answers “three” the sophisticated mereologist answers “seven”. 7. See James (1996: 117).

159

An essential characteristic of both of these examples is that different persons give different — and allegedly incompatible but equally true — descriptions of one and the same situation. The stories have their point of departure in a given situation familiar to everyone. Next conflict of belief or judgment arises about the correct description of the situation. Note that in the way the examples are set up different persons are confronted with one and the same situation. Furthermore, the given situation is not some mysterious reality lacking any determinate character and waiting for concept-users to impose some determinate structure on it. There are, quite on the contrary, features to the situation that are independent of the conceptual choices made by different thinkers. Take the second of Lynch’s examples. It is part of its very construction that there are three [not four, not five] marbles [not marble-fragments, not apples] in a bag [not outside of a bag, not in a bag containing marbles and pencils, not in a bag which is in a bag]. And it is equally part of the very construction that there are two [not one, not three] human beings [not dogs, not angels] looking into one and the same bag [not hallucinating a bag, not looking into different bags] each of whom answers [not just make senseless noises] a given question [not construct a question of their own] differently [not express the same proposition] — and so on and so on.8 That Lynch’s example relies on there being facts independent of the conceptual choices of the conflicting parties is surely not an accident. In order to give content to the claim that there are different ways of grouping things, or different ways of categorization, the sameness of the situation which different concept-users describe differently has to be admitted. The very claim that there can be more than one way of describing a given section of reality presupposes that there is a point of departure common to, and therefore not depending on, alternative descriptions.9 Hence, in order for there to be genuine conflict among the beliefs of two different parties reference to a situation with which both are confronted has to remain fixed, for otherwise conflict cannot even arise, let alone be discovered by anyone. That arguments from conflicting descriptions of one and the same situation are as a matter of principle unsuitable for supporting metaphysical 8. For other interesting points of critique see Klenthzos (2004: 214 f.). 9. Davidson (2001) 192 reminds us that for the notion of organizing to have clear meaning we have to refer to a diversity of objects of some sort. That is, if it is taken to imply that there are no objects independent of conceptual schemes then talk of conceptual schemes as ways of organizing the world (our experience of the world) is fatally flawed.

160

pluralism is even conceded by some pluralist thinkers. Arthur Blais has forcefully argued that in order to overcome the view that there can be only one true account of the way the world is it will not do to point out that there is a plurality of conceptual schemes, or of languages etc. What a pluralist, according to Blais, really has to reject is the view that different kinds of representation or description refer to the same set of objects or are about one and the same situation. Commenting on Hartry Field’s proposal that there might be many truths because the concepts we use when describing the world are not inevitable, Blais writes: Field’s suggestion is that we might use alternative predicates to differently describe the world of representation independent objects, but then there is still the one and only actual world that is allegedly described differently […]. I maintain that Field’s suggestion cannot yield a multiplicity of truths. If there were just one actual world that consisted of representation independent objects, there could only be one truth, and it would be the one, true, and complete description of these objects. […] Such objects would be distinct from one another by virtue of the properties that they possess, or fail to possess, independently of whatever concepts or predicates we happen to possess. The complete description of all the representation independent properties that all the representation independent objects possess would be the truth. (Blais, 1997: 14/15)

Blais’ comment is in full agreement with the result of our previous discussion, namely that in order to establish his position the pluralists first and foremost task is to argue in favor of FR. This point is of great importance. For to argue in favor of FR and subsequently in favor of the view that there is more than one conceptual scheme relative to which objects or facts exist will in fact change the way pluralism is to be characterized. Instead of saying that there are many ways the world is pluralism is rather to be stated as the view that there are many actual worlds. Such a commitment is in fact nothing but a consequence of the basic claims of pluralism. If every object and fact is relative to a conceptual scheme, and if there are different conceptual schemes, then for two different conceptual schemes C and D there are different sets of objects or facts comprising everything there is relative to C, and comprising everything there is relative to D, respectively.10 The set of objects or facts comprising everything there is 10. Note that the relativity claim is a modal claim according to which it is by their very nature that objects or facts exist only relative to conceptual schemes. That is, one cannot say of a fact F that it is relative to conceptual scheme C if F does not depend for its existence on C.

161

relative to C is the C-world, and the totality of objects or facts which are relative to D is the D-world. Such an ontology is certainly eccentric at the utmost; and beyond that I suspect that it is impossible to give an example of an alternative actual world — an example that does not just turn out to be a presentation of different views of the one world we happen to be parts of.11

LITERATURE Alston, William P. 1996: A Realist Conception of Truth. Ithaca/London. Alston, William P. 2001: A Sensible Metaphysical Realism. Marquette UP. Blais, Andrew L. 1997: On the Plurality of Actual Worlds. Amherst. Cappelen, Herman/Lepore, Ernie 2005: Insensitive Semantics. A Defence of Semantic Minimalism and Speech Act Pluralism. Oxford. Davidson, Donald 2001: “On the Very Idea of a Conceptual Scheme” in: Davidson, Donald: Inquiries into Truth and Interpretation. Oxford, 183–198. James, William 1996: Pragmatism. An New Name for Some Old Ways of Thinking. Cambridge (Mass.)/London, 1–166. Khlentzos, Drew 2004: Naturalistic Realism and the Antirealist Challenge. Cambridge (Mass.)/London. Lynch, Michael P. 1999: “Relativity of Fact and Content”, in: The Southern Journal of Philosophy 37, 579–595. Lynch, Michael P. 1998: Truth in Context. An Essay on Pluralism and Objectivity. Cambridge. Nagel, Thomas 1986: The View from Nowhere. New York/Oxford. Putnam, Hilary 1981: Reason, Truth and History. Cambridge. Putnam, Hilary 1992: Renewing Philosophy. Cambridge (Mass.)/London. Trigg, Roger 1980: Reality at Risk. A Defence of Realism in Philosophy and the Sciences. Brighton/Totowa. Van Inwagen, Peter 1998: “On Always Being Wrong”, in: Midwest Studies in Philosophy XII, 95–112.

11. Many thanks to Jan Opsomer for his helpful comments.

162

Grazer Philosophische Studien 73 (2006), 163–178.

SO-FAR INCOMPATIBILISM AND THE SO-FAR CONSEQUENCE ARGUMENT Stephen HETHERINGTON University of New South Wales Summary The consequence argument is at the core of contemporary incompatibilism about causal determinism and freedom of action. Yet Helen Beebee and Alfred Mele have shown how, on a Humean conception of laws of nature, the consequence argument is unsound. Nonetheless, this paper describes how, by generalising their main idea, we may restore the essential point and force (whatever that might turn out to be) of the consequence argument. A modified incompatibilist argument — which will be called the so-far consequence argument — may thus be derived.

1. The search continues, as we seek to ascertain whether causal determinism is compatible with freedom of action. An elegant suggestion has come from Helen Beebee and Alfred Mele (2002), who have argued that a Humean conception of laws of nature is enough to generate such a compatibilism.1 One notable manifestation of this compatibility is claimed to be the falsity of a premise in the consequence argument — which, as Beebee and Mele rightly observe, is currently ‘the most influential argument in circulation for incompatibilism’ (ibid.: 206). Unfortunately, though, we will see in this paper that incompatibilism is not as easily overcome as Beebee and Mele believe it to be. At best (and if a Humean conception of laws is correct), they have defeated the consequence argument’s standard version. However, as I will also show, it is possible to isolate an underlying form to their reasoning — one that allows us to resuscitate a modified form of the consequence argument. And this will be done on grounds that are recognisably, even if not classically, Humean. So, the result will be a somewhat Humean incompatibilism — not one that reflects Hume’s own considered 1. Their sort of idea is also advanced, albeit less confidently, by John Perry (2004: 237–41).

compatibilist conclusion, of course, but one that is strengthened by some relevantly Humean commitments. My goal will be to present and develop this modified incompatibilism, not to defend it. 2. Here is a brief and informal presentation by Peter van Inwagen (1983: 16) of the consequence argument:2 If determinism is true, then our acts are the consequences of the laws of nature and events in the remote past. But it is not up to us what went on before we were born, and neither is it up to us what the laws of nature are. Therefore the consequences of these things (including our present acts) are not up to us.

To this, Beebee and Mele respond by saying that a broadly Humean conception of natural laws, if correct, renders the consequence argument unsound. On that Humean conception (2002: 203), a law of nature is basically just a uniformity in nature.3 And any such uniformity includes ‘facts about the whole duration of the world, including (in worlds that have not yet ended) about the future’ (ibid.). Hence, where L is a conjunctive proposition reporting the world’s laws, a Humean conception regards L as being in part (and contingently) about the future. And here is where Beebee and Mele make their main inferential move (ibid.: 208): On a Humean conception of laws, just as facts about the future do not deprive us of present dual ability [i.e., the ability both to perform, and not to perform, a given action], facts about what laws there are do not deprive us of such ability, either, since the relevant feature of laws just is the fact that part of what laws describe is the future. Thus, a Humean — unlike van 2. The passage is quoted by Beebee and Mele (2002: 206). There have been several formalisations and suggested revisions of the consequence argument. In § 10, I will comment on one noteworthy revision of van Inwagen’s formal argument (which I will modify in § 8 and discuss in § 9). For now, I will follow Beebee and Mele in focussing upon the brief and informal version they quote. 3. Strictly, it is more than that. As Beebee and Mele allow (2002: 203), even Humeans should think so, seeking criteria with which to exclude ‘for example accidental regularities, single-case uniformities, and uniformities with non-existent subjects’. Accordingly, Beebee and Mele (ibid.) endorse the addition of a Ramsey-Lewis rider. When I need to refer to this rider, I will do so via a ‘ceteris paribus’ clause.

164

Inwagen — can reject the general claim that ‘if P is a law of nature, then no one can render P false’.4

After all, ‘ordinary future facts themselves do not impose limits on our abilities’ (ibid.: 207) — and Humean laws (that is, laws that instantiate the Humean conception) are in part future facts. Hence, Beebee and Mele are arguing that if laws are Humean, then to some extent it is up to us what the laws are. That kind of freedom of action can exist even if causal determinism obtains. Consequently, Humean compatibilism reigns; correlatively, a prominent argument for incompatibilism does not.5 3. Let us concede, for the sake of argument, that Beebee and Mele succeed in showing that Humean laws — insofar as these are partly about, and constituted by, the future — do not bind an agent who is acting right now. Beebee and Mele infer from this that if the Humean conception of laws is correct, then the incompatibilist’s challenge fails, framed as it is in terms of laws (plus the past). But that inference is far from mandatory; or so I will argue. Instead, we may infer that Humean laws were not essential to the incompatibilist’s challenge in the first place. There is a way to (1) retain the essence of the Humean conception, while (2) allowing there to be laws which might constrain agents who are about to act, so as to (3) retain, in turn, the essence of the incompatibilist’s challenge. I will do (1) and (2) in § 4, attending to (3) in § 5. 4. My first task, then, is to show that there is a way in which a Humean may conceive of laws as being able to constrain agents, notwithstanding the future’s inability to constrain them. Beebee and Mele think (p. 208) that only the past can limit an agent — and hence that Humean laws cannot do so. Now, it is true that discussions of this issue routinely formulate causal determinism as a thesis about whether, given the past and the laws, it is now possible to act differently to how in fact one will act. Accordingly, one possible implication of Beebee’s and Mele’s analysis is that if the laws 4. For this enclosed quotation, van Inwagen (1983: 63). 5. This compatibilist form of reasoning would not be available on all conceptions of laws of nature. Consider a necessitarian analysis such as David Armstrong’s (1983: Part II). On that approach, a given law is present in the world as a universal — constituted in its entirety as soon as it is constituted at all. Hence, it is never both (i) only partly constituted in the past and (ii) partly constituted in the future.

165

are Humean, then only the past — and not the laws — could bind an agent in that respect. That is the implication accepted by Beebee and Mele. But here is an alternative possibility: (1) There can be what amount to Humean laws wholly within the past (even if there are also what amount to Humean laws that are at least partly about the future). (2) These would be the laws that could constrain an agent who is about to act — to do, or not to do, X. (1) and (2) may be explained together. When I wonder whether I am free to do — or not to do — X, I am asking whether the world that is already in place allows me this freedom: Does the past allow only one possible future? Insofar as we are talking about a person who is about to act right now, the future beyond that action’s occurring may as well not even exist right now. At any rate, that future can be ignored, as we assess whether what is already in place will permit only one possible continuation of it.6 Nonetheless, this does not entail that no laws could be constraining the action. There is a way in which the past qua past can contain laws — Humean laws in particular. This is because the past includes whatever would be the Humean laws, ceteris paribus, if the world were to end at that moment of the agent’s acting. If the world were to end right now, for instance, then — ceteris paribus — there still would have been Humean laws within it. Let LSF be a conjunctive proposition reporting these — with the world being imagined, accordingly, to be about to end. I use the subscript ‘SF’ because I call these so-far Humean laws.7 They are the Humean laws insofar as the world has existed until now — and without the world being conceived of as continuing beyond now. Thus, so-far Humean laws are subjunctively constituted. They are what would have been the world’s Humean laws if the world was to end at a given time (such as now). But — relative to that time — they need not therefore be lesser regularities, lesser Humean laws, than are those ones that will in fact turn out — relative to the world’s entire temporal history — to be the Humean laws for the world.8 6. And ignoring the future in this way is especially Humean. If we apply Hume’s justly famous sceptical thoughts on the rational reach of inductive extrapolations, we should say that the agent does not know that there will even be a world beyond that present moment. (That point is about to be accorded greater significance.) 7. The subscript ‘SF’ is an indexical. It denotes whatever has so far been the past, in relation to whatever is the moment at which the agent is about to do, or not do, X. In much of my discussion, that moment will be the (shifting) present one. 8. It should be clear that not just any regularity that has in fact obtained until a given time

166

Hence, the so-far Humean laws might, or they might not, be the same as whatever will be the Humean laws in the actual ‘finished’ or completed world. (Consequently, let L — the usual representation of the world’s laws — therefore now be understood to be a conjunctive proposition reporting just the latter laws.) Even so, the so-far Humean laws are not subjectively constituted. They are not adequately described by locutions such as ‘the laws as I have believed them to be so far’. So-far Humean laws exist no less objectively, relative to a time (such as now), than do the standard Humean laws — those which would standardly be referred to as being the Humean laws. The latter could be called the world’s final (or ‘end of the day’) Humean laws. Fundamentally, though, they remain so-far Humean laws: they are simply specific instances of the generic kind that is the category of so-far Humean laws. For they, too, are Humean laws relative to a time, and insofar as it is hypothesised to be the world’s final moment. In their case, though, the time in question just happens in fact to be the world’s final moment. That is, the world’s final Humean laws are also so-far Humean laws. They are so-far Humean laws, relative to the world’s final moment in particular. The so-far Humean laws for the present moment thus need not be the same as the world’s final Humean laws. (And LSF is identical to L when the past designated by ‘SF’ encompasses the world’s entire history — that is, when ‘now’ is the world’s final moment.) So-far laws for the present moment are significant because, even if the world is not in fact about to end, LSF — not L (as Beebee and Mele have helped us to see) — remains relevant to an agent’s freedom, or lack of it, in acting right now. In reporting the so-far Humean laws, relative to the present moment, LSF is reporting what the Humean conception would classify as being the world’s laws, if the present moment were to be the world’s final moment.9 And, again, this is how we should conceive of the present moment, insofar as we are considering the agent who is about to act right now. It is a contingent truth (if it is a truth at all) that the world is not about to end, and hence that LSF will be supplanted at all — let alone ultimately by L — as a record of the world’s laws.10 That might not occur; will be a so-far Humean law for that time. Note 3 above tells us part of why this is so. (And I am about to say more on it.) 9. In effect, this means that the so-far laws at a given time t are what would be the laws if the course of the world beyond t were to be put to one side. I am not saying that the so-far laws at t are whatever uniformities (ceteris paribus) are supervening at t upon whatever part of the supervenience base for the laws (the final laws) is in place by t. (An anonymous referee alerted me to the need to clarify this point.) 10. Any Humean laws that develop in the future could themselves be constituted by events

167

and, again, a Humean in particular should not credit the agent, who is about to act, with knowing that it will occur. So, the agent acts under both the causal and the epistemic shadows cast only by the past — including the so-far Humean laws. Any later so-far Humean laws — either the final ones or, more generally, so-far laws for later moments — are irrelevant to the agent qua imminent agent right now. As I said, Beebee’s and Mele’s Humean compatibilism allows (ibid.: 208) that the past determines our actions — without conceding that the (Humean) laws do so. But Beebee and Mele have failed to notice the potential role, in limiting us as would-be free and moral agents, of whatever so-far Humean laws have been constraining the world until the present moment — where these need not be the same so-far Humean laws as those which the world (in its ‘final’ form) will have included if it survives far beyond the present moment. From the perspective of the entire world’s history, the latter are the Humean laws (and they are distinct from those reported in LSF). Yet that need not be the correct perspective to adopt in this setting. I conceded to Beebee and Mele that, insofar as Humean laws are partly about, and constituted by, the future, these laws do not constrain us as agents right now. However, we have found an extended sense in which Humean laws could constrain us right now, without their being about, and constituted by, the future. In order to see this, we have needed only to talk, more generally, of so-far Humean laws — rather than, more narrowly, of what are usually called Humean laws. In doing so, though, we are not introducing a new kind of law; what are usually called Humean laws are simply a special case within the category of what I am calling sofar Humean laws. Humean laws are are so-far Humean laws, relative to the world’s final moment in particular. However, whether or not an agent is able to perform a particular action freely and with moral agency should be assessed in relation to the moment of her acting11— which might well not be the world’s final moment. In general, so-far Humean laws are what potentially constrain the agent at that moment of acting; and it is possible that, at different moments of acting, as the world unfolds from one that are causally determined by events that have constituted whatever so-far Humean laws have existed in the past. So, the latter could even give rise — in a law-like way — to final laws that will supplant them. This does not entail, however, that the present moment’s so-far laws are not what would — on Humean grounds — be the world’s laws if the world were to end now. Accordingly, it does not entail that the present so-far laws are not what a Humean may regard as potentially binding the person who is now on the verge of acting. 11. I have further developed this kind of point in my (2003).

168

moment to the next, successively and correlatively different so-far Humean laws play this role. To focus only on so-far Humean laws for the world’s final moment (which is to say, the world’s ‘official’ Humean laws) is not to do justice to how an agent, when acting at a particular time, would potentially be constrained qua agent at that time. 5. Yet that misplaced focus is what traditional discussions of these issues have adopted. They talk only of the laws, overlooking the more general category of so-far (Humean) laws. Traditional incompatibilists, for example, have thereby made themselves vulnerable to Beebee’s and Mele’s argument. If laws are Humean, and if these issues are presented only in terms of the laws per se, then Beebee’s and Mele’s argument is a powerful one. For — as they rightly observe — the world’s (final) Humean laws do not limit the agent at the time of acting. However, once we notice both the possibility and the aptness of applying to agents-about-to-act the concept of a so-far (Humean) law, does Beebee’s and Mele’s objection become avoidable? In my terms, what Beebee and Mele have described is the causal irrelevance to the agent, when she is about to act, of the world’s final so-far Humean laws. But the world’s earlier so-far Humean laws could remain relevant. Will they also limit us as agents? Incompatibilists — at least those accepting what I am about to begin presenting as a suitably modified kind of incompatibilism — should claim so. For our generalised Humean conception of laws — namely, the concept of a so-far Humean law — will allow the incompatibilist’s consequence argument to be revised accordingly. To see this, we may adapt van Inwagen’s earlier formulation (in § 2) of the consequence argument. Doing so gives us (as follows) this informal version of what I call the so-far consequence argument: If so-far determinism is true, then our acts right now are the consequences of the so-far laws of nature and events in the remote past. But it is not up to us what went on before we were born, and neither is it now up to us what the so-far laws of nature are. Therefore the consequences of these things (including our present acts) are not up to us. What is so-far determinism? It stands to determinism as so-far (Humean) laws stand to the final (Humean) laws. If determinism is formulated (as it usually is) in this way,

169

The laws plus the past jointly entail whatever in fact is happening, then so-far determinism is to be formulated thus: The so-far laws plus the (rest of the) past jointly entail whatever in fact is happening.12 Then the so-far consequence argument stands to the so-far Humean laws as the (usual) consequence argument stands to the final Humean laws.13 And what could be termed so-far incompatibilism would be the conclusion of that argument. My claim right now, therefore, is that the traditional incompatibilist should consider moving to advocating a so-far incompatibilism — thereby becoming (as we might say) a so-far incompatibilist. Even if (traditional) incompatibilism as such were to fall, this would not entail the demise of so-far incompatibilism. 6. Beebee and Mele objected to the original consequence argument — specifically, to its premise that we are unable to do anything about the laws of nature. But their objection leaves intact the revised version of that premise — along with so-far incompatibilism. And so-far incompatibilism is no less threatening to the possibility, at a given moment, of free agency than traditional incompatibilism ever purported to be. Even if, as Beebee and Mele argue (ibid: 207–9), we are able to alter the (final) Humean laws by having the ability still to act differently to how in fact we will act, this does not entail that we are now able to alter the so-far Humean laws. These are now beyond our causal reach. Yet we are not thereby beyond theirs, when we act right now. Perhaps they will not end up being the world’s final Humean laws — that is, once the world is finished. Nonetheless, they have been Humean laws, ceteris paribus, for the world considered purely as it has been until now (and with no assumption of its continuing beyond now). And 12. Just as (standard) indeterminism is the denial of (standard) determinism, so-far indeterminism is the denial of so-far determinism. The applicability of the so-far consequence argument to this world is therefore not to be undermined by arguing for there being indeterminism in this world. Only so-far indeterminism is relevant in that way. And the so-far consequence argument hypothesizes so-far determinism as obtaining. That argument’s goal, after all, is just to show the logical incompatibility of (i) so-far determinism, relative to a given time, and (ii) anyone’s being able to act, at that time, in some way other than in fact they act. 13. This will become clearer in §§ 8–10, where I present the formal implications of my argument.

170

not only is this an appropriate way to conceive of Humean laws — namely, insofar as they are so-far Humean laws — when we are considering whatever might constrain an agent insofar as she is about to act; it is all that a so-far incompatibilist needs them to have been if the so-far consequence argument is to be maintained. Accordingly, it is all that an incompatibilist, in order to be threatening, needs to maintain. A so-far incompatibilist endorses the incompatibility of free action with so-far causal determinism, rather than with (the standard) final causal determinism. But at any moment of acting, so-far incompatibilism is no less (and no more) worrying or constraining than is traditional — what we may call final — incompatibilism. Are our actions causally determined by the past in general, including the so-far laws? Does the past, including its so-far laws, leave us with the ability to act otherwise than in fact we will act? By generalising the Humean conception of laws so as to talk of so-far Humean laws, we enable any would-be incompatibilist to reach for these revised questions, thereby evading Beebee’s and Mele’s conceptual clutches by becoming a so-far incompatibilist. 7. So, a Humean may allow that the so-far Humean laws, as reported in LSF, are helping to causally determine our actions — even if what will be the final Humean laws, as reported in L, are not doing so. And an incompatibilist may insist, by becoming a so-far incompatibilist, that LSF is helping to constrain our actions — even if she concedes (with Beebee and Mele) that L is not doing so. Beebee and Mele argued that the Humean conception of laws of nature gives us a Humean compatibilism. But in fact what underlies that conception — the more general concept of a so-far Humean law — allows us to strengthen the case for incompatibilism, by conceiving of so-far incompatibilism. To that extent and in that sense, therefore, a subtly (even if non-standardly) Humean kind of incompatibilism may be found to arise from Beebee’s and Mele’s attempt to articulate a Humean compatibilism. 8. So far, my argument has been presented informally (following the lead of Beebee and Mele), but it may gain strength from a more formal presentation. This section, then, contains a formal version of van Inwagen’s consequence argument (not using exactly his own notation), along with Beebee’s and Mele’s objection, followed by my formal revision of van Inwagen’s argument.

171

8.1. van Inwagen’s consequence argument (1983: 93–5; 2002: 159). Let the sentential operator ‘N’ be such that ‘Np’ says ‘p, and no one has or ever had any choice about whether p’. Adopt these two rules of inference: B C

Ƒp  Np N(p Š q), Np  Nq

And focus on three propositions — P, L, and A. P describes the world as a whole, at some time in the remote past; L describes the world’s natural laws; and A describes the obtaining of some actual state of affairs later than P.14 It is assumed that causal determinism implies (by being the universal generalisation of ) this thesis of strict entailment: (P & L) o A. Then the consequence argument proceeds as follows, after first assuming causal determinism (for conditional proof ): 1. 2. 3. 4. 5. 6. 7.

Ƒ((P & L) Š A) Ƒ(P Š (L Š A)) N(P Š (L Š A)) NP N(L Š A) NL NA

[From causal determinism] [1, standard propositional logic] [2, Rule B] [Premise: the fixity of the past] [3, 4, Rule C] [Premise: the fixity of the laws]15 [5, 6, Rule C]

8.2. Beebee’s and Mele’s objection formalised. Assume a Humean interpretation of L. Let t be some given present moment, short of the world’s ‘final’ moment, at which an agent x is bringing about A. Then L can include mention of correlations which (either as types or as instances) have not occurred by t. Assume causal determinism again: it entails that (P & L) o A. This consequence is equivalent to A’s being entailed by (P & Lpast & Lnon-past) — where the non-past amounts to the present (= t) plus the future ( t). In other words, given causal determinism, we have A’s being entailed by (i) P, in conjunction with (ii) the nomic correlations 14. van Inwagen does not include this ‘later than’ qualification in his presentation of the consequence argument. His not doing so reflects his adopting (1983: 65) a two-way (and thereby stronger) conception of causal determinism. On his conception, each of the past and the future (plus, in each case, L) entails the other. 15. Premises 4 and 6 are said by van Inwagen (2002: 159) to be ‘obviously true — no one has any choice about the past; no one has any choice about the laws of nature’.

172

that have already occurred, amongst those that conjointly constitute L, and (iii) the nomic correlations that have not already occurred, amongst those that conjointly constitute L. But the latter correlations — those reported in Lnon-past — cannot be constraining x at the time t of her doing A; only P plus Lpast could do that. Hence, for all that causal determinism proves to the contrary, x could be freely doing A at t (in the sense of not being constrained in performing that action at that time). Formally: the fact that (P & Lpast & Lnon-past) entails A does not entail that (P & Lpast) entails A. Given the Humean interpretation of L, therefore, we obtain a compatibilist implication — the compatibility of causal determinism with x’s freely doing A at t. And what (in the opinion of Beebee and Mele) is thereby revealed to be the faulty step in the consequence argument? It is premise 6. Beebee and Mele (2002: 209-10) are denying that NL. For insofar as x can be constrained at t only by P and Lpast, and insofar as causal determinism’s obtaining would therefore leave open the possibility of x freely doing A at t, the following possibility is also established — the possibility that, by freely bringing about A, x also freely brings about part of the content of Lnon-past. But Lnon-past is part of the content of L. (The former conjunction, Lnon-past, is a conjunct within the latter one, L.) Hence, even given causal determinism, at least part of the content of L need not be beyond the free control of x, acting at t. Thus, we have a6 — that is, aNL. 8.3. The so-far consequence argument. Even if Beebee and Mele are correct to deny that NL, I have argued informally in earlier sections that no would-be incompatibilist need yet admit defeat. Rather, she may revive enough of van Inwagen’s consequence argument (certainly its underlying spirit) for her purposes. She need only replace his talk of L with references to LSF. (And although, as I will explain, this would require a change of name on her part, such nominal reinvention is a small price to pay.) As we saw in § 4, LSF designates — at any given moment — the so-far laws as they are at that moment.16 Then the would-be incompatibilist may reach for so-far incompatibilism, by revising van Inwagen’s argument as follows. 16. Here is a formal reminder of some features, previously described informally, of sofar laws. (I continue assuming a Humean interpretation of laws.) Let tn = the world’s ‘final’ moment. Let iLSF = the so-far laws at time ti (i d n). Then nLSF = L. Let iL = the conjunction of all correlations recorded in L that have obtained by ti (i d n). Then L = nL. Hence, nLSF = L = nL. But, for all other ti (that is, all ti such that i < n), we have this result: a(iLSF l iL), where ‘l’ designates strict, not material, implication. A so-far law at ti (i d n) in this world is, ceteris paribus, a correlation that is a (final) law in any world W which is exactly like this world except that W ends at ti.

173

The revised argument begins with an assumption (again for conditional proof ) of so-far causal determinism. So-far causal determinism is a universal generalisation of — and hence tells us that — (P & LSF) o A. And the revised argument — the so-far consequence argument, I am calling it — proceeds like this: 1. 2. 3. 4. 5. 6. 7.

Ƒ((P & LSF) Š A) Ƒ(P Š (LSF Š A)) N(P Š (LSF Š A)) NP N(LSF Š A) NLSF NA

[From so-far causal determinism] [1, standard propositional logic] [2, Rule B] [Premise: the fixity of the past] [3, 4, Rule C] [Premise: the fixity of the so-far laws] [5, 6, Rule C]

This possible version of a so-far consequence argument alters van Inwagen’s original version only minimally: the references to L become references to LSF; nothing else is changed. Nonetheless, this new argument has a clear and immediate advantage over van Inwagen’s: it evades Beebee’s and Mele’s objection. Their rejection of premise 6 (the thesis that NL) is based upon reasoning which is inapplicable to the so-far consequence argument’s replacement premise 6 (the thesis that NLSF). That is, even if (for Beebee’s and Mele’s reasons) NL is false, this does not entail that NLSF is false. The reason for that disparity is simple (and as follows). Unlike L (on its Humean interpretation), LSF is wholly about the past: so-far causal determinism thereby respects Beebee’s and Mele’s guiding principle that only the past could ever constrain an agent’s actions. If (as so-far causal determinism implies) it is true that (P & LSF) o A, then x at t, in bringing about A, is left with no alternative to performing that action at that time. Yet recall that this constraint was not in place, given traditional causal determinism — namely, the thesis that (P & L) o A — and given the Humean interpretation of L. Thus, so-far causal determinism, unlike traditional causal determinism, leaves x with no alternative to doing A — even given that, as a general consideration, only the past can constrain an action when it is being, or when it is about to be, performed. 9. Here is a possible objection to the supposed significance of § 8’s so-far consequence argument:

174

The argument is being used to show that Beebee and Mele have not shown that so-far determinism is compatible with free action — and hence that they have not really revealed (traditional) determinism to have that compatibilist potential. But this use of the argument relies upon the idea that determinism is either committed to, or flows from, so-far determinism — so that the latter’s incompatibility with freedom of action either entails or is entailed by determinism (either entailing or reflecting its incompatibility with freedom of action). And this reliance undermines the argument, because neither of determinism and so-far determinism entails the other. It is true that neither of determinism and so-far determinism entails the other. However, that sort of logical intertwining is not being relied upon in my use of the so-far consequence argument. Although the world’s final laws are a special instance within the category of so-far laws, this does not entail that to be causally determined by, in part, those final laws is to be so-far causally determined by any of those same laws.17 At any moment (other than the world’s final one), being so-far causally determined involves being constrained by so-far laws, none of which need be preserved in the world’s final inventory of laws. Yet — and this is the key point — even if, at any such (non-final) moment, we are not being constrained by the final laws, a Humean should accept that we are still being constrained, if at all, by the so-far laws. These are, after all, the only Humean laws that have been constituted purely by the past as such — the past purely on its own, independently of whatever is to follow — up to that time; and it is (as Beebee and Mele note) only the past that a Humean can allow to be constraining us as agents. By recognising all of this on the Humean’s behalf, therefore, we make the modified consequence argument — the so-far consequence argument — the one that Humeans must confront. So-far determinism enters the story I am telling, not because of any logical links to traditional determinism, but simply because it is the kind of determinism, if any, that a Humean about laws should take seriously. Clearly, Beebee and Mele did not intend leaving themselves open to any incompatibilist adaptation of the Humean framework. Nonetheless, this is the result. They saw Humeanism about natural laws as a way of evading the standard consequence argument; and so it is (given its truth). But a Humean 17. This entailment would obtain only at the world’s final moment, for actions being done at that time.

175

compatibilist believes that all present actions, say, are necessitated by a past about which we can now do nothing; and even if, as Beebee and Mele argue (in the way that was explained in § 8.2), that past does not include the final laws, a Humean must accept its including so-far laws. Indeed, a would-be Humean compatibilist should say that these are integral to the past’s having whatever limitative power it has over us. Yet with that admission, the Humean is vulnerable to the so-far consequence argument — even if not to the traditional consequence argument. At least for a Humean about laws, therefore, this vulnerability obtains, regardless of whatever logical links do, or do not, obtain between determinism and so-far determinism. Strictly speaking, if so-far determinism constrains us, this is the past on its own doing so. (The so-far laws are just part of the past.) And because we cannot now alter the past, a would-be Humean compatibilism must fail. Inadvertently, then, Beebee and Mele have helped us to see that a Humean compatibilism, incorporating a Humeanism about laws, cannot be true. The so-far consequence argument has at least this import.18 10. An impressive, and somewhat formal, philosophical literature has been generated by van Inwagen’s vigorous defense of the consequence argument. However, I will not plunge into that body of work now. I do not need to do so, because my revision (in § 8) of van Inwagen’s argument, being so minimal in its formal modifications of his argument, should return us to 18. It shows how, if so-far determinism is true, no Humean can allow our present actions to be free. But is it possible that, for a Humean, so-far determinism is not true? In that case, too, Humean compatibilism would be false. Moreover, there is indeed good reason for a Humean to regard so-far determinism as false. What follows is a brief analysis of this point. With Humean laws being universal generalisations (ceteris paribus), determinism is applied thus: From ‘(x)(Fxat a time Š Gxa moment later)’ (this being some particular, and simple, law-statement) and ‘Faright now, it would follow that ‘Gain a moment’s time’. This is how there would be an entailment of future actions by the past plus the laws — such as is claimed to fall out of determinism. But if instead we are working only with so-far laws, we have this non-entailment: From ‘so-far-untilnow-(x)(Fxat a time Š Gxa moment later)’ and ‘Faright now’, it would not follow that ‘Gain a moment’s time’. That non-entailment obtains because, for a start, the stated quantificational range of this so-far law does not apply to a’s being G in a moment’s time. The so-far law describes the world only until now; the range of quantification falls short of applying to future actions. Also, this sort of limitation is inescapable because, as a Humean should accept, the so-far laws as they have been could be about to change. Perhaps they will, in a moment’s time, be replaced by at least somewhat different so-far laws. (This possibility is ever present, as the world continues taking shape, moment to moment; or so a Humean should agree.) Thus, all of this gives us a Humean argument against so-far determinism — and hence, once more, against Humean compatibilism.

176

the motivating origins of much of that literature. Nonetheless, the putatively improved consequence argument from Alicia Finch and Ted Warfield (1998: 522) merits comment. They accept Thomas McKay’s and David Johnson’s (1996) proof that van Inwagen’s rule C is invalid.19 Do Finch and Warfield therefore relinquish the consequence argument? No, they seek to improve it, by advocating this modified version: Replace rule C with rule C*: Np, Ƒ(p Š q))  Nq. Assume causal determinism (for conditional proof ). 1. Ƒ((P & L) Š A) [From causal determinism] 2. N(P & L) [Premise: the fixity of the past and of the laws] 3. NA [1, 2, Rule C*] Beebee and Mele would regard this argument (if they were to focus upon it) as being no more successful than van Inwagen’s original version. This time (presumably on the same grounds as they advanced against van Inwagen), they would reject premise 2. After all, if NL is false (as they argue), then so is N(P & L). But any such application of Beebee’s and Mele’s reasoning would be overlooking a possible modification to the Finch/Warfield argument. Specifically, we may modify the latter argument by talking of so-far laws, instead of the laws simpliciter. This would allow their argument to be revised, with minimal formal upheaval, in the following way: Assume so-far causal determinism (for conditional proof ). 1. Ƒ((P & LSF) Š A) [From so-far causal determinism] [Premise: fixity of the past and of the 2. N(P & LSF) so-far laws] 3. NA [1, 2, Rule C*] And even if Beebee’s and Mele’s reasoning implies its being false that N(P & L), this does not also entail its being false that N(P & LSF). An action’s being causally determined yet free (with this compatibilist possibility, left open on Beebee’s and Mele’s objection, allowing part of Lnon-past to be within x’s control) does not entail the action’s being so-far causally determined yet free. So, the spirit of incompatibilism is kept alive, once the relevance and power of so-far causal determinism is acknowledged. 19. Likewise, van Inwagen (2002:165) himself accepts this proof of C’s invalidity.

177

11. The traditional letter of incompatibilism is not thereby defended, of course, because I have argued that it is so-far incompatibilism, not traditional incompatibilism simpliciter, that survives. The latter is framed only in terms of causal determinism simpliciter, not of so-far causal determinism. Yet it is the power of so-far causal determinism, not of traditional causal determinism simpliciter, that has been seen to be impervious to the Beebee/Mele kind of objection. Does that limit the significance of this paper’s results? Not at all: if laws are Humean, and if only the past ever really constrains our actions, then it could only ever have been so-far incompatibilism, not incompatibilism as traditionally formulated, that would-be compatibilists needed to confront. A correlative so-far compatibilism is the most that a Humean could hope to establish; and the so-far consequence argument threatens to undermine that ambition.20

REFERENCES Armstrong, D. M. (1983). What is a Law of Nature? Cambridge: Cambridge University Press. Beebee, H. and Mele, A. (2002). ‘Humean Compatibilism’, Mind 111: 201–23. Finch, A. and Warfield, T. (1998). ‘The Mind Argument and Libertarianism’, Mind 107: 515–28. Hetherington, S. (2003). ‘Alternate Possibilities and Avoidable Moral Responsibility’, American Philosophical Quarterly 40: 229–39. McKay, T. and Johnson, D. (1996). ‘A Reconsideration of an Argument Against Compatibilism’, Philosophical Topics 24: 113–22. Perry, J. (2004). ‘Compatibilist Options’, in J. Campbell, M. O’Rourke, and D. Shier (eds.), Freedom and Determinism. Cambridge, Mass.: MIT Press, 231–54. van Inwagen, P. (1983). An Essay on Free Will. Oxford: Clarendon Press. — (2002). ‘Free Will Remains a Mystery’, in R. Kane (ed.), The Oxford Handbook of Free Will. New York: Oxford University Press, 158–77.

20. I appreciated the comments by John Fischer and Al Mele on a draft of this paper. The paper was improved greatly by questions and advice from an anonymous referee and the journal editors.

178

Grazer Philosophische Studien 73 (2006), 179–204.

METAEMOTIONS Christoph JÄGER & Anne BARTSCH University of Aberdeen & University of Halle Summary This paper explores the phenomenon of meta-emotions. Meta-emotions are emotions people have about their own emotions. We analyze the intentional structure of meta-emotions and show how psychological findings support our account. Acknowledgement of meta-emotions can elucidate a number of important issues in the philosophy of mind and, more specifically, the philosophy and psychology of emotions. Among them are (allegedly) ambivalent or paradoxical emotions, emotional communication, emotional self-regulation, privileged access failure for repressed emotions, and survivor guilt.

1. Here’s looking at you, kid! Emotions, it is often claimed, are either positive or negative. According to a popular view both in philosophical and psychological research on emotions, joy, pleasure, and pride, for example, stand in stark contrast in this respect to fear, anger, sorrow, disgust, shame, and guilt.1 This division has a long and distinguished career. Aristotle characterizes the passions (S…TK) as affections of the soul that are accompanied by pain or by pleasure. “Such are anger, pity, fear”, he contends, “and all similar emotions and their contraries” (Rhetoric 1378a). Related ideas reverberate in current appraisal theories of the emotions, which postulate a close link between the formation of emotions and appraisals of events and situations vis-àvis their significance for one’s personal ends and goals. Situations that we 1. See for example Gordon (1987), 27: “With a few exceptions, emotions are intuitively characterizable as ‘negative’ emotions, such as fear, embarrassment, and anger, or as ‘positive’ emotions, such as pride and gladness”. Similarly, William Lyons says in an influential study (1980, 91): “All emotions seem to include an evaluation, and an operative evaluation has to be either positive or negative”. Cf. also Green (1992), 171f., and statements in the same vein in Lazarus (1991) and Roseman (2001). See Solomon (2003) for a critical discussion of the positive and negative categories.

think promote our goals, it is argued, usually elicit positive emotions, while those that get in our way tend to produce anger, sadness, fear, or frustration. Some theories even declare this kind of situational appraisal to be a necessary and sufficient condition for the generation of emotions.2 On closer inspection this picture appears to be too simple. Often we find ourselves in the grip of emotions in which positive and negative qualities are inextricably bound together. The farewell scene in Casablanca is one of the saddest ever to appear on screen. But Rick’s unforgettable gesture, in which he lifts up Ilsa’s head and looks into her eyes, not only expresses deep grief and regret, but also tenderness, pride, determination, and patriotic pathos. Sometimes we wallow in sorrow; some pleasures are embittered with pains. Fears can be pleasant, and joys may be spoiled with regret. Apparently we often confront paradoxical emotions, emotions with neither an exclusively positive nor exclusively negative valence, but which oscillate between these poles. How can this be explained? At first glance an attractive explanation is that in such cases the valences of different aspects of the emotion-eliciting event diverge. If you indulge your sweet tooth but regret that you thereby violate your diet, you have mixed feelings. While you appreciate the hedonic qualities of your act, you dislike the fact that it doesn’t comply with some (self-imposed) norm. Spinoza advocates views in this direction about the “unsteady soul” (animus fluctuatio) that is plagued by ambivalent emotions. When we simultaneously hate and love something, he argues, the object of our emotion afflicts us with pain or sorrow (tristitia), but at the same time resembles something that usually evokes pleasure or joy (laetitia) in us (Spinoza, Ethics, III, 17).3 This account may be useful for some cases, but it cannot deal with others. Spinoza would be hard pressed to explain why people occasionally enjoy indulging in self-pity or are filled with “righteous indignation”. 2. Cf. for example Lazarus (1991). This also seems to fit our pre-theoretical conceptual schemes. Linguistic studies reported, in all languages that were investigated, positive and negative categories of emotion concepts. See Shaver et al. (1987); Shaver, Wu & Schwartz (1992). 3. “Si rem, quae nos Tristitiae affectu afficere solet, aliquid habere imaginamur simile alteri, quae nos aeque magno Latitiae affectu solet afficere, eandem odio habebimus, & simul amabimus.” A similar treatment has more recently been advocated by Patricia Greenspan (1980). Her key example is “friendly rivalry”: if a good friend wins over you in a situation of rivalry, you may feel pleased and at the same time be disappointed. For further distinctions regarding ambivalent emotions see also the helpful discussion in Lenzen (2004).

180

Here it would appear inappropriate to appeal to incongruent appraisals of different aspects of the events that elicit the ambivalent emotion. We feel scandalized by a person’s behavior if it appears morally wrong, or in some other respect normatively inappropriate, to us. Acts of this kind normally lack any aspect that we also appreciate. Yet people’s indignation often does seem to be interspersed with positive feelings, perhaps with tinges of pride or satisfaction with feeling scandalized — as one may see this as an authentic and normatively adequate manifestation of one’s moral values, sense of justice, moral superiority over the perpetrator, and so on. In such examples, and many others, a phenomenon begins to surface that has not received much attention in the philosophy of the emotions. Emotions, we wish to argue, can be intentionally directed at emotions. We shall call such higher-order emotions meta-emotions. Exploring this phenomenon will not only help us to explain many allegedly ambivalent or paradoxical emotions, but also to analyze complex emotional states where the valence of some higher-order emotion matches that of a lower-level one. Not only do we enjoy certain fears or feel shame about indulging in malicious joy (discrepant valence of first-level and second-level emotion); we are also happy about falling in love or ashamed of being jealous or envious (same valence of first level and second level emotion). In what follows we argue that meta-emotions of both kinds do indeed exist. We begin with some basic distinctions regarding ordinary first-level emotions and ask how they affect the construal of meta-emotions (section 2). We look at the intentional structure of meta-emotions (section 3), outline the core idea of appraisal theories, and argue that — contrary to first impressions — accounts from this camp support our point that emotions can be intentionally directed at emotions (section 4). Next we consider empirical findings that lend further credence to our model and can be interpreted as offering genetic hypotheses regarding the conditions under which meta-emotions are elicited (section 5). Finally (section 6), we discuss further applications of the concept. In addition to the proposed treatment of certain allegedly paradoxical emotions, our account may also elucidate phenomena such as emotional communication, emotional self-control, privileged access failure for repressed emotions, and survivor guilt.

181

2. Intentional and non-intentional emotions Many emotions have intentional objects, others don’t. Let’s call phenomena of the first type (such as John’s fear of a snake, or Martha’s pleasure about her birthday present) “intentional emotions”. A mental state or event is intentional in this sense if it involves, in Brentano’s words, “object-directedness” or a “relation to a content”. As Brentano illustrates, just as in imagination something is imagined and in judgment something is accepted or rejected, “in love something is loved, in hatred something is hated, in desire something is desired, and so forth” (Brentano 1955, book II, chapter 1, section 5). Pace Brentano, however, not all “psychical phenomena” have intentional objects. Neither do all kinds of affective states. Melancholy typically occurs as a kind of emotional background hum without specific intentional contents. Emotional states or processes of this stripe are called “moods”. Some emotion words exclusively denote intentional emotions; others — such as “happiness”, “angst”, “depression”, etc. — can, depending on the context, either refer to intentional emotions or to moods. This distinction does not preclude a mood subsequently picking up intentional objects and, vice versa, what started as an intentional emotion may lose its object-directedness and degenerate into a mood.4 Intentional emotions involve relations between a subject a, some emotional state or event E, some time t during which E occurs, and some intentional object o. The logical form of this relation is Et(a,o).5 The intentional object may be an individual object or event, or a group of objects or events. You can be frightened by your neighbor’s dog, or by twenty wild boars that suddenly cross your hiking trail. In the latter case the logical form is Et(a,®o1,o2, …, on¾). The object of an emotion may also be a proposition (as when you fear that the boars decide to stampede in your direction). In this case the emotion can be represented as Et(a,p). Can all other kinds of intentional emotions be reduced to propositional emotions? We don’t think so, but will put this question on the back burner in this essay. Note, moreover, that attributions of emotions do not generally 4. This has been stressed by Goldie (2000), 17f.; for a detailed discussion see chapter 6. Goldie also argues that the distinction between emotions and moods is rather a matter of degree, and that one should put the difference by saying that emotions “have more specific objects than moods” (17, our emphasis). 5. A similar analysis is proposed by Lenzen (2004), but not with time indices. As we shall see shortly, it is important for our purposes to relate ascriptions of emotional and meta-emotional states and processes to times.

182

allow for existential generalization. If Jim fears that Satan possesses his soul, it doesn’t follow that there is such a thing as Satan (or Jim’s soul). In general, the statement that a has, at t, an emotion E to the effect that b is F, does not entail that there actually is something at which a’s emotion is directed and that has the property F. Et (a,Fb) does not entail xEt (a,Fx).6 Intentional contexts create intensional contexts, and emotion operators thus face the familiar problems of “quantifying in”. As in the case of, for example, (alethic) modal and epistemic operators, quantifying from outside the scope of an emotion predicate into the emotional context is not generally allowed. It will be licensed only if further conditions are fulfilled. Thus, whereas the principle (P#) E t(a,Fb) Š xE t(a,Fx) is false (provided that b is allowed to represent an empty singular term), the following principle appears to be valid: (P) E t(a,Fb) & x(x=b & Fx) Š xE t(a,Fx) — if a has an emotion E to the effect that b is F, and something identical with b and having the property F does exist, then something about which a has the emotion E, to the effect that it is F, exists. Weaker forms of antecedent strengthening can do the trick as well. For example, it suffices to add that there is an entity which a believes is identical with b and has the property F. This is captured in the principle: (Pc) E t(a,Fb) & xBelt(a, x=b & Fx) Š xE t(a,Fx) One central task of a theory of emotions is to spell out, for each emotion, such conditions of “quantifying in”, i.e. conditions under which de dicto attributions (such as Et(a,Fb)) do entail the corresponding de re attributions (such as xEt(a,Fx)), and vice versa. We don’t have the time to pursue these issues here. Suffice it to say that, due to the features of first-order emotions just sketched, second-order emotions don’t generally allow for existential generalizations regarding the objects of the corresponding firstorder emotions. 6. Other points are that Et (a,xFx) does not entail xFx, and that Et (a,x(Fx Š Gx)) does not entail x(Fx Š Gx). We are interpreting the quantifier as ranging over existent objects only.

183

What has been said so far poses two questions: (i) What kinds of emotion may occur in higher-order forms? (ii) What kinds of emotion can constitute intentional objects of meta-emotions? We see no reason to deny that first-level emotions of whatever type may cause all kinds of secondary affective states, including moods, agitations, and bodily feelings. But we shall restrict our use of the term “metaemotion” to intentional emotions. Since moods lack specific intentional objects, they cannot a fortiori be directed towards objects constituted by emotions. An emotional state such as “meta-melancholy” would be a conceptual impossibility in our account. Intentional forms of joy, fear, guilt, shame, etc., by contrast, can occur as meta-emotions. Typical examples are angstlust or feeling guilty over having engaged in malicious joy. As for the primary level, moods, (non-intentional) agitations and so forth can be the objects of meta-emotions. For example, you may hate the feeling of (unspecific, non-intentional) boredom, or be afraid of falling into a state of depression. Obviously, emotions can also be directed at other people’s emotions. You may be glad about your grandma’s happiness that she has won the lottery. Yet we restrict the notion of meta-emotion to intrapersonal cases: if S has a meta-emotion, its intentional object is an intentional or non-intentional emotion had by S. Note that this restriction has the consequence that not all emotions are suitable candidates for the role of meta-emotions. For example, emotions such as envy or jealousy, which are typically directed towards others and may well become intentional objects of one’s feeling guilty or ashamed, will not normally occur as intrapersonal meta-emotions. What about “negative” cases? Among the intentional objects of S’s emotion can also be the proposition that S fails to have a certain emotion. You may feel happy about, proud of, or guilty over having emotionally tuned out in some situation or with regard to some emotional disposition you used to have.7 Since in such cases there is no emotional state or disposition at the primary level, however, we shall again deny such states the accolade of “meta-emotion”. Negative cases may also occur at the meta-level. You may not be happy, or not be grumpy about experiencing (or not experiencing) certain joys or sorrows. Since in such cases as well emotions are absent at one level, these states don’t merit the title “meta-emotions” either. 7. “‘It was really interesting to see what the inside of a stomach looked like, or what human brains looked like’, police, rescue workers, medical staff, soldiers, and others sometimes say. Later on, they may feel guilty for looking at the body without emotion” (Matsakis 1999, 61).

184

Mentioning emotional dispositions brings us to one last preliminary distinction. Emotion predicates have dispositional and “occurrent-mentalstate” readings. They may refer to (i) emotions that are currently experienced; (ii) complex mental states that stretch over a longer period of time and involve a number of linguistic, behavioral, and physical liabilities; or (iii) personality traits. The latter, too, involve long-term tendencies to think, act, and feel in certain ways. Let us refer only to (ii) as the “longterm disposition” sense of emotion terms (cf. Alston 1967). When Jules says that Jim is enthusiastic (or happy or jealous) he may mean that Jim is currently experiencing these emotions; that at present Jim has strong tendencies to experience them; or that he is generally an enthusiastic (or happy or jealous) kind of person. These distinctions also apply to meta-emotions and their emotional objects. We may currently experience pleasure, annoyance, or fright regarding an emotion that we currently experience. Or we may enjoy or dislike detecting in ourselves some longterm emotional dispositions or character traits. Likewise at the meta-level. For example, one may either currently experience repulsion with respect to “weak” emotions, such as fear or sorrow (as they are currently experienced or thought to lurk around in some dispositional form). Or, we may have long-term dispositions at the meta-level or even harbor personality traits that make us dislike having, or being disposed to have, certain emotions at the primary level. (For example, cynics are often said to have a dispositional aversion to “weak” emotions.) 3. The logical structure of (ascriptions of ) meta-emotions What has been said so far suggests that the general logical form of (an ascription of ) a meta-emotion is Et (a,A t c a).8 E may, but need not, be of the same type as A, and t may, but need not, be identical with tc. Depending on the emotional object (A t c a) of the meta-emotion, the latter may also be analyzed in more fine grained ways. The structure of (an ascription of ) a meta-emotion the object of which is an intentional emotion is Et (a,Ectc (a,o)). Such meta-emotions can be further broken down into two groups, depending on whether or not the lower-level emotion is propositional. Jim may be annoyed about his fear of the neighbor’s dog, 8. A represents an affective state. The “negative” cases (which we don’t count as metaemotions proper) have the form: E t (a,aA tca), aE t (a,A tca), and aE t (a,aA tca).

185

or about his fear that the dog is going to bite him. If the emotional object is propositional the structure of the meta-emotion can be represented as Et (a,Ectc (a,p)). Does Et (a,A tc a) entail A t c a, and Et (a,Ect c (a,o)) entail Ect c (a,o)? No. The emotional objects of a meta-emotion need not exist. This follows, for example, from the fact that even occurrent emotions have dispositional components, and that we lack privileged access to our mental and behavioral dispositions. A teenager’s pride and pathos of what she believes is her eternal love does not entail that her love is in fact eternal. It doesn’t even entail that it is love at all. Moreover, there is the phenomenon of emotional repression. Repressors tend to misinterpret their emotions or, in certain circumstances, even fail to notice them at all.9 Meta-emotions are thus not generally veridical. Nor do we wish to claim that, vice versa, every emotion, in any circumstance, elicits some meta-emotion. Just as Et (a,A t c a) does not entail A t c a and Et (a,Ectc (a,o)) does not entail Ect c (a,o), At a does not entail E t c (a,At a), and Ect c (a,o) does not entail Et (a,Ect c (a,o)). Our model doesn’t exclude higher-order emotions beyond the second level, but it doesn’t threaten to carry us into infinite hierarchies of emotions. At how many levels, it may be asked, can emotions occur in human (and potentially also nonhuman) subjects? Can we feel guilty for enjoying the thrills of a bullfight? Or be ashamed of enjoying a good cry during a cheap tear-jerker? It would seem so. Exactly at how many levels meta-emotions may occur is a question that cannot be answered with purely philosophical means, however, and we shall set it to one side here. What has been said so far does, nonetheless, allow for the following general characterization of metaemotions: Meta-emotions are occurrent or dispositional intentional emotions had by some subject S at some time (or some period of time) t, taking as their objects at least one of S’s own dispositional or non-dispositional affective states or processes that occur, or are believed by S to occur, at (or during) t or some (period of ) time prior or later than t. It should be evident from this characterization that meta-emotions are, in our view, genuine emotions. They are no sui generis species of mental 9. Cf. our discussions in Jäger & Bartsch (2002), further developed in Jäger & Bartsch (2006). See also Jäger (2002) and the more detailed discussion of the limits of privileged access in Jäger (1999), chapters 1–3.

186

states. What distinguishes them from first-level emotions is simply the fact that their intentional objects are emotions. But just as higher-order beliefs are, apart from their higher-order structure, “normal” beliefs and higher order volitions, apart from their higher-order structure, “normal” volitions, higher-order emotions are “normal” emotions. So far we have been looking into the general structure of meta-emotions and their ascriptions. But exactly what kinds of mental entity are meta-emotions, and how are they elicited? 4. Meta-emotions in light of appraisal theories Until the second half of the twentieth century, both rationalist and empiricist traditions in the philosophy of the emotions were dominated by socalled “feeling theories”. According to feeling theories, the essence of an emotion is its phenomenal quality. Such views face a number of well known problems, however, which have led most authors over the last decades to abandon them.10 Some authors have reacted by suggesting radical forms of cognitivism in the analysis of emotions. Emotions, they claim, can be reduced to evaluative judgments.11 If such theories were correct, metaemotions would simply be judgments (of a certain kind) about judgments (of a certain kind); they would consist in higher order epistemic acts and the resulting attitudes. But reductive forms of cognitivism throw out the baby with the bathwater. For one thing, it does seem essential to the concept of emotion that emotions have phenomenal qualities, even though feeling theories may overemphasize this aspect. Moreover, the judgments and beliefs involved in an emotion are neither necessary nor sufficient for having it. My fear of spiders does not entail that I judge them (whether consciously or not) to be dangerous or in some other way threatening to my well-being. Furthermore, emotions typically involve other proposi10. Exceptions are Damasio (1994); the theory of “Embodied Appraisals” suggested by Jesse Prinz and his careful defense of the James-Lange theory against some standard objections (Prinz 2003, 2004); and Jenefer Robinson’s thesis that each emotion is the product of “affective appraisals” linked with bodily experiences (Robinson 2004). For helpful discussions on the difficulties that traditional feeling theories face, see the classic investigations of Kenny (1963) and Alston (1967); and Lyons (1980), chapter 1; Deigh (1994); Betzler (2003); and Voss (2004). 11. See for example Solomon (1993), 126; Solomon (1988); and Nussbaum (1990) or (2004), 196, where she summarizes her position by claiming that “emotions can be defined in terms of judgment alone”.

187

tional attitudes as well, for example volitions. Feelings of shame or sorrow, for instance, also incorporate wishes and desires to the effect that similar situations will not occur in the future. In light of these and other considerations12 we favor a view of emotions as mental phenomena with at least five components: (i) phenomenal qualities, (ii) volitions and desires, (iii) evaluative judgments and beliefs, (iv) physiological reactions, and (v) behavioral dispositions, including dispositions to behave in certain linguistic, facial, and bodily ways.13 Some of these components (maybe all of them) may also be described in terms of their functional role. It seems plausible to analyze them as systematic relations between inputs to the emotional system, system changes that are caused by these inputs, and corresponding emotional outputs.14 It would be beyond the scope of this paper to spell out the details of this model and apply each aspect to the notion of meta-emotion. We shall thus be selective and in the remainder of this essay focus on the cognitive components. What role, if not that envisaged by reductive forms of cognitivism, do judgments and beliefs have for emotions and meta-emotions? Detailed proposals regarding this role can be found in psychological appraisal theories of the emotions, as developed and empirically tested by Richard Lazarus, Klaus Scherer, Nico Frijda, and others. Such theories are best interpreted as investigating conditions under which emotions are generated. The most important appraisal criteria proposed by appraisal theories include (a) the perceived “novelty” of, or change in, a situation (its “event-likeness”); (b) its (hedonic) “valence”; and (c) its perceived relevance to the subject’s personal ends and goals. In addition, criteria are proposed such as (d) normative adequacy, (e) “agency”, (f ) controllability, and (g) what is often called “certainty” regarding the question as to whether or to what degree an emotionally relevant event or situation is or will be present. The core idea of appraisal theories is that emotions are formed on the basis of characteristic patterns of appraisal. Anger, for example, comes up if situational changes are perceived, and experienced as unpleasant, coun12. David Rosenthal (1998), for example, argues that emotions differ from propositional acts and attitudes in that the former can be verbally expressed without being conscious. 13. A related approach is an integrative definition of emotion proposed by Kleinginna & Kleinginna (1981), which, in psychological contexts, has gained wide currency. For multi-component approaches within philosophical theories, see for example Goldie (2000), or Ben-Ze’ev (2004). 14. This has been suggested by Achim Stephan (2004).

188

ter-productive with respect to one’s goals, and normatively inappropriate. Moreover, the subject might feel that s/he or someone else is responsible for the situation and that the situation could be, or could have been, controlled. We may regret that the sun is not shining; such facts too may elicit emotions. But we cannot literally be angry about such kinds of events (unless some deity is perceived to be responsible for them), for they cannot be controlled and don’t have normative dimensions. We may however be angry about the fact that rain is dripping through our recently renovated roof. This event may be perceived as a situational change, as unpleasant, not being in line with our goals and desires, and normatively inappropriate (since we believe the roofer was professionally and legally obligated to do a proper job). Moreover, we may feel that we are in a good position to control and change the situation, e.g. by having the roofer redo his job, or by employing someone else to fix the problem. Current psychological appraisal theories indirectly support the thesis that emotions can be objects of emotionally relevant appraisals. Consider for example the meta-emotional state of being angry at oneself for one’s fearfulness. In the spirit of psychological appraisal theories this meta-emotion could be reconstructed as follows. The emotion has become salient for the person (situational change/“novelty”); he is convinced that it is an instance of fear, and that it is his fear (certainty, “agency”); he experiences it as unpleasant (hedonic valence), and believes it to obstruct his goals — for example to pass a test of courage, or to make a public appearance that he believes is important for his career (goal relevance). Moreover, he may regard his fear to be normatively inadequate (normative adequacy) and believe that he could overcome it (controllability). An analogous picture suggests itself for meta-emotions with discrepant first- and second-level valence. If you feel ashamed about having indulged in malicious joy, this may plausibly be explained in terms of the following situational appraisals. You realize the emotion as your joy and acknowledge that you are responsible for it (change/novelty, certainty, “agency”); you acknowledge that the emotion has positive hedonic qualities for you (valence), but regard this to be normatively inadequate (normative adequacy). Furthermore, you believe that your malicious joy does not promote some of your goals, be it the “character goal” never to engage in this kind of emotion or the goal to present yourself as a sensible, empathetic person in the presence of a potential friend or lover (goal relevance). Finally you may also think that you can, at least to a considerable degree, control this emotion (controllability).

189

These reflections still leave various questions unresolved. The factors just outlined play an important role in current discussions, yet there is scope for debate about how exactly a potential set of necessary and sufficient conditions should be constructed. Does each situation that triggers an emotion of whatever kind involve every single criterion mentioned above? Are they jointly sufficient? Can other forms of appraisal also become a breeding ground for meta-emotions? We must leave these questions for some rainy days, and for future empirical research. The following general conclusion appears justified, however. If situational appraisals do indeed figure in the formation of emotions (which hardly seems deniable), and if this happens only roughly along the lines appraisal theories suggest, many emotions which are prima facie puzzling lose their paradoxical taint. The reason, we suggest, is that both positive and negative emotions can constitute intentional objects of higher-order emotions whose valence diverges from that of the lower-level emotion. Let us see whether we can further fortify this hypothesis with empirical evidence. 5. Empirical findings that suggest the existence of meta-emotions To date, meta-emotions have not systematically been studied from an appraisal theoretical perspective.15 There are a number of interesting investigations, especially in the field of media psychology, which nonetheless suggest that appraisal criteria that have so far only been explored for other (mainly environmental) events and situations can also be applied to emotions. Change/novelty. As already noted, most appraisal theories assume that situations must be perceived as constituting a relevant change in the subject’s environment (i.e. as “event-like”) in order to elicit an emotion. This may with justice be regarded to be the most fundamental step in emotion-related appraisals. The perception of notable changes is supposed to give rise to a state of readiness for emotion. Depending on the outcome of further appraisal processes, this “emotional readiness potential” may 15. Seminal works on higher-order affective states, however, are the studies of Peter Salovey, John Mayer and co-workers on the meta-experience of mood, the studies of Mary Beth Oliver on sad film preference, and the studies of John Gottman and coworkers on the influence of parental meta-emotions on child development (cf. Mayer & Gaschke 1988; Mayer & Stevens 1994; Salovey et al. 2002; Oliver 1993; Oliver, Weaver, & Sargent 2000; Gottman, Katz, & Hooven 1997).

190

or may not develop into a full blown emotion (Ellsworth 1994; cf. also Scherer 1984). Can emotions themselves be perceived as notable situational changes? Research on “sensation seeking” suggests that strong emotions can satisfy a desire for salient and intense experience. Sensation seeking has primarily been studied in the context of horror films and other portrayals of violence.16 The strength of the sensation seeking motive has been shown to influence subject’s emotional preferences. Sensation seekers preferred emotionally intense media stimuli, regardless of whether these elicited positive or negative emotions. Participants with a weak sensation seeking motive, by contrast, preferred neutral and positive stimuli (Zuckerman 1979; Zaleski 1984). This suggests that sensation seekers seek experiences of feeling something at all. What some people greet as welcome variations in the monotony of their everyday emotional life, others may perceive as a threat to their emotional equilibrium. In both cases however emotions seem to stand out for their subjects as significant changes in their everyday stream of consciousness. Valence/pleasantness-unpleasantness. Emotions can be pleasant or unpleasant. Feeling exulted is nicer than being grumpy. The critical question for our argument is whether hedonic valence can be regarded as an independent feature of meta-emotional appraisal. Research on “mood management” suggests that it can. According to Dolf Zillmann (2000), for example, the experience of an emotion as being pleasant or unpleasant is a function of the subject’s previous state of arousal. People typically prefer an intermediate level of arousal and consequently prefer emotions that help them to attain, or remain in, such an intermediate state of arousal. People who are bored and under-aroused enjoy emotionally arousing media stimuli, whereas over-aroused or stressed people prefer soothing media offers. These results suggest that emotions such as fear or suspense can indeed be experienced as being either pleasant or unpleasant, depending on the subject’s previous frame of mind. Goal relevance. Many authors consider the relevance of events and situations for people’s goals, needs, and desires to be the central criterion of emotional appraisal. Can emotions themselves become objects of teleological appraisal? There is empirical evidence that supports this hypothesis. Apparently anger is often welcome in conflict situations because it helps 16. Cf. Zuckerman (1979) and Zuckerman (1994), 27, where he characterizes sensation seeking as “the seeking of varied, novel, complex and intense sensations and experiences, and the willingness to take … risks for the sake of such experience”.

191

to command respect. In their research on news preferences, Silvia Knobloch-Westerwick and Scott Alter (2006) found indirect evidence for this. Participants of their study were angered by a confederate. One group of participants was then offered an opportunity to pay back the provoker. Male participants in the “pay back condition” preferred bad news that was likely to sustain their bad moods. As predicted, this effect did not occur in female participants. In the condition without an opportunity to take revenge both sexes preferred good news. This interaction of condition and gender was explained in terms of gender specific norms and social expectations regarding anger and anger expression. Women are not expected to be aggressive, whereas it is often considered “unmanly” not to fight back when provoked. It may be more in accord with the social goals of women, as well as those of men in no-revenge situations, to dissipate anger. But in the male/pay back condition, anger was goal-conducive in the experiments. According to leading appraisal theories, application of the three appraisal criteria considered thus far provides a sufficient condition for the formation of emotions. The point of the above considerations was to demonstrate that there is empirical evidence that people do indeed apply these criteria to emotions, at least under certain experimental conditions. Whereas these first three criteria are regarded as constituents of all emotions, further appraisal criteria, like “agency”, “control”, “certainty”, and “norm compatibility”, are often thought to be responsible for the differentiation of specific emotions. Is there also evidence to the effect that meta-emotions have appraisal content that goes beyond a simple positive-negative evaluation of first-order emotions? Normative appraisal. We can, and actually do, appraise emotions visà-vis their social and moral propriety. There is a long research tradition, pioneered by Paul Ekman, which deals with culturally defined display rules for emotion (Ekman & Friesen 1969; Ekman 1972; Buck 1984). Furthermore, Arlie Hochschild has shown that not only the external expression but also the inner experience of emotions, can become an object of normative appraisal (Hochschild 1983; cf. also Averill 1991; and Planalp 1999). Hochschild coined the term “feeling rules” to refer to the normative aspect of subjective emotional experience. Feeling rules have been studied most frequently in a professional context. They can be explicitly laid down in “company philosophies”, as in the case of flight attendants or Disneyland workers, who are expected never to get angry with their customers, no matter what happens (Hochschild 1983). In other cases, feeling

192

rules are “unwritten laws”, as with construction workers, among whom fear of accidents — not only talk about accidents, or behaving as if one is afraid — is a taboo (Haas 1978). In both cases, emotions are subjected to normative appraisal, concerning not only their outward expression, but also their private experience. What about the criteria “agency”, “controllability”, and “certainty”? The agency criterion concerns the question who or what caused the emotion-eliciting event. It seems plausible to assume that emotions are also evaluated in this respect. A person who watches a horror film is normally cognizant that doing so will cause emotions such as fear, horror, or disgust in her, and that she is responsible for creating, or subjecting herself to, a situation that is likely to evoke such emotional experiences. We are not aware, however, of any empirical studies concerning people’s sense of causal agency in emotion elicitation. There is ample evidence, however, that people appraise emotions with respect to their controllability. A great number of conscious and unconscious strategies for coping with emotion have been studied in this regard, e.g. by Richard Lazarus.17 Finally, emotions would seem to be appraised also with respect to the subject’s certainty about the situation. Sometimes we are unclear, initially at least, about our feelings. Are we tired or depressed, angry or jealous, driven by admiration or infatuation? John Mayer and coworkers developed a “meta-mood-scale” that measures, among other things, the extent to which people think they are clear about their feelings. Participants could report on this scale, in reliable patterns, how certain they were about their feelings (Mayer & Gaschke 1988). We have argued that each of the appraisal criteria which are widely thought to produce emotions when applied to non-emotional events and situations, can be applied to emotions themselves. We cannot provide conclusive arguments for this thesis. Yet there is considerable empirical evidence that the criteria in question are applied to emotions, at least under certain experimental conditions.18 Irrespective of the specific set of criteria discussed above, it seems justified in light of what has been said 17. See, for example, Lazarus (1991). Lazarus himself, however, does not consider the person’s sense of control over his or her emotions as a second-order criterion of appraisal, but subsumes it under the first-order appraisal of control over the emotioneliciting situation. 18. A more comprehensive overview of research literature supporting the assumption of second order appraisals in the context of media psychology can be found in Bartsch et al. (2006).

193

to supplement our characterization of meta-emotions with the following thesis about the generation of meta-emotions: Meta-emotions are elicited when a person appraises his or her own emotions in light of emotionally relevant appraisal criteria. 6. Further applications: emotional self-awareness, art and entertainment, survivor guilt Our account of meta-emotions sheds light on a number of perennial questions from the philosophy and psychology of emotions and the philosophy of mind. We already have seen how it bears on the problem of allegedly paradoxical emotions. In the final pages of this essay we want to put some more flesh on our proposal by looking at three further examples: emotional self-awareness and self-regulation, emotions in art and entertainment, and survivor guilt. Self-awareness and self-control of emotions. In seminal work of the psychologist John Mayer and co-workers (cf. Mayer & Gaschke 1988), the concept of meta-emotion is closely related to the notion of “emotional intelligence”.19 Emotional intelligence is construed by these authors as a person’s capacity to experience and express emotions in self-conscious and self-controlled ways. It would be worthwhile to elaborate this idea further and to analyze the exact function of meta-emotions in self-awareness and self-control of emotions. For instance, it seems intuitively plausible that meta-emotions influence people’s willingness and motivation to pay conscious attention to their first-order emotional states. Emotions that are judged to be enjoyable, goal conducive, morally adequate, or otherwise desirable should be more welcome as objects of self-awareness and selfreflection than emotions associated with pain, failure, moral scruples, and so on. In our view, research on emotional repression provides empirical evidence for this role of meta-emotions in emotional self-awareness. As we have discussed in more detail elsewhere (Jäger & Bartsch 2002; 2006), people’s tendency to repress emotions — i.e. to avoid conscious awareness of emotional arousal — depends on the perceived threat that the emotions pose to their self-esteem (cf. Mendolia 1999). 19. Mayer and Gaschke also talk about “meta-moods”; this does not refer to intentional meta-moods however, but to what they call the “meta-experience” of moods.

194

Meta-emotions have also been related to other forms of emotional selfregulation. John Gottman and coworkers (Gottmann, Katz & Hooven 1997) concluded from their research on emotional communication in families that positive meta-emotions help to regulate negative first-order emotions. People with positive meta-emotions about anger and sadness feel rarely out of control with these emotions, compared to people with corresponding negative meta-emotions. Children who grow up in families with a generally positive attitude towards emotions are able to recover more quickly from physiological arousal, are less vulnerable to negative daily moods, and engage in cooperative peer interaction more often than children who grow up in emotion avoiding or emotion dismissing families. If such functions of meta-emotions in self-awareness and self-control of emotions could be further substantiated this would have interesting implications for theories of self-consciousness and personhood.20 Art and Entertainment. Another field where the concept of meta-emotion can be fruitfully applied is psychology of entertainment and art appreciation. Both art and media entertainment often revolve around emotions that are not desired by most people in everyday life. What Mary Beth Oliver (1993) called the “paradox of the enjoyment of sad films” pertains not only to the enjoyment of tear-jerkers, but also to people’s taste for horror films, thrillers, and a variety of other genres renowned for eliciting strong negative emotions in their recipients. Oliver proposed to solve this paradox by invoking meta-emotions. She was able to show that feeling sad is central to what drama-fans appreciate about sad films. Oliver observed that in participants with a preference for dramas the reported degree of sadness while watching sad film segments was proportional to the reported degree of enjoyment. Comparable results were found in a study on horror films that one of us conducted. Participants liked horror films more if they reported that they liked the emotions they experienced during the film.21 Why then do people enjoy and appreciate emotions like sadness, fear, and disgust in the context of media use, but rarely so in everyday life? It can be speculated that media entertainment is designed in such a way 20. Another interesting question in this context is the relation between metaemotions and empathy, and the significance of this relation for altruistic behavior. Empathy has emotional dimensions, and emotions about other people’s emotions may influence one’s intra-personal meta-emotions, and vice versa. For a helpful discussion of the emotional aspects of empathy and their significance for altruistic motivation see Sherman (1998). 21. Results from an unpublished study conducted by Anne Bartsch and Dennis Storch (Bartsch & Storch, in preparation).

195

as to invite positive meta-emotions by offering a wide range of emotionrelated gratifications.22 Entertainment media foster a state of mild and pleasant arousal, or provide viewers with “kicks” of extraordinarily strong emotion. They make emotions intelligible and controllable for the viewer, and offer moral justifications for the experience and expression of emotions. Conventions of typical entertainment genres can to a large extent be understood as implicit “viewing contracts” that make the gratification potential of genre-typical emotions predictable for the viewer. Survivor guilt. Another interesting application of the concept of metaemotion is people’s apparently irrational feeling of severe guilt connected with events for which they bear no, or only little, responsibility. A striking example is (the experience of 23) survivor guilt. Survivor guilt comes in different varieties. According to leading authors in the field, so-called existential survivor guilt is an amalgamation of guilt and related feelings about staying alive in situations where others died, or about being less harmed than other people with whom the subject identifies. As the American psychologist and psychotherapist Aphrodite Matsakis characterizes it, so-called existential survivor guilt is “the guilt you feel because you are alive, healthier, or otherwise better off than someone else” (Matsakis 1999, 40, 69). Content survivor guilt, by contrast, is guilt resulting from something you did to stay safe or alive, including guilt over adopting coping mechanisms such as denial, repression, rationalization, lying, and so forth, or “having certain thoughts and feelings during trauma or stressful ordeal” (Matsakis 1999, 58; cf. also Williams 1987). The boundaries between the two forms are fuzzy, since the fact that a person survived a tragic event may have resulted from something s/he did. But this need not detain us here. To keep matters within reasonable bounds in the present context, we shall focus on paradigm cases of existential survivor guilt. Let’s narrow down the territory further by considering a particular kind of this guilt, namely what may be called existential survivor guilt proper, i.e. guilt over having survived in a situation where others died. This kind of survivor guilt may again be broken down into two forms. (i) In some cases the person who feels guilty played a role in the tragic incident. A good example of this is the case of Wan Dang, one of the pro22. This has been suggested by Bartsch et al. (2006), and Bartsch and Viehoff (2003). 23. We shall henceforth often drop this qualification and use the term “survivor guilt” in a sense that doesn’t imply that the subject actually is guilty of anything.

196

test leaders of the 1989 Tiananmen Square demonstrations for democracy in China. When hundreds of demonstrating students and other Chinese civilians were shot by the Chinese army, Wan Dang felt, as he later confessed, guilty for this incident — despite his being fully aware that “the main responsibility was not ours [but] with the government who actually did the killing” (The Sun, 1998, A3, quoted from Matsakis 1999, 53). In such forms of survivor guilt, which are connected with “role related responsibility guilt”, the extent of the experienced guilt is typically not warranted by the subject’s actual responsibility. Still, such feelings may not be entirely inappropriate. Small as it may have been, the subjects did have their share in the causal chains that led to the tragic events, and selfassessments of the correct degree of responsibility are characteristically difficult in such cases (especially for those who were involved). We shall therefore turn to another type of existential survivor guilt, which appears even more puzzling. (ii) Often survivor guilt is experienced in connection with situations that have nothing whatever to do with one’s own actions, and in which one knows that this is so. This may be called “existential survivor guilt proper without perceived responsibility”, or for short: “survivor guilt without responsibility”. It is especially this phenomenon, we suggest, in which meta-emotions play a crucial role. Consider survivors of natural catastrophes, car or ship accidents, plane crashes, and so forth, who know they couldn’t have done anything about the incident and couldn’t have saved a single life. Often such people still experience survivor guilt. If the intentional objects of these feelings were the events themselves, or the fact that the survivors had been spared, their emotions would appear entirely irrational. But, as has recently been argued by David Velleman (2004), such a diagnosis would leave something out. Feelings of guilt without perceived wrongdoing, Velleman insists, can be rational and warranted. There is much to be said for this diagnosis. Yet Velleman’s explanation, while it points in the right direction, may be found less convincing. According to Velleman, the fact that you haven’t wronged anyone doesn’t imply that you have no grounds for feeling guilty. Instead, “it may only show that we need to interpret our feelings [of guilt] more carefully, as anxiety about warranting envious resentment” (Velleman 2004, 247f.). There can be grounds for envying those who are more fortunate, Velleman thinks, “or for resenting those whose good fortune is undeserved; and so … a beneficiary of good fortune may rationally feel anxiety about providing others with ground for resentment” (ibid.). Now, the envious

197

resentment that survivors fear obviously cannot be felt by those who were (maybe literally) in the same boat, but did not survive. Velleman thus argues that third parties can feel “resentment on behalf of the deceased. … And a survivor can rationally feel anxiety about providing grounds for such vicarious or sympathetic resentment” (ibid.). However, can this form of anxiety really be construed as (an experience of ) guilt? Even if it can, would not the subject have to believe that s/he provides good grounds for resentment? And why should this be so? An explanation in terms of metaemotions, we believe, provides a more plausible and simpler explanation of survivor-guilt-without-responsibility. In Velleman’s account, too, this peculiar kind of guilt is explained as an emotion about emotions. But Velleman construes the phenomenon as an interpersonal form of higher-level emotion. The truth, we contend, is that survivor guilt is an intrapersonal meta-emotion. Survivors experience positive emotions about their fortune (such as gratitude and relief ). But at the same time they feel guilty about having these positive emotions. In light of the tragic losses the event caused for others, the survivors feel — consciously or unconsciously — that any positive emotions related to the situation are (normatively) inappropriate. According to this view, survivor guilt is not a special form of anxiety, but genuine guilt. However, it is not (the experience of ) guilt over some “wrongdoing” in the ordinary sense, but guilt over being swamped with positive feelings despite the fact that grief and sorrow alone are perceived to be appropriate. This explanation fits nicely into the picture psychologists and psychotherapists draw of the phenomenon. For example, Matsakis explicitly points out, on the basis of extensive experience with post-traumatic clients, that “survivor guilt does involve gratitude” (1999, 50). She reports the story of Paul and Karen whose car was struck by a truck when they were driving home one night. Karen was severely injured; Paul was not. Paul wondered why he was spared injury while his wife had to endure numerous surgeries and much pain. ‘It should have been me. And if it couldn’t have been me, why couldn’t the injuries at least have been divided equally between us?’ Paul asked others. Paul meant what he said. … On the other hand he was glad he had been spared Karen’s ordeal. … Paul felt guilty about being grateful he wasn’t hurt, just as guilty as he felt about not having been injured. (Matsakis 1999, 50)

This is not an instance of “existential survivor guilt proper”, as we dubbed it (for Karen survived); but naturally people also feel glad for having been

198

spared from death, instead of feeling gratitude for not suffering from some (non-lethal) injury. The psychological dispositions of survivors may also elicit meta-emotions about positive experiences and emotions that are not directly related to the survival. As in the case of Mitchell: “The tornado destroyed almost every home in the neighborhood except mine”, he reported. “Six people died. The people next door lost their baby and everything they owned. The night after that tornado hit our town, my wife and I had the best sex we’d had in years. Is there something wrong with me? Why am I glad I’m okay when others around me are hurting?” (Matsakis 1999, 39). Here Mitchell doesn’t talk about guilt he feels for being glad that he has been spared, but for enjoying pleasures and positive emotions while others he feels close to are thrown into extremely desperate and painful situations. Speaking more generally, survivor guilt may involve the feeling that “you should not go on with life, or at least not enjoy success or happiness” (Matsakis 1999, 43). What started as a negative meta-emotion towards positive emotions directly experienced in connection with one’s survival (like gratitude and relief ) may in such cases develop into more general negative meta-emotions towards enjoying positive emotions at all. However, the crucial point in all these cases, and those discussed above, is that we are clearly dealing with emotions whose intentional objects are other emotions. Let us sum up. We have presented a number of phenomenological and ordinary language considerations that suggest the existence of meta-emotions. We have looked into the conceptual nature of the latter, and outlined the basic structure of ascriptions of meta-emotions. Next, we have squared our thesis with psychological research on emotion and, drawing on specific empirical findings, formulated a general genetic hypothesis about the conditions under which meta-emotions arise. In this context, our focus was on appraisal theories. How, it may be asked, can meta-emotions be treated within other approaches? An answer to this question must wait.24 But, in light of what has been said in this essay, the following conclusion appears justified: Whatever format an acceptable theory of emotions adopts, it should be equipped to deal with meta-emotions. This is a constraint for any theory. This claim gains further support because, apart from the suggested treatment of certain ambivalent emotions, the applications discussed in the 24. Bartsch (forthcoming) examines how meta-emotions might be treated in neurological, script-theoretical, and social constructivist accounts of emotion.

199

present section show that meta-emotions are not merely a philosophical construction. We are dealing with reality here.25

REFERENCES Alston, W. P. 1967: “Emotion and Feeling”, in: P. Edwards (ed.), The Encyclopedia of Philosophy, Vol. 2, New York/London. Aristotele, The Art of Rhetoric: Greek/English, London, Cambridge, Mass.: William Heinemann LTD, Harvard UP, 1967. Averill, J. R., 1991: “Emotions as Episodic Dispositions, Cognitive Schemas, and Transitory Social Roles: Steps Toward an Integrated Theory of Emotion”, in D. Ozer, J. M. Healy & A. J. Stewart (eds.), Perspectives on Personality, Vol. 3a, London: Kingsley, 139–176. Bartsch, A. (forthcoming): “Meta-Emotionen und ihre Vermittlung im Film”, in J. Eder, K. Fahlenbrach, & A. Bartsch (eds.), Audiovisuelle Emotionen, Köln: Halem, 2007. Bartsch, A., & R. Viehoff, 2003: “Meta-emotion: In Search of a Meta-account for Entertainment by Negative Emotions”, Siegener Periodikum zur Internationalen Empirischen Literaturwissenschaft 22, 309–328. Bartsch, A., R. Mangold, R. Viehoff & P. Vorderer (2006): “Emotional Gratifications During Media Use — An Integrative Approach”, Communications 31, 261–278. Bartsch, A., & D. Storch (in preparation): “Meta-emotions During Film and Music Video Reception. Development of a Scale for Measuring Affective Appraisal of Emotions as a State Variable”. Ben-Ze’ev, A., 2004: “Emotions as a Subtle Mental Mode”, in Solomon (2004), 250–268. Betzler, M., 2003: “Vernunft und Leidenschaft”, in Stephan & Walter (2004), 262–286. Brentano, F., 1955: Psychologie vom empirischen Standpunkt, Hamburg 1955. Brosschot, J. F & E. Janssen, 1998: “Continuous Monitoring of Affective-autonomic Response Dissociation in Repressors During Negative Emotional Stimulation”, Personality and Individual Differences 25, 69–84. Buck, R., 1984: The Communication of Emotion. New York: Guilford Press. 25. For valuable comments on earlier versions of this paper we wish to thank Nina P. Azari, Christian Beyer, Dieter Birnbacher, Carl Plantinga, an anonymous referee for the Grazer Philosophische Studien, and especially Darrell Rowbottom.

200

Damasio, A. R., 1994: Descartes’s Error — Emotion, Reason, and the Human Brain, New York: Avon Books. Deigh, J., 1994: “Cognitivism in the Theory of Emotions”, Ethics 104, 824–854. DeSousa, R., 1987: The Rationality of Emotion, Cambridge, Mass: MIT Press. Döring, S. A., & V. Mayer (eds.), 2002: Die Moralität der Gefühle, Berlin: Akademie Verlag. Ekman, P., 1972: Emotions in the Human Face, New York: Pergamon. Ekman, P. & R. J. Davidson (eds.), 1994: The Nature of Emotion, Oxford: Oxford University Press. Ekman, P. & W. V. Friesen, 1969: “The Repertoire of Nonverbal Behavior: Categories, Origins, Usage, and Coding”, Semiotica 1, 49–98. Ellsworth, P. C., 1994: “Levels of Thought and Levels of Emotion”, in Ekman & Davidson (1994), 192–196. Frijda, N. H., 1986: The Emotions, Cambridge: Cambridge University Press. — 1994: “Emotions Require Cognitions, Even If Simple Ones”, in Ekman & Davidson (1994), 197–202. Frijda, N. H. & M. Zeelenberg, 2001: “Appraisal: What is the Dependent?”, in Scherer, Schorr & Johnstone (2001), 141–157. Goldie, P., 2000: The Emotions — A Philosophical Exploration, Oxford: Clarendon Press. Gordon, R. M., 1987: The Structure of Emotions, Cambridge et al.: Cambridge University Press. Gottman, J. M., L. F. Katz, & C. Hooven, 1997: Meta-emotion: How Families Communicate Emotionally, Mahwah, NJ: Erlbaum. Green, O. H., 1992: The Emotions. A Philosophical Theory, Dordrecht: Kluwer Academic Publishers. Greenspan, P., 1980: “The Case of Mixed Feelings: Ambivalence and the Logic of Emotion”, in Amélie Oksenberg Rorty (ed.), Explaining Emotions, University of California Press, 1980, 223–250. Haas, J., 1978: “Learning Real Feelings: A Study of High Steel Ironworker’s Reactions to Fear and Danger”, in J. Haas & W. Saffir (eds.), Shaping Identity in Canadian Society, Scarborough, Ontario: Prentice Hall, 227–245. Hatzimoysis, A. (eds.), 2003: Philosophy and the Emotions (Royal Institute of Philosophy Supplements), Cambridge: Cambridge University Press. Hochschild, A. R., 1983: The Managed Heart: Commercialization of Human Feeling, Berkeley: University of California Press. James, W., 1884: “What is an Emotion?”, Mind 19, 188–204. Jäger, C., 1999: Selbstreferenz und Selbstbewußtsein. Paderborn: mentis Verlag. — 2002: “Introspektion und emotionale Abwehr”, Erwägen, Wissen, Ethik 4 (2002), 542–544.

201

Jäger, C., & A. Bartsch, 2002: “Privileged Access and Repression”, in Döring & Mayer (2002), 59–80. — 2006: “Emotionale Verdrängung und Autorität der Ersten Person”, Philosophie der Psychologie 4 (Internet publication), 1–23. Kenny, A., 1963: Action, Emotion and Will, London: Routledge & Kegan Paul. Kleinginna, P.R., & A.M. Kleinginna, 1981: “A Categorized List of Emotion Definitions, with Suggestions for a Consensual Definition”, Motivation and Emotion 5 (1981), 345–379. Knobloch-Westerwick, S. & S. Alter, 2006: “Mood Adjustment to Social Situations Through Mass Media Use: How Men Ruminate and Women Dissipate Angry Moods”, Human Communication Research 32, 58–73. Lazarus, R. S., 1991: Emotion and Adaptation, Oxford: Oxford University Press. — 1994: “Appraisal: The Long and the Short of It”, in Ekman & Davidson (1994), 208–215. — 2001: “Relational Meaning and Discrete Emotions”, in Scherer, Schorr & Johnstone (2001), 37–67. Lenzen, W., 2004: “Grundzüge eine philosophischen Theorie der Gefühle”, in K. Herding & B. Stumpfhaus (eds.), Pathos, Affekt, Gefühl. Die Emotionen in den Künsten, Berlin, New York: Walter de Gruyter, 80–103. Lyons, W., 1980: Emotion, Cambridge, Cambridge University Press. Matsakis, A., 1999: Survivor Guilt, Oakland, CA: New Harbinger Publications, Inc. Mayer, J. D., & Y. N. Gaschke, 1988: “The Experience and Meta-experience of Mood”, Journal of Personality and Social Psychology 55, 105–111. Mayer, J. D., & A. A. Stevens, 1994: “An Emerging Understanding of the Reflective (Meta-)Experience of Mood”, Journal of Research in Personality 28, 351–373. Mendolia, M., 1999: “Repressor’s Appraisals of Emotional Stimuli in Threatening and Nonthreatening Positive Emotional Contexts”, Journal of Research in Personality 33, 1–26. Nussbaum, M., 1990: Love’s Knowledge, Oxford, New York: Oxford University Press. — 2004: “Emotions as Judgments of Value and Importance”, in Solomon (2004), 183–199. Oliver, M. B., 1993: “Exploring the Paradox of the Enjoyment of Sad Films”, Human Communication Research 19, 315–342. Oliver, M. B., J. B. Weaver, & S. L. Sargent, 2000: “An Examination of Factors Related to Sex Differences in Enjoyment of Sad Film”, Journal of Broadcasting and Electronic Media 44, 28. Planalp, S., 1999: Communicating Emotion: Social, Moral, and Cultural Processes, Cambridge: Cambridge University Press.

202

Prinz, J., 2003: “Emotion, Psychosemantics, and Embodied Appraisals”, in Hatzimoysis (2003), 69–86. — 2004: “Embodied Emotions”, in Solomon (2004), 44–58. Robinson, J., 2004: “Emotion: Biological Fact or Social Construction?”, in Solomon (2004), 28–43. Roseman, I. J., 2001: “A Model of Appraisal in the Emotion System: Integrating Theory, Research and Applications”, in Scherer, Schorr & Johnstone (2001), 68–91. Rosenthal, D. M., 1998: “Consciousness and Its Expression”, in P. A. French & H. K. Wettstein (eds.), The Philosophy of Emotions (Midwest Studies in Philosophy XXII ), Notre Dame: University of Notre Dame Press, 1998, 294–309. Salovey, P., L. R. Stroud, A. Woolery, & E. S. Epel, 2002: “Perceived Emotional Intelligence, Stress Reactivity, and Symptom Reports: Further Explorations Using the Trait Meta-Mood Scale”, Psychology and Health 17, 611–627. Schachter, S., and J. P. Singer, 1962: “Cognitive, Social, and Physiological Determinants of Emotional States”, Psychological Review 69, 379–399. Scherer, K. R., 1984: “On the Nature and Function of Emotion: A Component Process Approach”, in K.R. Scherer & P. Ekman (eds.) Approaches to Emotion, New York: Wiley, 293–318. — 2001: “Appraisal Considered as a Process of Multilevel Sequential Checking”, in Scherer, Schorr & Johnstone (2001), 92–120. Scherer, K. R., A. Schorr & T. Johnstone (eds.), 2001: Appraisal Processes in Emotion, Oxford, New York: Oxford University Press. Schwartz, G. E., 1990: “Psychobiology of Repression and Health: A Systems Approach”, in J. Singer (ed.) Repression and Dissociation, Chicago: University of Chicago Press, 405–434. Shaver, P., J. Schwartz, D. Kirson, & C. O’Connor, C., 1987: “Emotion Knowledge: Further Exploration of a Prototype Approach”, Journal of Personality and Social Psychology 52, 1061–1086. Shaver, P., S. Wu, & J.C. Schwartz, 1992: “Cross-cultural Similarities and Differences in Emotion and Its Representation: A Prototype Approach”, in M.S. Clark (ed.), Emotion, Newbury Park, CA: Sage, 175–212. Sherman, N., 1998: “Empathy and Imagination”, in P. A. French & H. K. Wettstein (eds.), The Philosophy of Emotions (Midwest Studies in Philosophy XXII), Notre Dame: University of Notre Dame Press, 1998, 82–119. Solomon, R. C., 1993: The Passions: Emotions and the Meaning of Life, Hacket Publishing. — 1988: “On Emotions as Judgments”, American Philosophical Quarterly 25, reprinted in id., Not Passion’s Slave: Emotions and Choice, Oxford, New York: Oxford University Press, 92–113.

203

— 2003: “Against Valence (‘Positive and Negative Emotions’)”, in id., Not Passion’s Slave: Emotions and Choice, Oxford, New York: Oxford University Press, 162–177. Solomon, R.C. (ed.), 2004: Thinking about Feeling: Contemporary Philosophers on Emotions, Oxford: Oxford University Press. Spinoza, Ethics: Ethica — Ordine Geometrico demonstrata, Latin/German, Stuttgart: Reclam, 1977. Stephan, A., 2004: “Zur Natur künstlicher Gefühle”, in Stephan and Walter (2004), 309–324. Stephan, A., & H. Walter (eds.), 2004: Natur und Theorie der Emotion, Paderborn: mentis, first ed. 2003. Velleman, J. D., 2003: “Don’t Worry, Feel Guilty”, in Hatzimoysis (2003), 235– 248. Voss, C., 2004: Narrative Emotionen, Berlin, New York: Walter de Gruyter. Weinberger, D. A., 1990: “The Construct Validity of the Repressive Coping Style”, in J. Singer (ed.), Repression and Dissociation, Chicago: University of Chicago Press, 337–386. Williams, T., 1987: “Diagnosis and Treatment of Survivor Guilt”, in Posttraumatic Stress Disorders: A Handbook for Clinicians, ed. T. Williams, Cincinnati, Ohio: Disabled American Veterans. Zaleski, Z., 1984: “Sensation Seeking and Preference for Emotional Visual Stimuli”, Personality and Individual Differences 5, 609–611. Zillmann, D., 2000: “Mood Management in the Context of Selective Exposure Theory”, in M. F. Roloff (ed.), Communication Yearbook 23, Thousand Oaks, CA: Sage, 103–123. Zuckerman, M., 1979: Sensation Seeking: Beyond the Optimal Level of Arousal, New York: Wiley. — 1994: “Behavioral Expressions and Biosocial Bases of Sensation Seeking”, New York: Cambridge University Press.

204

Grazer Philosophische Studien 73 (2006), 205–219.

THE PARADOX OF NONBEING 1 William H. HANSON University of Minnesota

In their admirably clear and very useful First-Order Modal Logic, Melvin Fitting and Richard L. Mendelsohn present thorough semantic and proof-theoretic treatments of several languages in which many philosophical puzzles of ongoing interest can be stated with clarity and precision.2 It is thus surprising and somewhat disappointing that their discussion of what they call the Paradox of NonBeing fails to make use of the formal tools that they have developed with such care. I shall attempt to remedy this situation by using Fitting and Mendelsohn’s own formalism to clarify the matters they discuss. I hope thereby to make at least some progress in understanding certain longstanding philosophsical problems. Specifically, I formalize their arguments and use their semantics to show that: (a) Actualism is a consistent position, despite their claims to the contrary; (b) The position they call ‘Deflationism’ is a caricature of actualism and is a straw man; (c) Their semantic framework contains the means for devising a plausible actualistic resolution of the Paradox of NonBeing. Central to Fitting and Mendelsohn’s presentation of the Paradox of NonBeing3 is the following argument (168–169): (A) If an individual (call him “John”) denies the existence of something, then John refers to what he says does not exist; (B) Things which do not exist cannot be referred to or mentioned; no statement can be about them. the following conclusion is drawn: 1. I wish to thank Alana Yu and an anonymous referee for Grazer Philosohpische Studien for helpful comments on earlier versions of this paper. 2. All page references are to Fitting and Mendelsohn (1998), unless otherwise indicated. 3. ‘Paradox of non-existence’ would have been a better term, since the authors distinguish existence from being, the paradox they discuss concerns existence rather than being, and they recognize (175) that being has its own distinct, though analogous, paradox. (I plan to discuss this latter paradox, a genuine paradox of non-being, in another paper.) To maintain continuity with their presentation, however, I shall use their term throughout this paper.

(C) If John denies the existence of something, then what John says does not exist does exist. Fitting and Mendelsohn say that this argument is valid but that its conclusion is paradoxical and must be rejected. And they think there is a genuine puzzle involved in seeing how to reject at least one of the premises in order to avoid being committed to the conclusion. The difficulty involved in solving this puzzle, in coming up with a plausible rationale for rejecting either (A) or (B), is the core of their Paradox of NonBeing. They follow Berlin (1949–50) in calling those who accept (A) and deny (B) Inflationists, and those who deny (A) and accept (B) Deflationists. The authors devote sections 8.3–8.6 of their book (167–178) to discussion of the Paradox and related matters. They themselves are Inflationsts, and in the sections just mentioned they strive to show the advantages of Inflationism over Deflationism. Unfortunately their discussion, couched as it is in ordinary English, is not as precise as it might be, and it is marred by their apparent failure to recognize that Inflationists and Deflationsts (who are species of possibilists and actualists, respectively) speak different languages. Yet these problems can be avoided, and the underlying issues significantly clarified, when the argument on which the Paradox is based is stated in the language of first-order modal logic with identity that the authors have so painstakingly developed. Consider the following abbreviations of the key terms in (A)–(C): Ex: Rxy: Dxy: Sxy: j:

x exists x refers to y x denies the existence of y x says of y that it does not exist John

Using these symbols, a first attempt at formalizing the authors’ argument might be: x(Djx Š (Rjx š Sjx)) ™¡x(™Ex š yRyx)4 x(Djx Š (Sjx š Ex)) 4. I symbolize only the first part of (B), but this will not affect the subsequent discussion. Notice that symbolizing the first part of (B) in what is perhaps the most straightforward way, namely as x(™Ex Š ™¡yRyx), stacks the deck against Deflationists. For this formula, claiming as it does that a thing which doesn’t exist in the actual world can’t be referred to in any possible world, is intuitively less plausible than the one given in the text.

206

I don’t think this formalization is quite right, however, because the first premise and the conclusion are non-modal formulas. Although (A) and (C) are non-modal sentences of English, I suspect the authors’ intent is better captured by symbolizing them as the necessitations of the formulas given above. (Otherwise they make merely contingent claims, which is not what one expects from metaphysicians with the authors’ views.) When this is done, and when the second premise is presented in an equivalent form, the argument becomes: (A*)  x(Djx Š (Rjx š Sjx)) (B*)  x(™Ex Š ™yRyx) (C*) x(Djx Š (Sjx š Ex)) It is easy to see that (C*) follows from (A*) and (B*) in any standard quantified modal logic based on a normal modal propositional logic.5 Although the authors’ argument is clearly valid, they argue for its validity as follows (169): This argument is logically straightforward. It requires little more than transitivity of implication. We need only to rephrase premise (B) to read (D) If John refers to or mentions something, Ks, then Ks exist. Then, as an instance of (D), taking “Ks” to be “what John says does not exist,” we have (E) If John refers to or mentions what he says does not exist, then what John says does not exist does exist. And, from (A) and (E), by transitivity, we reach the desired conclusion (C). Since the authors are working in English, perhaps some clarity is gained by this maneuver. And when (D) and (E) are formalized as follows6, the inferences the authors endorse are indeed correct. 5. A normal propositional modal logic is one that contains the system K, which is the deductive closure of all truth-table tautologies and the axiom (p Š q) Š ( p Š q), under modus ponens, necessitation, and uniform substitution. Because of the authors’ special treatment of individual constants, (A*) and (C*) are not well formed in their systems. Yet their systems validate the argument from (A*) and (B*) to (C*) after we have universally generalized on ‘j’ within the scope of the modal connective in (A*), and likewise in (C*). Specifically, the formula (C**) ( xy(Dyx Š (Syx š Ex))) is a consequence of (B*) and (A**) ( xy(Dyx Š (Ryx š Syx))), when (A**) and (B*) are taken as local assumptions in the authors’ varying domain and constant domain first-order extensions of K. Equivalently, the corresponding conditional of this argument is valid in all varying domain models (and hence valid in all constant domain models). See 21, 95–99, 101–105. 6. Again I take the authors assertions to be in the apodeictic mode.

207

(D*) x(Rjx Š Ex) (E*) x((Rjx š Sjx) Š Ex) (E*) does follow from (D*) (which in turn follows from (B*)), and (A*) and (E*) together yield (C*), in any standard quantified modal logic based on a normal modal propositional logic, including the authors’ constant domain and varying domain systems.7 Having presented the argument from (A) and (B) to (C), the authors use it, and further elaborations of it, in an attempt to show that the position of Deflationists, whom they call “the dominant group in twentieth century Anglo-American philosophy” (169), is inconsistent. There are, however, serious problems with their claims and arguments. I shall focus on these in three steps. First, their criticism of the Deflationist position contains several errors. In exposing these errors I show that the arguments which they claim establish the inconsistency of Deflationism fail. Second, the Deflationist is a straw man. It is clear that in criticizing Deflationism, the authors have in mind actualism. Actualists hold that only actual things exist or have being in any sense, and they probably are “the dominant group in twentieth century Anglo-American philosophy”, though they needn’t be Deflationists in the authors’ sense of that term. The authors’ Inflationism, on the other hand, is a species of possibilism. The authors believe in non-actual objects, they distinguish between existence and being, and, while they attribute being to all the objects of their ontology, they attribute existence only to actual objects (175–178). Both actualism and the authors’ version of possibilism are consistent positions, but each must be understood in terms of its own principles. To show this I shall present two models drawn from the authors’ own semantics. I first give a constant domain model which embodies the possibilist principles of the authors’ Inflationism, since it makes (A*) true while making (B*) and (C*) false. Next I give a varying domain model that embodies actualist principles and in which (A*), (B*), and (C*) are all true, thus showing that actualists need not be Deflationists. The fact that these two models assign different ranges to the quantifiers and different extensions to predicates shows that the languages of actualists and Inflationists are significantly different and that these differences must be taken seriously in understanding the two positions. Third, I show that the two models can be extended to deal with individual terms (even non-designating terms) in ways that preserve the consistency of both actualism and Inflationism, and I argue for the plausibility of the extended model of actualism. 7. More precisely, in the authors’ systems it is (D**) ( xy(Ryx Š Ex)) and (E**) ( xy((Ryx š Syx) Š Ex)), which stand in the logical relations mentioned in the text. The situation is thus analogous to the one described in footnote 5.

208

1. The inconsistency of Deflationism has not been proved Consider first the authors’ claim that Deflationism is inconsistent. They begin their criticism by attacking (B) which, by definition, Deflationists accept. Specifically, they tell us (173): (B) says that the expression “things which do not exist” cannot refer to things which do not exist. Yet (B) does succeed, via this expression, in indicating what it is that cannot be referred to. The first sentence is false, since (B) says nothing about any linguistic expression. And the second sentence is doubtful, since in (B*) the quoted expression is rendered by the antecedent of a universally quantified conditional. The Deflationist will say that since there are not and could not be things which do not exist, the quantified conditional is vacuously true in every possible world, and its necessitation, (B*), is therefore true. Nevertheless the authors claim to have shown that (B) is “self-defeating” and “ill-constructed” (173). Having rejected (B) for the dubious reasons just discussed, the authors say that Deflationists should replace it with (D), which they call “the conditional form of premise (B)” (173). On their analysis, however, (D) turns out to be even worse than (B). Here is their argument for this conclusion (174). Consider the following instantiation of (D): (F) If John refers to or mentions apples, then apples exist. Among the statements governed by the condition laid down in (F) is (F) itself. For, if (F) is true, then since in the very claiming of (F), apples are referred to, apples exist. In other words, (F) is self-referential (a property it shares with (B)). Self-reference is not of itself pernicious. Although (F) has no untoward consequences, another instantiation of (D), namely, (G) If John refers to or mentions things which do not exist, then things which do not exist do exist. has the contradictory consequence: Things which do not exist do exist, since (G) itself makes reference to things which do not exist, (D), then, fares even worse than (B), for whereas (B) was merely self-defeating, (D) implies a patent contradiction. The last move in this argument is clearly wrong. For if

209

Things which do not exist do exist is contradictory, it follows only that (G), which appears to be a universally quantified conditional, has a contradictory consequent, not a contradictory consequence. But the authors may have something more subtle in mind. Perhaps their idea is that the antecedent of (G) is the offending party, for it says that John refers to things which do not exist, and this may be thought an impossible task. So perhaps this is the basis of their claim that “(G) itself makes reference to things which do not exist”. If so, they are again mistaken. Just because the antecedent of (G) says that John refers to things which do not exist, it does not follow that one who asserts (G) does so. A quantified conditional with an antecedent that makes an impossible claim is not itself stained by the sin of that claim, but merely vacuously true (and necessary). It appears that the mistake the authors made in dealing with (B), noted above, is repeated here with respect to (G). Once again matters will be clarified if the crucial sentences are properly formalized. If we add Ax: x is an apple, to our list of abbreviations of key terms, we can formalize (and modalize) (F) and (G) as follows: (F*)  x((Rjx š Ax) Š Ex) (G*) x((Rjx š ™Ex) Š Ex). Thus formalized, it is easy to see that (F*) is a simple consequence of (D*) and that (G*) is logically equivalent to (D*). And we’ve already seen that (D*) is a consequence of (B*). Furthermore, (B*), (D*), (F*), and (G*) are perfectly consistent, both individually and collectively, as I will presently show.

2. Models for Inflationism and actualism; the Deflationist is a straw man As I mentioned earlier, the authors’ Inflationism is a species of possibilism. Like many other possibilists they distinguish those objects that exist from those that have being, the former constituting a proper subclass of the latter. In formal semantics they let their quantifiers range over the latter class, and they use an existence predicate to distinguish, at each possible world, the things that exist at that world from those that do not. The actual world, the one we inhabit, is just one of many worlds, and the actual objects are those that exist here. But there

210

are many merely possible, non-actual objects (Pegasus, perhaps, and the golden mountain) within the range of these possibilists’ quantifiers, and that range is unvarying from world to world. 8 While Inflationism, so described, is an ontologically rich theory, it has a close cousin that is richer still. For while the authors include merely possible objects in their ontology, they draw the line at impossible objects, expressly excluding Meinong’s round square and similar impossibilia from the realm of being (177). So Inflationism, a genuine kind of possibilism, is to be distinguished from a stronger Meinongian theory, which might be called ‘Meinongian possibilism/ impossibilism’. The authors’ Inflationism is also distinct from the kind of possibilism advocated by David Lewis (1986). Lewis postulates many different possible worlds, each inhabited by completely different objects, and he rejects the distinction between existence and being. He also rejects world-relative predication, and he espouses an indexical theory of actuality. I discuss neither Meinongian nor Lewisian theories in this paper. From now on the only kind of possibilism I shall be concerned with is Inflationism, as characterized above. Inflationism is best formalized using the authors’ constant domain semantics.9 A constant domain first-order normal model is a structure M = ¢G, R, D, I ², where G is a non-empty set (of worlds), R is a relation on G (the relative possibility relation), D is a non-empty set (of individuals), and I is an interpretation function (assigning values to non-logical constants of the language) (95–96). The model is constant domain because there is just one domain of individuals, D, for the entire model. Quantifiers range over the same set of individuals at each world. It is normal 10 because I assigns the identity relation on D to the nonlogical constant ‘=’ (141). Given this formal machinery, Inflationism can be understood as commitment to a preferred or intended constant domain model MP = ¢GP, RP, DP, IP², where GP really is the set of all possible worlds, RP is the “real” relation of relative possibility among possible worlds, DP is the set of all actual and non-actual individuals11, and, for an interpreted first-order language, IP assigns the correct extensions, at worlds, to non-logical constants of the language. For example, given our interpretations of the predicates specified above, IP assigns to ‘A’, at 8. These views make the authors possibilists of the first kind on the classification given by Linsky and Zalta (1994: 435–436). 9. Indeed the authors express a preference for constant domain semantics (275). 10. This use of the term ‘normal’ to characterize first-order modal models should not be confused with its use in ‘normal modal propositional logic’. See footnote 5. 11. It is not implausible to think that Inflationism is committed to there being more possible worlds, more actual and non-actual individuals, etc., than can be accommodated in any set. Although this is an issue that Inflationists should address, it is irrelevant to the matters discussed in this paper.

211

each world w, the set of things that are apples in w, to ‘R’ the set of ordered pairs of things such that the first member of each pair refers to the second member of that pair in w, etc. Since MP is a constant domain model, quantifiers always range over all actual and non-actual things, no matter the world at which they are evaluated. In the Inflationist terminology, quantifiers range over everything that has being. But Inflationists also hold that not every object that is (i.e., that has being) is actual at each possible world, and they say that the objects that are actual at world w exist in w. So IP assigns to ‘E’, at each world w, the set of things that exist in w.12 Given the foregoing, it is not difficult to see that (A*) is true at the actual world — and indeed at all worlds — of MP13. For Inflationists hold that we are able to refer to all things that have being, even things that do not actually exist, such as Pegasus. We can thus truly deny the existence of such things, truly say of them that they do not exist. So if John denies the existence of something, actual or non-actual, Inflationists will claim that he thereby refers to it and says of it that it doesn’t exist, and MP embodies this view. These same considerations show that the remaining sentences (B*) through (G*) are false throughout MP. Actualists, on the other hand, will want to express their position by means of varying domain semantics (101–103). A varying domain first-order normal model is a structure M = ¢G, R, D, I ², where G and R are as in constant domain models. Now, however, different members of G can have different domains of individuals associated with them. This is accomplished by making D not a set of individuals, but rather a function that takes each w  G to a non-empty set, D(w), the set of individuals that exist in w. Thus different members of G can be thought of as having different individuals existing in them. The domain of the model, D(M )14, is the union of the domains of all the members of G. I is an interpretation function, as before, assigning values to non-logical constants of the language, these values being set-theoretic constructs of members of D(M ). Quantifiers are now world-relative. When a sentence is evaluated at a world w, 12. When dealing with constant domain semantics, the authors use the symbol ‘İ’ as the primitive existence predicate (106). I shall continue to use ‘E’ in this capacity. 13. Strictly speaking, and as was noted in footnote 5, (A*) is not well formed in the authors’ systems. On their approach individual constants are used only in conjunction with predicate abstraction. We needn’t concern ourselves with these complications at this point, however, for it is easy to show that the result of universally generalizing on ‘j’ within the scope of the modal connective in (A*) (i.e., the formula xy(Dyx Š (Ryx š Syx)), which was called (A**) in footnote 5) is true at each world of MP. My subsequent claims about the truth values of (C*) –(G*) should be understood, similarly, as about (C**)–(G**), from which ‘j’ has been eliminated by a similar generalization. 14. The authors call D(M ) the domain of the frame, the frame of a model ¢G, R, D, I ² being ¢G, R, D².

212

the quantifiers in that sentence are taken to range over D(w), not D(M ). Again, since these models are normal, I assigns the identity relation on D(M ) to the non-logical constant ‘=’. Finally, since different individuals can exist at different worlds, and since quantifiers are world-relative, the existence predicate, ‘E’, is defined as: Ex =df y(y = x)15. Although the world-relative quantifiers of varying domain models reflect the view that only actual things exist, actualists have a problem describing a preferred varying domain model that embodies their position in the way that MP embodies Inflationism. For those elements of the domain of such a model that do not exist at the actual world of that model would seem to be things that, according to actualists, do not exist in any sense. Thus they cannot be referred to in a metalanguage actualists can accept and use. Similarly, it is not clear how actualists are able to make use of possible worlds, since they deny the existence of the elements of GP, the possible worlds recognized by Inflationists. Actualists have tried to avoid these problems by using things that are acceptable in their ontology in place of the objectionable entities of the Inflationist. Some actualists hold that there are possible worlds and possible individuals but construe them as unproblematic actual entities — maximal consistent sets of propositions (or of states of affairs) which describe “a total way things might have been”, and special properties (individual essences), respectively (Plantinga: 1976). Alternatively, actualists may deny that there are possible worlds and possible individuals in any sense and use abstract entities, such as pure sets, to play the roles that possible worlds and possible individuals play in MP (Menzel: 1990, 1991). Either way, actualists substitute actual objects for Inflationists’ merely possible objects. I shall assume that surrogates of this kind are available to actualists in constructing their intended varying domain model. The preferred or intended varying domain model for actualists, then, is MA = ¢GA, RA, DA, IA², where GA is the set of all possible-world surrogates, RA is the “correct” relation of relative possibility among these world surrogates, DA is the function that takes each world-surrogate, w, in GA to the non-empty set DA(w) of individuals and individual-surrogates existing in w, DA(MA), the union of the domains of all the world-surrogates, is the set of all actual individuals and individual surrogates16, and, for an interpreted first-order language, IA assigns 15. When dealing with varying domain semantics, the authors use ‘E’ as the defined existence predicate (164). I shall continue to use ‘E’ in this capacity. 16. Actualists may face cardinality problems in constructing MA similar to those, mentioned

213

the correct extensions, at world-surrogates, to non-logical constants of an interpreted language.17 It is easy to see that all of the formalized sentences (A*) – (G*) are true at the actual world surrogate — and indeed at all world surrogates — of MA.18 Consider first the sentence (A*). At the actual world surrogate of MA quantifiers range over all and only actual things, and (A*) thus says of each such thing that if John denies its existence he both refers to it and says of it that it does not exist. This is unproblematically true. The fact that, according to actualists, John will always be mistaken in denying the existence of an actual thing is irrelevant. Sentence (C*) is true for a similar reason, plus the fact that actualists define ‘Ex’ as ‘y(y = x)’. And this definition of the existence predicate is also responsible for the fact that each of the remaining sentences (B*), (D*) – (G*) is not only true at all the world surrogates of MA but logically true — true at each world of each varying domain model. The foregoing shows that actualists need not be Deflationists. The authors define Deflationists as those who accept (B) but deny (A) in order to avoid accepting (C). Yet (C) is unacceptable only to those who allow quantifiers to range over things that fall outside the extension of the existence predicate. Inflationists do this, but actualists do not. Given the actualist’s understanding of quantifiers and existence, (A), (B), and (C) can all be accepted without any qualms. The authors’ assumption that everyone will want to join them in rejecting (C) is a bit of parochialism on their part. They apparently fail to recognize that actualists and Inflationists speak different languages with different ontological commitments, and they expect actualists, who recognize only things that exist, nevertheless to in footnote 11, that possibilists face in constructing their preferred model, MP. But again, these problems are irrelevant to the matters discussed in this paper. 17. It may be wondered whether actualistic surrogates really are available in sufficient abundance to permit construction of models like MA that are both fully actualistic and adequate for the purposes of modal semantics. This is a disputed matter. A related question, narrower and more precise, can be obtained by focusing on the triples consisting of just the M, R, and D elements of MP and MA. Let MP* be ¢GP, RP, DP ², and let MA* be ¢GA, RA, DA(MA )². The Inflationist can then ask whether MA* is isomorphic to MP*. The actualist will see this as the question whether MA* would be isomorphic to MP*, if only there were such a thing as MP*. Inflationists may claim that the answer is negative and that therefore no actualistic semantics for modal discourse is adequate. Actualists may answer positively or negatively, but deny that a negative answer is fatal for actualistic semantics. Fortunately, I need not address these questions in this paper, since I am focusing mainly on Fitting and Mendelsohn’s claim that actualism is inconsistent. Even if MA* is only an “approximation” of MP*, it is sufficient to show that the authors’ attribution of inconsistency to actualism is mistaken. For a comprehensive discussion of the question of actualistic surrogates for possibilia, and related matters, see Fine (2003). 18. A remark similar to that of footnote 13 applies to evaluating sentences containing individual constants in varying domain semantics.

214

join them in finding (C) unacceptable. Thus they invoke the Deflationist, who is now seen to be a straw man.

3. Adding individual terms to the language At this point the reader may protest that even if the foregoing is correct, it does not solve all the puzzles associated with the authors’ Paradox of NonBeing because there has been no consideration of individual terms like ‘Pegasus’ and ‘the round square’. Ironically, the authors themselves provide a treatment of individual terms that can be used by actualists to extend their theory in a consistent and plausible way, thus undercutting the motivation for Inflationism. The authors’ approach to individual terms is based on their use of predicate abstraction. For any formula ) and variable v, ¢Ov. )² is a predicate abstract which can be applied to any term, t, to create a new formula ¢Ov. )²(t) (196– 197). Only variables, however, can combine directly with primitive predicates to form atomic formulas. Individual constants and other terms combine with atomic predicates only by means of predicate abstraction. Thus, given the following abbreviations: Hx: x is a horse p: Pegasus, we can say that Pegasus is a horse by writing ‘¢Ox. Hx²(p)’, but not by writing ‘Hp’, which the authors consider ill formed. Their treatment of individual constants is nevertheless inclusive and generous. They allow individual constants to designate different individuals at different worlds, to fail to designate at some worlds, and even to designate at a world an individual which does not exist at that world (231). Similarly the extension of a predicate at a world may include individuals which do not exist there (101–104). Thus Inflationists can hold, as do the authors (243), that at the actual world in their intended model MP, the designation of ‘p’ is Pegasus, and the extension of the predicate ‘H’ includes Pegasus, even though Pegasus does not exist at the actual world. ‘¢Ox. Hx²(p)’ is thus true at the actual world of MP. More generally, as long as a term t designates at a world w, a sentence of the form ¢Ov. )²(t) can be either true or false at w. If t does not designate at w, ¢Ov. )²(t) is false at w (232).19 The authors also define three useful predicate abstracts, starting with the abstract ‘D’ for denotation: 19. In this respect the authors’ approach bears some similarity to that of Menzel (1991).

215

D(t) =df ¢Ox. x = x²(t), for any term t. Although ‘D(t)’ doesn’t literally say anything about the term ‘t’, they tell us to “[t]hink of D(t) as intended to assert that t designates”. (If ‘t’ doesn’t designate at a world, no application of a predicate abstract to ‘t’, not even the abstract for self-identity, yields a truth at that world.) They also introduce predicate abstracts for existence, ‘E’, and non-existence, ‘E’, as follows: E(t) =df ¢Ox. E(x)²(t) E(t) =df ¢Ox. ™E(x)²(t), where ‘E’ is the existence predicate.20 It is now easy to show that D(t) { [E(t) › E(t)] is valid in all models, whether constant domain or varying domain.21 It is not difficult to see that this semantic apparatus provides the means by which Inflationists can consistently deny the existence of beings like Pegasus. Using the three abstracts defined above, Inflationists can say (truly, at the actual world of their intended model MP ) that Pegasus has being, lacks existence, and has non-existence, as follows: D(p)

(Pegasus is self identical and thus has being; roughly, ‘Pegasus’ designates.) ™E(p) (Pegasus lacks existence.) E(p) (Pegasus has non-existence.) These sentences are true at the actual world of MP because ‘p’ designates Pegasus at that world even though Pegasus does not exist there. So even when nondesignating proper names are present, the Inflationist has no problem avoiding familiar puzzles, as the authors rightly contend. 20. These definitions and the gloss on ‘D(t)’ are on 233. (See also 220 and 164.) Recall that ‘E’ is primitive under constant domain semantics (106), but that ‘Ex’ is defined as ‘y(y = x)’ under varying domain semantics (164). As noted previously, the authors use the symbols ‘İ’, and ‘E’, respectively, to express these two notions of existence. To maintain continuity with earlier sections of this paper, I use ‘E’ for both, indicating in the text which is intended. 21. The authors prove this for varying domain models (234), and it is not difficult to see that it also holds for constant domain models, ‘Ex’ being understood differently in the two cases, as noted in the previous footnote. Indeed it holds, in both kinds of models, for any predicate abstracts ‘P’ and ‘P’ which are defined like ‘E’ and ‘E’, respectively, but with any predicate ‘P’ playing the role of ‘E’.

216

But the authors’ semantic framework also provides actualists a consistent and plausible way of handling these sentences, although the authors do not acknowledge this. For notice that actualists can also say (truly, at the actual world of their intended model MA) that Pegasus lacks existence, as follows: ™E(p) (Pegasus lacks existence.) This holds because ‘p’ does not designate at the actual world (or indeed at any world) of MA, and so the application of any predicate abstract to ‘p’ yields a falsehood. Of course this also means that both ™E(p) (Pegasus lacks non-existence.) and ™D(p) (Pegasus lacks self identity; roughly, it is not the case that ‘Pegasus’ designates.) are true at the actual world of MA. The feature of the authors’ semantics responsible for the truth of all three of these sentences, namely that no predicate abstract is true of any term at a world at which that term does not designate, may seem counterintuitive. Against this feature it could be objected that at least vacuous predicates, like self-identity, should be truly predicable of all terms, even non-designating terms. The objector might claim that even the actualist, who denies that ‘Pegasus’ designates, should at least hold that ‘¢Ox. x = x²(p)’ is true, and indeed logically true. But I don’t find this intuition compelling. Indeed I think it is wrong. A sentence of the form ‘¢Ox. )²(p)’ can plausibly be interpreted as ‘Pegasus is such that it is )’. But if the name ‘Pegasus’ does not designate, then every such sentence is properly judged to be false, even if ) is self-identity. Since ‘Pegasus’ does not designate, there is no way Pegasus is. Thus I want to defend the authors’ semantic framework, which provides actualists a consistent way to state their views even when non-designating proper names are present. Since each of the three sentences displayed in the preceding paragraph is true at the actual world of MA, actualists do indeed have a way of handling empty names that is consistent, and consistent with their principles.22 22. The treatment of ‘Pegasus’ in this paragraph and the preceding one is taken from the authors’ discussion of a term (‘George Washington’s eldest son’) which, in a temporal model, does not designate at any world (237–238). The individual constant ‘p’, which abbreviates ‘Pegasus’, behaves in an exactly analogous way in the actualist’s intended model MA.

217

It must be admitted, of course, that all three of these sentences will be true at all worlds in MA, and hence necessarily true at the actual world of MA. For even though actualists will recognize flying-horse surrogates at various worlds of MA, they will deny that any of them are Pegasus, in order to avoid well-known difficulties. So, in particular,

™E(p) (Necessarily Pegasus lacks existence.) will be true for the actualist, although it is not true at the actual world in the Inflationist’s MP. Thus, if it is thought that no sentence which uses a proper name in denying existence can be necessarily true, Inflationism will seem to have at least one advantage over actualism. Yet this latter view is not plausible. For consider Ptah, the Egyptian god of creation who allegedly called each thing into being by uttering its name. It is arguable that there could not be such a being, that ‘Ptah’ necessarily does not designate.23 So if the Inflationist’s language contains names of impossible individuals, like ‘Ptah’, a sentence with the same form as ‘™E(p)’ will be necessarily true on his theory.24 Thus if ‘Ptah’ is, as it seems, a genuine proper name, the alleged advantage of Inflationism on this point vanishes. The authors of First-Order Modal Logic claim that actualism is inconsistent. Yet their own semantic framework allows construction of a consistent interpretation of actualism that avoids the puzzles associated with the Paradox of NonBeing. And this result holds for languages containing non-designating individual terms. While there are unresolved disputes between actualists and possibilists,25 I have shown that Fitting and Mendelsohn’s charge that their version of possibilism, Inflationism, is superior to actualism because the latter is inconsistent can safely be rejected. I hope also to have convinced the reader that formalizations of actualism such as that embodied in MA should be taken seriously.

REFERENCES Anderson, C. Anthony. 1984. “General Intensional Logic”. Handbook of Philosophical Logic. D. Gabbay and F. Guenthner, eds. Kluwer. vol. 2: 355–385. 23. Anderson (1984: 357) suggests that ‘Ptah’ is a proper name that could not designate. Using ‘Ptah’, rather than ‘the round square’, allows us to (attempt to) speak of impossible objects without abandoning proper names for definite descriptions. 24. As noted in footnote 22, the authors explicitly accept the tense-logic analogue of this result (237–238). 25. See footnote 17.

218

Berlin, Isaiah. 1949–50. “Logical Translation”. Proceedings of the Aristotelian Society. NS 50: 157–188. Fine, Kit. 2003. “The Problem of Possibilia”. The Oxford Handbook of Metaphysics. Michael J. Loux and Dean W. Zimmerman, eds. Oxford University Press. 161–179. Fitting, Melvin and Richard L. Mendelsohn. 1998. First-Order Modal Logic. Kluwer. Lewis, David. 1986. On the Plurality of Worlds. Blackwell. Menzel, Christopher. 1990. “Actualism, Ontological Commitment, and Possible World Semantics”. Synthese. 85: 355–389. — 1991. “The True Modal Logic”. Journal of Philosophical Logic. 20: 331–374. Linsky, Bernard and Edward N. Zalta. 1994. “In Defense of the Simplest Quantified Modal Logic”. Philosophical Perspectives, 8, Logic and Language. James E. Tomberlin, ed. Ridgeview Pub. Co. 431–458. Plantinga, Alvin. 1976. “Actualism and Possible Worlds”. Theoria. 42: 139–160.

219

Grazer Philosophische Studien 73 (2006), 221–230.

FLEXIBLE PROPERTY DESIGNATORS Dan LÓPEZ DE SA University of St. Andrews

The simple proposal about rigidity for predicates can be stated thus: a predicate is rigid if its canonical nominalization signifies the same property across the different possible worlds. I have tried elsewhere to defend such a proposal from the trivialization problem, according to which any predicate whatsoever would turn out to be rigid. Benjamin Schnieder (2005) aims first to rebut my argument that some canonical nominalizations can be flexible, then to provide five arguments to the effect that they are all rigid, and finally to propose a general explanation of why they are all rigid. I show first why my argument has not been rebutted, then why Schnieder’s five arguments for their rigidity all fail, and finally why the alleged “explanation” cannot be such, as the facts alluded to are neutral with respect to the rigidity or flexibility of the nominalizations.

1. Rigidity for predicates: the simple proposal A singular term is rigid if it signifies1 the same object across the different worlds. What it is for a predicate to be rigid? The simple proposal about rigidity for predicates is the following: a predicate is rigid iff it signifies the same property across the different possible worlds. This was arguably suggested by Kripke (1980) himself, and indeed seems to be tacitly assumed in discussions in philosophy of mind, philosophy of science, or metaethics. In order to allay the misgivings some people have about predicates signifying properties, the proposal can be restated via canonical nominalizations of predicates. If P is a predicate, its canonical nominalization, P-ing, is (roughly) the expression that results from P by replacing the first verb it contains by its gerund form. So ‘being water’, ‘being the substance instances of which fall from the sky in rain and fill the lakes and rivers’, ‘running’ and ‘exercising the way José prefers’ are the canonical nominalizations of the predicates ‘is water’, ‘is the substance instances of which fall from the sky in rain and fill the lakes and 1. Following Gödel, I use ‘signification’ as a word for what in German is called ‘Bedeutung’.

rivers,’ ‘runs’ and ‘exercises the way José prefers.’ A predicate is rigid, the simple proposal has it, iff its canonical nominalization signifies the same property across the different possible worlds. The simple proposal is in effect quite a natural one, as its critics also concede. However, it has been claimed to suffer from what I call the trivialization problem: any predicate whatsoever would turn out to be trivially rigid, according to the proposal. I have argued elsewhere2 that the simple proposal can overcome the trivialization problem, by exploiting intuitions about the actual truth-values of identity statements involving nominalizations. In a recent paper, ‘Property Designators, Predicates, and Rigidity’, Benjamin Schnieder (2005) aims first to rebut my argument that some canonical nominalizations can be flexible, then to provide five arguments to the effect that they are all rigid, and finally to propose a general explanation of why they are all rigid. In what follows I show first why my argument has not been rebutted (section 2), then why Schnieder’s five arguments for their rigidity all fail (section 3), and finally why the alleged “explanation” cannot be such, as the facts alluded to are neutral with respect to the rigidity or flexibility of the nominalizations (section 4).

2. The flexibility of some nominalizations defended According to the defender of the simple proposal, some (canonical) nominalizations can signify different properties in the different worlds. Plausibly, for instance, ‘exercising the way José prefers’ signifies running with respect to the actual world, but swimming with respect to an appropriate counterfactual world. But now consider the (abundant) property exercising-the-way-José-prefers, which is (stipulated to be) had by something in a world iff it is the way of exercising José prefers in that world. Notice that the assumption that ‘exercising the way José prefers’ rigidly signifies this property instead of flexibly signifying the different sports José prefers in the different worlds has the same consequences regarding both the actual and counterfactual truth-values of sentences like ‘Pedro exercises the way José prefers’. Furthermore, one might suggest, if there is such a property as exercising-the-way-José-prefers, isn’t it the obvious candidate for ‘exercising the way José prefers’ to (rigidly) signify? How could the simple proposal and the flexibility of ‘exercising the way José prefers’ then be defended? This challenge is what constitutes the trivialization problem, as the worry would reappear with respect to any putative candidate of a flexible predicate whatsoever: how could its flexibility be defended given that an (abundant) 2. In (López de Sa, 2001, 2003). See also (LaPorte, 2000) and (Salmon, 2003, 2005) for congenial proposals.

222

property, tracking the actual and counterfactual extensions of the predicate, will always be available to be rigidly signified?3 The strategy I advocate for overcoming the trivialization problem exploits intuitions about the actual truth-values of identity statements involving nominalizations. To the extent to which one has intuitions that some such statements are (contingent but) true, one can provide the required reason for defending the flexibility of one of the nominalizations involved. For if both were rigid, the famous Kripkean argument would entail that the statements are necessary if true. Hence, even if the alternative candidate (abundant) property is always available, one can have reasons for holding that the nominalization does not (rigidly) signify it. The following seem to me to be precisely cases at hand, intuitively (merely contingent) true identity statements: Running is exercising the way José prefers. Being water is being the substance instances of which fall from the sky in rain and fill the lakes and rivers. Having the color Sònia likes best is being blue. According to Schnieder, these are intuitively correct claims to make — and in effect contingent ones. What he challenges, however, is that they are identity statements. He offers some nice examples of statements of the same form that are naturally understood as conveying (in the appropriate contexts) relations 3. The trivialization problem was considered in (Salmon, 1982) and (Linsky, 1984). For recent influential versions of it, see (Schwartz, 2002, 268–9) and (Soames, 2002, 250–1). Talking about general terms — involved in (some of ) the predicates — rather than predicates themselves, (Linsky, 2006) offers a Lewisian “double indexing” model for representing how some of them might be flexible. Each expression can be associated with a function that assigns to each possible world the property (or kind) the expression signifies with respect to that world — which in turn can be modelled as a function that assigns to each possible world the extension of the property with respect to that world. As I argue in (López de Sa, 2006), this by itself does not solve the corresponding trivialization problem. I don’t think Linsky would disagree, given that he alludes to the relevant considerations being forthcoming once the expressions are embedded in contexts involving modal or other intensional operators, (Linsky, 2006, 661). Although he does not elaborate on this, the thought might be similar to the one I mention in the text just below. In his reply to Linsky, Soames (2006) concedes that a distinction can be drawn between rigid and flexible general terms (such as ‘blue’), and derivatively between the predicates which involve them (such as ‘is blue’). The proposal is not as general as mine, given that not every predicate comes from a general term by attaching the copula (see ‘runs’). Besides, and as we are about to see, it is my view that essentially the same kind of consideration that can be used for distinguishing between rigid and flexible general terms is also available for the case of canonical nominalizations of predicates. My thanks to an anonymous referee for this journal for suggesting that I should comment here on Linsky’s and Soames’ recent papers.

223

between the properties signified other than identity, sometimes perhaps mere extension subsumption. One of them is the following: And as to being loyal, what is that? It is being truthful! It is being faithful! (Joseph Conrad, The Arrow of Gold). I agree: some statements of the same form are indeed sometimes naturally understood as conveying (in the appropriate contexts) relations between the properties signified other than identity.4 But the defense I submitted, contrary to what he seems to hold, is not incompatible with acknowledging this. For it to work it suffices that statements of the form do sometimes convey identity, and that mine are cases thereof.5 More in particular, it suffices that my examples above can be understood such that one might have intuitions about their being contingently true by contrast with Having a heart is having a liver; Being in pain is having C-fibers firing; which are, intuitively, false or necessary. I am perfectly aware that my appeal to intuitions here can be contested. What it is important to appreciate is that the trivialization problem is overcome if one could have such intuitions, regardless whether we do in fact do — or, indeed, perhaps merely that it is not trivially false that we might do. For then one can motivate in terms of them the (possible) flexibility of a predicate, even in the presence of a property which tracks actual and counterfactual extensions of the predicate. If, contrary to what I think, all our candidate flexible predicates turn out to be rigid, then far from being a trivial consequence of the simple proposal, this would be a substantial result about the intuitions we do in fact have regarding the cases in question.6 And all this is, of course, compatible with the claim that by ‘is’ people sometimes convey relations other than identity. 4. One might feel that Schnieder is nonetheless overstating his claim when he says that “the ‘is’ usually plays a different role [than conveying identity] in such statements. … [Conveying identity] is just not the common function of the ‘is’ in similar statements” (Schnieder, 2005, 230–1, my emphasis), on the basis of this and a couple of other examples. 5. Notice that my defense is compatible with, but is not committed to, the view that the relevant statements are ambiguous between one and other kind of meaning. An alternative towards which I am more sympathetic would have it that the literal meaning is given by the strict identities, and that the rest of the relations are (sometimes) conveyed as a matter of nonliteral, conversational or conventional implicature. I am indebted here to the comments by one referee for this journal. 6. In my view, the case can be made for the stronger, more ambitious conclusion that the

224

3. The five arguments for global rigidity rebutted Take ‘having the color of ripe tomatoes’ as a paradigm for a flexible nominalization, if any is. According to the defender of the simple proposal, this can flexibly signify (in actuality) the property being red, instead of rigidly signifying the property having-the-color-of-ripe-tomatoes. In other words, it is true to say (in actuality) that having the color of ripe tomatoes is (identical to) being red. Schnieder offers five arguments against this claim. I reject them in turn. Argument 1: [The Kripkean test case for rigidity requires that ‘Having the color of ripe tomatoes might not have been having the color of ripe tomatoes’ have a true reading — DLdS] What about the property of having the color of ripe tomatoes? Could it have failed to be the property of having the color of ripe tomatoes? To me this question sounds odd; I cannot see what a positive answer should be meant to assert. (Schnieder, 2005, 232, my emphasis) Reply: The question is not odd (concerns of verbosity aside), and it is straightforward enough what a positive answer should be meant to assert: that having the color of ripe tomatoes might have been being blue instead. Or at least, this is what the contender will fairly claim, given that, as we have seen, it is precisely appealing to intuitions about the contingent truth of the identification that she motivates her view. Argument 1 is therefore obviously not an argument for the rigidity of nominalizations that is independent of the issues we have discussed in the preceding section. Argument 2 : Having the color of ripe tomatoes is a relational property. Whatever has this property has it in virtue of standing in a certain relation to ripe tomatoes. Being red, however, is not a relational property. Or even if it is, it is at least not a property which things have in virtue of standing in a certain relation to ripe tomatoes. So, by Leibniz’ Law, being red cannot be identical to the property of having the color of ripe tomatoes. (Schnieder, 2005, 232, my emphasis) Reply: It is hardly contestable that by ‘in virtue of ’ philosophers might mean to convey a number of different relations in different contexts: epistemic, causal, semantic, constitutive, among perhaps many others. I guess there is a sense in which it is perfectly acceptable to say that whoever is the Pope is the Pope in virtue of holding a certain position, being the Head candidate examples are actually flexible. For this and further discussion, see (López de Sa, 2003).

225

of the Catholic Church. In that sense it is certainly also acceptable to say that whatever has the property of having the color of ripe tomatoes does so in virtue of standing in a certain relation to ripe tomatoes. But this being true in the relevant sense falls short of supporting the emphasized contention about relationality — in much the same way as the claim about the Pope does not require that a relation to the position be built into the nature of the person signified! There might very well be stronger senses for the locution, so that the ‘in virtue of ’ claim suffices for the property being relational. I take it that whatever has the property having-the-color-of-ripe-tomatoes does so in virtue of standing in a certain relation to ripe tomatoes, in this stronger sense. But merely asserting that the same holds for having the color of ripe tomatoes amounts to merely asserting that ‘having the color of ripe tomatoes’ signifies having-the-color-ofripe-tomatoes. And this is precisely what was at stake. Argument 3: Having the color of ripe tomatoes is a property possessed in virtue of standing in the relation of sameness of color to ripe tomatoes. Since ripe tomatoes are red, things stand in the said relation to ripe tomatoes in virtue of being red. So, it is correct to say that things have the property of having the color of ripe tomatoes in virtue of being red. However, things are not red in virtue of being red (‘in virtue of ’ indicates some kind of explanatory relation and a thing’s being red is surely not self-explanatory). So again, there is something truly predicable of the one property while not so of the other: the things which have it, have it in virtue of being red. Leibniz’ Law yields the non-identity of the two properties. (Schnieder, 2005, 232–3, my emphasis) Reply: Again, there is arguably a sense in which it is perfectly acceptable to say that someone is the Pope in virtue of being the Head of the Catholic Church, although it would be odd, and not very explanatory, to say that he is the Head of the Catholic Church in virtue of being the Head of the Catholic Church. In this sense, the things Schnieder says that I have highlighted are also clearly acceptable. But this sense better not suffice for non-identity, on pain of the Pope not being the Head of the Catholic Church, after all! Argument 4 : My shoes are brown and thus they lack the property of having the color of ripe tomatoes. But my shoes would have possessed this property (while my red shirt would not) if ripe tomatoes had been brown. Hence there is a property, namely the property of having the color of ripe tomatoes, which is possessed by my shirt, while under some counterfactual circumstances it would have been possessed by my shoes. … There is such a property, we say, and we refer to it with the term ‘having the color of ripe tomatoes’. This property

226

cannot be the property of being red — for that property would not have been possessed by my shoes, if ripe tomatoes had been brown (tomatoes having a different color would not have affected the color of my shoes). (Schnieder, 2005, 233, my emphasis) Reply: Schnieder’s shoes lack both having-the-color-of-ripe-tomatoes and the property of being red, and hence lack the property of having the color of ripe tomatoes, whichever of the two is signified (in actuality) by the nominalization involved. The shoes would have possessed the property having the color of ripe tomatoes if they were brown, again whichever it is of being brown or havingthe-color-of-ripe-tomatoes that the nominalization would signify (in the envisaged situation). Furthermore, there is certainly a property which is possessed by Schnieder’s shirt, while in some counterfactual circumstances it would have been possessed by Schnieder’s shoes, namely having-the-color-of-ripe-tomatoes, which is not any of the colors. So far so good. Now merely asserting that this is the one we refer to by the term ‘having the color of ripe tomatoes’ amounts to merely asserting that ‘having the color of ripe tomatoes’ signifies having-thecolor-of-ripe-tomatoes. And again this is precisely what was at stake. Argument 5 : Whoever thinks in the above case that ‘having the color of ripe tomatoes’ non-rigidly designates the property of being red (since, after all, red is the color of ripe tomatoes) should for reasons of parity hold the parallel claim that ‘to be the virtue that Socrates was most famous for’ non-rigidly designates the property of being wisdom (since, after all, wisdom is the virtue Socrates was most famous for). But now we may notice that (1) To be the virtue that Socrates was most famous for is only an accidental feature of wisdom. After all, Socrates could have been most famous for his piety than his wisdom. But whoever thinks that ‘to be the virtue that Socrates was most famous for’ denotes the property of being wisdom cannot account for the truth of (1). Being wisdom, evidently, is an essential, not an accidental, property of wisdom. So the two property designators differ in reference. (Schnieder, 2005, 233, numbering altered) Reply: The expression ‘is wise’ is a predicate, whose canonical nominalization is ‘being wise’, which arguably (rigidly) signifies the property being wise, i.e. wisdom. By contrast, ‘has the virtue Socrates was most famous for’ flexibly signifies the different virtues, being wise, being pious etc. Socrates might be most famous for — or so the defender of the simple proposal would probably hold. Schnieder’s cases in Argument 5 are different. The ‘is’ in ‘is the virtue Socrates

227

was most famous for’ or ‘is wisdom’ is that of identity, not of predication, and hence similar to ‘is (identical to) Ratzinger’ or ‘is (identical to) the Head of the Catholic Church’. It is not clear what one should say about the expressions ‘being (identical to) the virtue Socrates was most famous for’ and ‘being (identical to) the Head of the Catholic Church’. But suppose, for the sake of the argument, that one holds that they flexibly signify (respectively) the property of being wise and of being Ratzinger in actuality — as Schnieder seems to think the defender of the simple proposal should do. Consider the following line of thought: (2) Being the Head of the Catholic Church is only an accidental feature of Ratzinger. After all, he could have been a Buddhist all his life. But again, intuitions concerning (2) better be compatible with Ratzinger being (identical to) the Head of the Catholic Church! Mutatis mutandis for the virtuous case. All in all, none of the five arguments is compelling.

4. The general “explanation” rejected Schnieder devotes his final section to providing a general explanation of why all nominalizations are rigid. This involves the following principle, “which reveals both the reference conditions for canonical property terms and the nature of the properties denoted” (Schnieder, 2005, 236): (P) Being F (to be F, F-ness) is the property which is essentially such that it is possessed by all and only Fs. Suppose this is right. According to Schnieder, if principle (P) is correct, it straightforwardly explains why canonical property designators are all rigid. (Schnieder, 2005, 237) This is not so. To show this, I assume that a nominalization is flexible, and we see that the corresponding instance of (P) is nonetheless true. Suppose then that ‘having the color of ripe tomatoes’ flexibly signifies being red in actuality. Hence having the color of ripe tomatoes is being red. Now being red is the property which is essentially such that it is possessed by all and only red things, and red things are all and only the things that have the color of ripe tomatoes. Hence having the color of ripe tomatoes is the property which is essentially such that

228

it is possessed by all and only the things that have the color of ripe tomatoes. Hence the instance of (P) is true. (Furthermore, this line if argument is reproducible with respect to any counterfactual situation. Suppose we consider a world w in which ripe tomatoes are blue. ‘Having the color of ripe tomatoes’ flexibly signifies being blue wrt w. Hence ‘having the color of ripe tomatoes is being blue’ is true wrt w. Now ‘being blue is the property which is essentially such that it is possessed by all and only blue things’ is true wrt w, and ‘blue things are all and only the things that have the color of ripe tomatoes’ is true wrt w. Hence ‘having the color of ripe tomatoes is the property which is essentially such that it is possessed by all and only the things that have the color of ripe tomatoes’ is true wrt w. Hence the instance of (P) is true wrt w. Hence the instance of (P) is indeed necessarily true.) Therefore, there can be no argument in terms of (P) for the rigidity of all canonical nominalizations.7

Conclusion I conclude that, contrary to Schnieder’s contention, the simple proposal can indeed overcome the trivialization problem.8

7. Schnieder (2005, 237) says that (P) can also be put more formally as follows: (P*) Being F = the property x such that  y( y has x l y is F). If ‘x’ here could be understood flexibly, then (P*) would seem an appropriate regimentation of (P), and the remarks in the text would apply. Otherwise (P*) does settle the issue, but by fiat. If ‘x’ is required to be rigid, (P*) amounts to the claim that any canonical nominalization does signify the property whose extension is the extension of the predicate in each possible worlds. That is to say, it asserts that in particular ‘having the color of ripe tomatoes’ does signify having-the-color-of-ripe-tomatoes. But it merely asserts it, it does not explain it or provide grounds for it. Something along the lines of (P) might be supported by Schnieder’s contention about understanding of the nominalization depending on understanding on the embedded general term (2005, 236) — or so, at least, one can take for granted for the sake of the argument. But obviously this does not suffice for vindicating the considered reading of (P*). 8. Thanks are due to Benjamin Schnieder for very stimulating discussion, and anonymous referees for useful suggestions. Research funded by projects HUM2004-05609-C02-01 (MEC) and BFF2002-10164 (ESF), and grant EX2004-1159 (MEC). Thanks to these institutions and to Mike Maudsley for his linguistic revision.

229

REFERENCES Kripke, S. (1980). Naming and Necessity. Oxford: Blackwell. LaPorte, J. (2000). “Rigidity and Kind”. Philosophical Studies, 97, 293–316. Linsky, B. (1984). “General Terms as Designators”. Pacific Philosophical Quarterly, 65, 259–76. — (2006). “General Terms as Rigid Designators”. Philosophical Studies 128, 655– 67. López de Sa, D. (2001). “Theoretical Identifications and Rigidity for Predicates”. In J. M. Sagüillo et al. (eds.), Formal Theories and Empirical Theories. Santiago de Compostela. — (2003). “Rigidity for Predicates and the Trivialization Problem”. (Unpublished.) — (2006). “Rigidity, General Terms, and Trivialization”. Proceedings of the Aristotelian Society. (Forthcoming.) Salmon, N. (1982). Reference and Essence. Oxford: Basil Blackwell. — (2003). “Naming, Necessity, and Beyond”. Mind, 112, 475–92. — (2005). “Are General Terms Rigid?” Linguistics and Philosophy, 28, 117–34. Schnieder, B. (2005). “Property Designators, Predicates, and Rigidity”. Philosophical Studies, 122, 227–41. Schwartz, S. P. (2002). “Kinds, General Terms, and Rigidity: A Reply to LaPorte”. Philosophical Studies, 109, 265–77. Soames, S. (2002). Beyond Rigidity: The Unfinished Semantic Agenda of Naming and Necessity. Oxford: Oxford University Press. — (2006). “Reply to Critics”. Philosophical Studies 128, 711–38.

230

Grazer Philosophische Studien 73 (2006), 231–235.

Alfred NORDMANN: Wittgenstein’s Tractatus: An Introduction. Cambridge: Camridge University Press 2005. xi + 234 pp. £ 15.99. The first thing to say about this book is that it is not an introduction to the Tractatus. It is not necessary to know just what an introduction to the Tractatus might look like in order to know that this is not one; it is enough to note that the book sets out to present an interpretation of Wittgenstein’s early work, but does not attempt to provide an interpretative narrative that is systematically grounded in the details of the text. It would not be possible to give this book to a novice in the hope that it would help to guide him or her through the text, illuminating individual remarks, and showing how they build on one another to achieve Wittgenstein’s philosophical aims. What the book provides are some suggestions for how the work of detailed interpretation might be approached, and it is hard to believe that anyone who is not already familiar with both the text and with the controversy between Cora Diamond, James Conant and Peter Hacker would be able to make much of it. This does not, of course, entail that this is a bad book on the Tractatus, only that it is not a good introduction to it. How good the book will be found to be depends on how compelling the reader finds the interpretative strategy that it recommends. Does the book present a promising interpretative approach? The book includes a wide range of interpretative claims and it is not always easy to see how they hang together. It begins by placing the Tractatus in the critical tradition of philosophy and it includes a very brief outline of ideas from Kant, Lichtenberg and Hertz that are thought to have influenced Wittgenstein. Nordmann introduces the idea of a critique of language and the idea that the relation between representations and their objects is an “internal” relation. Both of these ideas are held to be fundamental to the Tractatus. They are held to be the source of Wittgenstein’s idea that the limits of language must be uncovered from within, and

that uncovering the limits of language is the solution to the problem of philosophy. This critical picture is complicated by the fact that the relation between logic and language has been made an issue by Frege’s and Russell’s claims that the propositions of logic are maximally general truths. Wittgenstein’s Kantian commitments lead him, it is suggested, to reject this idea and to insist that logic has no subject matter. Nordmann argues that this idea is essential to the claim that the relation between language and the world it represents is internal: logic shows the logical form that is shared by language and the world insofar as they are internally related to one another. To think that logic has a subject matter is to mistake the logical for the empirical. Wittgenstein’s “fundamental idea that the ‘logical constants’ are not representatives; that there can be no representatives of the logic of facts’ (TLP 4.032) is held to be connected with the fundamental Kantian concern to show that the relation between language and the reality it depicts is internal or immediate. All of these ideas are very difficult and they touch on many of the topics that preoccupy Wittgenstein in the remarks on logic and language that make up the principal part of the Tractatus. However, precisely insofar as they are very difficult, they require a careful explication grounded in the demanding details of the text itself. This section of the book is not so much a recipe for interpretation as a list of the principal ideas that any interpretation needs to clarify and locate within the context of an understanding of Wittgenstein’s overall project. This is what Nordmann does not provide. These topics are covered in less than 50 pages (something less than a quarter) of the book. None of these topics is treated at any length, or with any sense of its intrinsic obscurity, or even with any sensitivity to the question whether Wittgenstein succeeds in realizing his critical aims of staying within language and does not fall into metaphysical dogmatism. The remainder of the book focuses on the much more general issue of the status of Wittgenstein’s remarks, and the question how an author’s philosophi-

cal aims could be achieved through the medium of remarks which the author himself aims to show are nonsensical. It is in connection with these more general questions that Nordmann’s principal interpretative claims are made. One of Nordmann’s central interpretative claims is that the Tractatus takes the form of a reductio argument in which the starting hypothesis is that every thought we have can be expressed in speech. (Nordmann concedes that the hypothesis is not stated in the Tractatus, but he traces it to a — somewhat tendentiously interpreted — remark in the Notebooks.) The ideas about representation, logic and language that are presented in the first fifty pages of the book are held to form the background to the reductio argument, insofar as they show that language succeeds when it represents states of affairs that can either exist or fail to exist. The investigation of empirical language shows that propositions with sense are contingent propositions, which can be either true or false, or which describe possible states of affairs: ‘The sense of a sentence is therefore limited from the beginning to what is depictable’ (p. 60). A corollary of the starting hypothesis is that the statement that nothing is inexpressible in speech is not nonsensical. Yet this statement does not depict a possible state of affairs. Every statement with sense describes a possible state of affairs. Thus we are led to a contradiction and thereby to what Nordmann sees as the conclusion of the Tractatus, the rejection of our starting hypothesis: ‘There is indeed the inexpressible. This shows itself; it is the mystical’ (TLP 6.522, Ogden’s translation). Thus Nordmann places himself in opposition to the so-called resolute reading of the Tractatus put forward by Diamond and Conant, in holding that the main point of the book is precisely to deny that the limits of language coincide with the limits of thought: there is the inexpressible in speech. He argues that Wittgenstein recognizes four categories of sentence: (1) well-formed sentences with sense (empirical propositions); (2) well-formed sentences that are senseless (the propositions of logic); (3) ungrammatical sentences that lack sense (pure nonsense sentences); (4) ungrammatical sentences that are not senseless (sentences that show what

they cannot say). The remarks of the Tractatus, sentences of metaphysics, and sentences of ethics, all belong to this fourth category. In this way, Nordmann argues, we can recognize that Wittgenstein’s remarks are nonsense (have no sense), and recognize it on the basis of the reductio argument, and yet avoid the paradoxical claim that Diamond and Conant are committed to: that the remarks by means of which we achieve our insight are senseless gibberish. Although they say nothing, they show what they cannot say; and it is through coming to grasp what they show that the reader comes to understand the author and recognize that his remarks are nonsense: they fail to express what the author means. However, Nordmann recognizes that he avoids one problem only to be presented with another: What does it mean to say that an ungrammatical proposition that says nothing, or is nonsense, is not senseless? How are we to understand the idea of a sense that cannot be expressed? A large part of Nordmann’s book is taken up with the attempt to make this idea intelligible and to show that Wittgenstein embraces it. It’s at this point that the instability of the interpretative approach that Nordmann favours begins to make itself manifest. Nordmann recognizes that whatever sort of sense a sentence of the fourth category has, it cannot be the same sort of sense as the kind that is expressed by an empirical proposition with sense; that is to say, it cannot be a thought that is true or false. At times he is inclined to try to exploit an ambiguity in what he sees as Wittgenstein’s conclusion — ‘There is indeed the inexpressible’ — and hold that it leaves open whether there is something we are silent about, or whether it is simply that not everything we desire to say is expressible in speech. The latter is clearly a much weaker claim. The former requires us to make sense of two kinds of senses (the expressible and the inexpressible), whereas that latter leaves it open whether there is any sense that we are attempting to express, and in particular leaves open the possibility that there is none, that our words are plain nonsense. This latter thought seems to be the one that Wittgenstein expresses in ‘A Lecture on Ethics’, when he acknowledges and expresses respect for a desire ‘to go beyond the world and that is

232

to say beyond significant language’, but does not commit himself to the idea that the nonsense we thereby essentially come out with shows something that cannot be said. Rather, he remarks that: ‘This running against the walls of our cage is perfectly, absolutely hopeless. Ethics so far as it springs from the desire to say something about the ultimate meaning of life, the absolute good, the absolute value, can be no science. What it says does not add to our knowledge in any sense. But it is a document of a tendency in the human mind which I personally cannot help respecting deeply and I would not for my life ridicule it’ (Philosophical Occasions, p. 44). However, the very weakness of the thought expressed here means that it cannot serve Nordmann’s interpretative ends, for it does not appear to recognize his alleged fourth category of sentence: sentences that have a sense even though they are nonsense. Thus, Nordmann appears to go against his own intuition that Wittgenstein does not commit himself to a something that is inexpressible, by arguing for the idea that nonsense sentences may have a sense that cannot be said but which shows itself. However, when we look at the details of Nordmann’s attempt to make the idea of a sense that cannot be said intelligible, he seems to slide between two alternative and incompatible ideas. On the one hand, in his official account of the problematic notion of sense, he seems to want to respect the idea that the kind of sense he is concerned with cannot be equivalent to a thought that is true or false. The result is a weak account of the sense of nonsense sentences on which their having a sense is simply a matter of their prompting certain responses in us. For example, the sentence ‘I am safe whatever happens’ may prompt feelings, ideas and images in the mind of the speaker, or in the mind of someone who hears it uttered. However, even if Nordmann is right that this is the case, it seems to be a matter of the sentence’s utterance having certain mental accompaniments; these mental accompaniments do not connect with the logical structure of the sentence and they do not obviously warrant the idea of either cognitive or semantic content. Perhaps we would speak of understanding someone who said these words, but this is more a question of

understanding his or her desire to say them, of entering imaginatively into the desire or feeling that prompts them. However, the idea that the words themselves have a form of cognitive or semantic content that would warrant the idea that they have a sense that belongs to the same genus as the sense of an empirical proposition has not been made out. It seems right, as Nordmann suggests, to compare the sentence ‘I am safe whatever happens’ with a gesture, or to see them as expressive of a feeling or an image, but none of this amounts to the idea, which elsewhere he seems inclined to defend, that there is something inexpressible (a thought), as it were, behind a nonsense sentence, which shows itself, which is conveyed, but which cannot be said. Rather, it seems only to suggest that a gesture, or a piece of music, or a painting may be apt to express a feeling or an image that cannot be expressed in speech; it may also be that gestures are sometimes made by means of language, when cognitive content is subordinate to the images and feelings that our words prompt. Acknowledging this does not require us to recognize a second species of sense, or a class of sentences that are nonsensical but not senseless, or hold that there are thoughts that lie beyond the limits of language. The official account of the sense of nonsense sentences is, then, significant but philosophically innocuous. However, it also offers Nordmann very little when it comes to his understanding of the status of the remarks of the Tractatus. For it suggests that, although the remarks may prompt feelings, ideas and images in the mind of a reader, they cannot convey anything that could properly be called a thought or substantial philosophical insight. Rather, Nordmann’s official account of the sense of nonsense sentences would seem to indicate the view of Diamond and Conant: that the point of the remarks is that a reader realizes that the images that are prompted by a sentence such as ‘The world is all that is the case’ create a mere illusion of sense (of thought). Indeed, Nordmann sometimes seems inclined to adopt this view. However, the idea that the remarks of the Tractatus are plain nonsense that merely prompt characteristic images and feelings in us appears to conflict with Nordmann’s desire to express the

233

point of the work in the form of a proposition that we are brought to accept as the result of the reductio argument: ‘There is indeed the inexpressible’. Although the remark is held to be nonsensical, what it shows but does not say is understood as a conclusion, as something that contradicts Diamond and Conant’s claim that there is only one kind of nonsense, as an insight that is conveyed by the work, and so on. In this case, Nordmann is led into trying to say what these words allegedly cannot say, and thus into doing what Ramsey and Russell accuse Wittgenstein of doing: saying an awful lot about what he claims cannot be said. Moreover, on Nordmann’s interpretation, it is only the background claim, that only empirical propositions that describe states of affairs have sense, which prevents us from taking the words that he uses to express the point of the work at face value. But in that case, why isn’t the conclusion that the background assumption is wrong? Wittgenstein’s account of representation may work for some sentences with sense (namely those that express empirical propositions), but it doesn’t cover other areas of meaningful discourse, including metaphysics or philosophy. Wittgenstein’s pessimistic conclusion concerning the status of his own remarks is just an indication that he over-generalizes his theory of semantic content. (The status of the remarks that express this theory of representation is not something Nordmann ever considers and it is quite unclear whether he also takes these remarks to belong to his fourth category.) Nordmann’s apparent ability to tell us the point of the book would seem to encourage the idea that not only empirical propositions express thoughts. It is central to his interpretation of the work that it has a point that is conveyed, but not said, by the remark ‘There is the inexpressible’, but the most that his official account of the sense of nonsense sentences permits is that this sentence prompts images and feelings in a reader. Nordmann’s attempt to state the point of the remarks of the Tractatus contrasts sharply, not only with his official account of the sense of nonsense, but also with Wittgenstein’s reflections on ‘I’m safe whatever happens’ in ‘A Lecture on Ethics’. Wittgenstein makes no attempt to explicate or interpret what thought some-

one who says these words conveys; he seems to regard the words as ‘absolutely hopeless’. It is simply that he wants to record his recognition and respect for the human desire that prompts the utterance of these senseless words: the desire to go beyond the limits of language and speak of “what is higher”. This seems, to some extent at least, a softening of the view of the Tractatus, in which he enjoins silence, and of the letter to Fricker, in which he appears to commend his having avoided “gassing”. However, there is nothing in it that suggests Wittgenstein accepts that there are thoughts that these senseless words are trying but failing to say, and which a hearer must somehow cotton on to. The final impression, then, is that Nordmann provides an account of a distinction between mere gibberish (‘Piggly wiggle tiggle’) and nonsense (‘I’m safe whatever happens’) that is expressive of a gesture or feeling or which prompts ideas and images in us. However, the account does not clearly amount to the articulation of a second species of a single genus, the sense of a sentence, and does not therefore warrant the claim to have made intelligible the idea of thoughts that lie beyond the limit of language. Moreover, it is an account that seems ill-suited to cope with philosophical remarks which Nordmann wants to interpret as drawing a conclusion concerning the limits of language, which he is led into trying to state. It may be that a second major interpretative claim — that the remarks of the Tractatus have the form of aphorisms — is intended to save the appearance of a contradiction. Applied to the remarks on logic and language the claim is not an immediately compelling one, and the suggestion is almost impossible to assess in the absence of the interpretation itself, i.e. in the absence of a detailed narrative grounded in the details of the text. If the very idea that the work is aphorisitic precludes this — i.e. if it means that the work is only intended to prompt ideas in us without making any claims - then we are back once again with the apparent contradiction that Nordmann finds himself driven to present the point of the work in the form of a stated conclusion. The instability of the interpretative approach leaves us hovering between the idea that its remarks prompt

234

a purely subjective play of ideas and images and the idea that they convey an allegedly ineffable truth that Nordmann nevertheless manages to state. It will be an unfortunate consequence if Diamond’s and Conant’s important work on the Tractatus leads interpreters to think that it is possible to investigate the point of Wittgenstein’s early work by focusing on the single issue of whether there is more than one kind of nonsense, or on the single question of the status of Wittgenstein’s remarks. There is nothing in their work that warrants the idea that the Tractatus can be approached any way other than through the details of the text, yet it is possible to see that there is a danger that the sheer power of the question they raise — how to read the text in a way that does justice to TLP 6.54 — might encourage interpreters to focus directly on the end of the work and on what is involved

in throwing away the ladder, and recognizing that Wittgenstein’s remarks are nonsense. The idea that it is a test of a satisfactory reading of the Tractatus that it succeeds in taking Wittgenstein at his word in TLP 6.54 is, I believe, a salutary one, but if the result is that even a book that calls itself an introduction to the work focuses almost exclusively on the issue of the status of nonsense, and of whether (or whether Wittgenstein believed) that nonsense could show what it does not say, then it will have had the undesirable effect of distracting interpreters from the topics that principally preoccupy Wittgenstein: the nature of a proposition, the nature and status of logic, of the propositions of mathematics, and of laws of nature, the relation between the subject and the world, and the nature of ethics.

235

Marie McGINN University of York

Grazer Philosophische Studien 73 (2006), 236–238.

Severin SCHROEDER: Wittgenstein: The Way Out of the Fly-Bottle. Cambridge: Polity 2006. 288 pp. ISBN 0745626165. $ 24.95 (€ 22,50). This is an excellent new introduction to Wittgenstein. It is highly recommendable as an introduction to Wittgenstein for graduate students and advanced undergraduates. However, it is also highly recommendable as an introduction to Wittgenstein for professional philosophers (and dedicated non-professionals), since Schroeder is likely to challenge at some point one’s current conception of Wittgenstein. This is in large part due to the very close reading Schroeder offers of Wittgenstein’s two primary texts (the Tractatus LogicoPhilosophicus and the Philosophical Investigations), but it is more fully due to the insightful and innovative exegesis Schroeder offers of Wittgenstein’s critique of the inner-object model of psychological terms. This particular critique is the crowning achievement of Wittgenstein’s rejection of referentialism, and the rejection of referentialism is the main theme of this book. It helps Schroeder to explain both the relationship of the later Wittgenstein to the early Wittgenstein and the problems of a contemporary post-analytic philosophical scene that has failed to remind itself of the insights Wittgenstein’s critique offers. The book proceeds in chronological fashion. It begins with a philosophically oriented account of Wittgenstein’s early years (up to the publication of the Tractatus) and the cultural context of those years. This section includes an account of the moral sensibility of the young Wittgenstein (a point to which I shall return at the end of the review). Schroeder then presents the Tractatus, after which he pauses briefly to recount Wittgenstein’s activities after the publication of that text up through his resignation from Cambridge in 1947 (by which time Part I of the Philosophical Investigations had been written). A brief account of the events that led Wittgenstein to return to philosophy is also included. The next and largest portion of the book is dedicated to the Philosophical Investigations. This

is followed by a brief account of Wittgenstein’s final years and a chapter on Wittgenstein’s influence on later philosophy. This final chapter focuses in particular on the unjustified waning of his influence and the role played in that waning by Quine, Davidson, Kripke, and Putnam. The strength of Schroeder’s book as an introduction to Wittgenstein is that it is lucid and accessible: it presents the material not only in a way that is clearly written and cogently organized (so that it is very accessible), but it also presents the material in the fullness of its density and complexity (and so it is quite lucid). Of course, there are those who will not be pleased with Schroeder’s approach. Schroeder explicitly dispenses with the “New Wittgenstein” and he makes a point of explaining Kripke’s failure to understand the later Wittgenstein. There is also no “Wittgenstein-in-continuity”, meaning that there is no sustained look at Wittgenstein’s middle period or the continuities that could be said to unify the early and later work. Finally, there is no “third Wittgenstein”, meaning that Schroeder does not posit that On Certainty marks a further, third stage in Wittgenstein’s thought. The latter two (the absence of a middle and late Wittgenstein) are fully understandable omissions in an introduction. The former two (the rejection of the New Wittgenstein and the Kripkean Wittgenstein) are justified rejections. Schroeder is fully attuned to both the conceptual and aesthetic dimensions of Wittgenstein’s primary texts (the Tractatus and the Investigations), and it is precisely this attunement that leads to some of the most original and insightful aspects of his explanation of Wittgenstein’s thought. For instance, there is the concept of a synoptic representation (übersichtliche Darstellung), which is a tool Wittgenstein used in order to make it easier to survey a particular philosophical puzzle. A synoptic representation does this by presenting a simpler picture of what is puzzling us so that we may focus more clearly on that which leads to our confusion. The primitive language-games are an example, and much is

236

often made of the language-games and their centrality to Wittgenstein’s later thought. However, Schroeder’s introduction to Wittgenstein rests upon the premise that what appear to be synoptic representations often lack the sort of simplicity that should be the primary feature of such representations. This lack of clear ‘synopticity’ results in part from the way in which the Investigations is constructed and from Wittgenstein’s desire to not “spare other people the trouble of thinking” (PI viii). In fact, in some cases the pictures that the synoptic representations are intended to represent are missing. Thus, much of the work in interpreting Wittgenstein is dedicated to the (re)construction of such pictures. It is Schroeder’s success in fully and clearly constructing these targeted pictures which is one of the main achievements of the book. Schroeder applies this constructive method to the elucidation of Wittgenstein’s approach to three different puzzles, each of which stems from a referentialist picture of language: the paradox of the instantaneous experience of complex contents, the puzzle of first-person authority, and the problem of skepticism about other minds. Each of these puzzles is rooted in a “corollary of referentialism” that Schroeder labels “the Inner-Object Conception”, according to which “psychological terms stand for objects (states or occurrences) in the mind” (185). In other words, a psychological word (such as a sensation term) must refer to some object, and this object must be in the mind. At the start of the Investigations, the referentialist picture that is targeted is the one offered by the Tractatus, and Schroeder notes that a common reading of the Investigations claims that this negative, critical part of the Investigations is followed by the development and presentation of positive doctrines (151). However, Schroeder, in analyzing the dissolution of the three referentialist puzzles, is able to persuasively demonstrate that Wittgenstein does not go on to present positive doctrines but instead realized that referentialism was more deeply rooted than he had thought. Perhaps the best example of all this is in Schroeder’s treatment of the private sensation diary. The diary example from PI § 258 presents the case of a man (about whom Wittgen-

stein speaks in the first person) who keeps a diary in which he writes ‘S’ every time he feels a certain sensation for which he has no “natural expression” (§ 256). In other words, he has no criterion for the correct application of this sign, such that whatever seems right to him is right, and this in turn means (according to § 258) that we cannot talk about what is right or wrong. A common temptation here might be to conclude the following: First, the inner object account is mistaken, since it leads to the absurd conclusion that we can be mistaken about our sensations and yet never even know that we are mistaken. Second, since there is no way to determine if there is an inner object that corresponds to ‘S’, there is something deeply mistaken about what the diarist is doing. Third, we therefore need different — or clearer — criteria by which to judge our own sensations. Schroeder makes clear what is wrong with this thought. The rejection of the inner-object model of sensation does not mean that we need some better criterion or set of criteria by which to judge our talk of sensation. To respond to the disquietude that the sensation diary opens up by wanting better criteria is a sign that we are still deeply in thrall to the inner-object conception of psychological terms. We do not need criteria by which to judge what is objectively occurring in a person because our correctness in such situations is built into language. In other words, our language is built around the fact that people are generally both sincere and correct in their talk of what they are feeling. And when others tell us what they are feeling, we are not concerned with the correctness of their reports on the objective occurrences in their bodies. Instead, we are concerned with “what they feel: with what it feels like to them. The state of their body interests us only indirectly, insofar as it is likely to affect the way they feel, now or in the future. But where the two data contradict each other, it is the felt state that matters” (215). We are interested in the experience of the other person, not the physiological state of affairs. A similar lesson is found in Schroeder’s elucidation of Wittgenstein’s non-causal account of reasons. If the inner-object conception is assumed at some level, and if what

237

causes us to do things are connections that can be revealed by neurophysiological investigation, then, in giving reasons as causes for our behavior, we could easily be mistaken (meaning that those reasons might not correctly capture the underlying, objectively-present, neurophysiological causal connections). Furthermore, this causal account of reasons fails to address the deeper puzzle of how we can be certain about our reasons for having acted as we did. As Schroeder explains, for Wittgenstein it is not a matter of explaining actions by offering causes. Instead, the reasons an agent offers for why s/he did something “may give us insight into the agent’s character” (233). In other words, when we ask after reasons, we are asking after the person and not the neurochemistry, since it is the person and his/her experience that interests us. This is no small claim, since in making it Schroeder is charging much of contemporary post-analytic philosophy with continuing to think in accord with a discredited referentialism. This charge is made explicit in the final chapter, and Schroeder notes his debt here to P. M. S. Hacker’s Wittgenstein’s Place in Twentieth-Century Analytic Philosophy (1996). Hacker’s claim, in a nutshell, is that although the later Wittgenstein fully discredited referentialism (on which the Tractatus relied), it is in fact not the spirit of the later Wittgenstein but rather the “Tractarian spirit” that continues to exert influence on contemporary post-analytic philosophy. If this is so, then Wittgenstein’s investigations are incomplete — at least if these investigations were intended to undo the power of the referentialist picture.

And this takes us back to the ethical sense of things. As Schroeder noted in the first section of his book, the moral view of the young Wittgenstein had some distinct aspects, two of which were non-cognitivism and individualism: morality could not be given a rational foundation, and morality was based not on reasons but rather was rooted in a person’s character. These two features seem to be present in the later Wittgenstein — especially in his account of why there is no criterion for the self-ascription of sensations and of why we offer reasons for what we did. In both cases, limits are drawn to what can be rationally (and scientifically) determined, and these limits open up the space in which a person’s character is made manifest. If this is so, the conception of philosophy offered by Wittgenstein (both early and late) is consistent with Wittgenstein’s moral sensibility. Further, in treating philosophy as bound up with referentialism, we deny ourselves the means by which we express and reflect upon our actual everyday experience, and we destroy some of the autonomy that we accord ourselves as (moral) persons. Perhaps the claim I wish to make here is too strong, but it seems that Wittgenstein might wonder if, in abnegating the conception of philosophy he worked so hard to develop, we don’t face the danger of abnegating something fundamental about ourselves. Schroeder, in introducing us anew to this Wittgenstein, rejects such an abnegation and thus does us a great service.

238

Brian ARMSTRONG College of St. Benedict / St. John’s University

Grazer Philosophische Studien 73 (2006), 239–243.

Willem A. deVRIES: Wilfrid Sellars. Chesham: Acumen 2005. 337 + XIII S. ISBN 1844650391. (Paperback) £ 14.99. Das philosophische Interesse an der Philosophie von Wilfrid Sellars hat in den letzten Jahren stark zugenommen. Vor allem im Zusammenhang der Rezeption der Arbeiten von John McDowell und Robert Brandom erlebt Sellars’ Philosophie eine Renaissance und er selbst kommt als einer der großen Klassiker der analytischen Tradition wieder verstärkt in den Blick. Dabei gerät bisweilen in den Hintergrund, dass diese Vermittlung durch Autoren mit einem dominanten eigenen philosophischen Programm nicht unproblematisch ist: Denn hier werden naturgemäß eher diejenigen Aspekte von Sellars’ Denken in den Vordergrund gerückt, die sich für die eigene philosophische Agenda als besonders fruchtbar erweisen bzw. mit dieser zumindest vereinbar sind oder aber im Gegenteil Aspekte, die vor dem Hintergrund der eigenen Position am leichtesten als defizitär zu charakterisieren sind. Ersteres gilt vielleicht eher für Brandoms sprachpragmatischen Ansatz, der insbesondere die inferentialistischen Elemente in Sellars’ Philosophie betont; letzteres lässt sich im Zusammenhang von McDowells Kritik an Sellars’ Konzeption beobachten, wie er sie etwa in seinen ‚Woodbridge-Lectures‘ von 1998 formuliert. Umso erfreulicher ist es, dass Willem deVries mit seiner Einführung in Sellars’ Denken nun einen systematischen Überblick über die (theoretische) Philosophie von Sellars vorgelegt hat, dessen Verfasser sich selbst zurücknimmt und es dem Leser auf diese Weise erlaubt, einen gleichsam unverstellten Blick auf das Werk dieses herausragenden Autors zu werfen. Dies gelingt deVries insbesondere dadurch, dass er, erstens, Sellars selbst ausführlich zu Wort kommen lässt und, zweitens, Sellars’ Werk ausdrücklich als systematisches Ganzes in den Blick rückt, innerhalb dessen seine zahlreichen Beiträge zu verschiedensten Teildisziplinen der Philosophie zu verorten sind. Für eine angemessene Beurteilung des philosophischen Ranges von Sellars’ Werk

ist dieser Blick auf das systematische Ganze unerlässlich, weil Sellars, wie Jay Rosenberg in seinem Beitrag zur Stanford Encyclopedia of Philosophy betont, mehr als die meisten anderen Klassiker der analytischen Philosophie, ein systematischer Denker war. Es ist deVries hoch anzurechnen, dass er als erster versucht hat, die zentralen Teile dieses systematischen Ganzen in ihrem Zusammenhang darzustellen. Dass deVries sich dabei nicht auf die theoretische Philosophie beschränkt, sondern wenigstens in seinem 9. Kapitel („Practical reason“) dem Leser auch zentrale Themen von Sellars’ praktischer Philosophie nahe bringt und im abschließenden 10. Kapitel, das mit „The necessity of the normative“ überschrieben ist, in einen systematischen Zusammenhang stellt, verdient dabei eigens hervorgehoben zu werden. Ein Glücksfall für den Leser ist der vorliegende Band auch deshalb, weil sich mit deVries einer der besten Sellars-Kenner an die Aufgabe einer einführenden Gesamtdarstellung gewagt hat. DeVries ist selbst ein Schüler von Sellars, hat zahlreiche Artikel zu dessen Philosophie verfasst und ist Koautor eines ausführlichen Kommentars zu Sellars’ Klassiker „Empiricism and the Philosophy of Mind“. (Der Kommentar zu EPM wurde gemeinsam mit Timm Triplett verfasst und ist erschienen unter dem Titel Knowledge, Mind, and the Given. Reading Wilfrid Sellars’s “Empiricism and the Philosophy of Mind” bei Hackett Publishing Company, Indianapolis 2000.) Durch die eingängige, aber gründliche Art und Weise der Darstellung bietet deVries mit seinem Buch nicht nur demjenigen Hilfestellung, der sich einen ersten Überblick über Sellars’ Philosophie verschaffen und auf die Sellars-Lektüre vorbereiten möchte, sondern leistet gleichzeitig einen anregenden Beitrag zu zahlreichen Fragen, die in der Sellars-Forschung diskutiert werden. Dass diese Beiträge zum Teil kontrovers sind, versteht sich von selbst und ich möchte im Weiteren einige Punkte herausgreifen, die mir selbst nicht unproblematisch scheinen. DeVries beginnt seinen Überblick mit einer Skizze der methodologischen Impli-

239

kationen des synoptischen Blickes auf die Welt, den Sellars von der Philosophie einfordert: Philosophie muss die grundlegenden Strukturen des manifesten Weltbildes genauso in den Blick nehmen, wie die Möglichkeiten und Grenzen des sich entwickelnden wissenschaftlichen Weltbildes. Da sich diese Möglichkeiten und Grenzen der Entwicklung des wissenschaftlichen Weltbildes immer auch in Abhängigkeit von und im Vergleich mit dem manifesten Weltbild zeigen, ist der synoptische Blick auf beide Weltbilder wesentliche Voraussetzung für das ‚reflektierte Heimischwerden‘ in unserer Welt, das Sellars für die zentrale Aufgabe der Philosophie hält. Von welchen Voraussetzungen Sellars sich bei seiner Ausgestaltung dieses synoptischen Blicks leiten lässt, erläutert deVries gleichfalls in diesem 1. Kapitel: Es sind Naturalismus, (wissenschaftlicher) Realismus und Nominalismus – alles zusammen aber immer als Teil eines Bildes, das auch die Wahrheit hinter den zurückgewiesenen Gegenpositionen sichtbar werden lässt. Im 2. Kapitel beschäftigt er sich mit Sellars’ inferentialistischer Sprachphilosophie, insbesondere dessen metasprachlicher Auffassung von Bedeutungsaussagen. DeVries macht durch seine weiteren Ausführungen plausibel, warum die Sprachphilosophie an so früher Stelle erläutert werden muss: Sie ist Voraussetzung für die Auseinandersetzung mit der Sellarsschen Umsetzung seiner nominalistischen Überzeugungen und natürlich für das Verständnis von Sellars’ Konzeption der Intentionalität. Und sie ist darüber hinaus auch Grundlage zentraler methodologischer Annahmen, die im 3. Kapitel diskutiert werden, und, wie deVries im 6. Kapitel aufzeigt, für Sellars’ wissenschaftstheoretische Überzeugungen. Die Diskussion der methodologischen Fragen im 3. Kapitel beschäftigt sich mit dem, was man als die transzendentalphilosophischen Aspekte von Sellars’ Methodologie bezeichnen könnte: Mit der Frage nach dem Status von Kategorien, von apriorischen Aussagen und (leider nur kurz) mit der Frage, was Sellars’ Ansicht nach die ‚transzendentale Aufgabe‘ der Philosophie ist (deVries, 64 ff.). Diese kursorische Diskussion wird, wie ich glaube, dem Ausmaß nicht ganz gerecht, in dem Sellars bereit ist, seinem eigenen Phi-

losophieren eine transzendentalphilosophische Methodologie zugrunde zu legen. Zunächst ist hier festzustellen, dass deVries Sellars eine Auffassung von Transzendentalphilosophie zuschreibt, die er selbst folgendermaßen formuliert: „Transcendental philosophy, on this view, … is reflection on the most general norms and structures constitutive of cognitive engagement with the world“ (deVries, 66). Er bezieht sich dabei auf Sellars’ Verteidigung von Kants Vorhaben einer transzendentalen Psychologie gegen Vorwürfe, dieser betreibe selbst nur schlechte, ‚rationale‘ Psychologie. Sellars stellt in diesem Zusammenhang eine Analogie her zu dem, was er als transzendentale Sprachwissenschaft bezeichnet. Diese versucht “… to delineate the general features that would be common to the epistemic functioning of any language in any possible world” (Sellars, “Some Remarks on Kant’s Theory of Experience” Journal of Philosophy 64 (1967), 646). Einer solchen transzendentalen Sprachwissenschaft geht es also um die Untersuchung bzw. das Auffinden der in der Sprache selbst implizit und explizit immer schon vorhandenen Normen. Entsprechend beschäftigt sich die transzendentale Psychologie mit den – natürlich sehr allgemein beschriebenen – wesentlichen Eigenschaften, die jedes beliebige Begriffssystem haben muss, das Wissen von einer Welt erzeugt, deren Teil es selbst ist (vgl. Sellars ebd.). Transzendentale Psychologie versucht also, diejenigen Regeln oder Normen zu bestimmen, denen jedes derartige Begriffssystem immer schon unterworfen ist – weil es andernfalls nicht das sein könnte, was es ist: Mit Kant gesprochen wären andernfalls die Bedingungen seiner Möglichkeit nicht erfüllt. Sellars selbst geht es nun, wie der Kontext dieser Bemerkungen meiner Ansicht nach deutlich macht, genau wie Kant darum, die Eigenschaften reflektierend zu erfassen, die jedes beliebige Begriffssystem haben muss, das Wissen von einer Welt erzeugt, deren Teil es selbst ist – um die Bedingungen der Möglichkeit eines derartigen Begriffssystems. Damit kommt aber ein Aspekt von Sellars’ Methodologie in den Blick, den deVries vernachlässigt: die Notwendigkeit bestimmter extrem allgemeiner Strukturen unseres intentionalen Bezugs.

240

Das wird vor allem auch in seiner Diskussion von Sellars’ Auffassung kategorialer Strukturen deutlich. Diese allgemeinen kategorialen Strukturen können wir nicht ohne weiteres revidieren oder selbst zum Gegenstand der Kritik machen. Sie sind vielmehr Voraussetzung dafür, dass wir zu einem gegebenen Zeitpunkt eine kritische Distanz zu unserem Begriffsrahmen einnehmen können: Wir konzentrieren uns gleichsam auf die allgemeinsten formalen Strukturen dieses Begriffsrahmens und können so künftige Entwicklungen imaginieren. (Vgl. Sellars, Science and Metaphysics (1967), 140. In dieser Hinsicht weicht Sellars m. E. deutlich von Davidsons Position ab, die deVries für kompatibel zu halten scheint; vgl. deVries, 65.) Dass die faktischen Übergänge von einem Begriffsrahmen, einem Weltbild, zu einem anderen, die formal-kategoriale Struktur seiner Ansicht nach nicht einfach auflösen können, macht Sellars auch klar, wenn er davon spricht, dass solche Entwicklungen „new ways of schematizing categories“ (Sellars, Science and Metaphysics (1967), 49) einschließen, d.h. von kategorialen Nachfolgerbegriffen, die in systematischem Zusammenhang zu ihren weniger angemessenen Vorläufern stehen müssen. Der Dynamik der Begriffsentwicklung sind also zwar, wie deVries zu recht betont (deVries, 64), prinzipiell auch die Strukturen unterworfen, von denen wir glauben, dass jedes beliebige Begriffssystem sie besitzen muss, das Wissen von einer Welt erzeugt, deren Teil es ist. Doch gelten hinsichtlich dieser Aspekte unseres Begriffssystems strenge Bedingungen der Kontinuität zwischen Begriffssystemen, die sie vor anderen Aspekten dieser Systeme auszeichnen. Indem man Sellars’ eigene metaphorische Ausdrucksweise aufgreift, könnte man sagen, dass neue Arten und Weisen der Schematisierung von Kategorien die Kontinuität der Kategorien selbst nicht berührt. DeVries fährt im 4. Kapitel fort mit einer luziden Darstellung von Sellars’ Nominalismus. Hier werden insbesondere Missverständnisse hinsichtlich der Radikalität von Sellars’ Position überzeugend ausgeräumt: Sellars klingt, so macht deVries deutlich, nicht deshalb an manchen Stellen so, als hielte er das Reden über abstrakte Entitäten für unpro-

blematisch, weil er etwa eine abgeschwächte, bloß partielle Version des Nominalismus vertreten würde. Vielmehr schlägt er eine nominalistische Reinterpretation solcher Aussagen vor, die diese scheinbar problematischen Aussagen recht verstanden zu (im Sinne einer nominalistischen Position) unproblematischen macht. Die Analyse des Wissensbegriffs und Sellars’ Kritik am Mythos des Gegebenen im 5. Kapitel baut auf dem Kommentar zu EPM auf und erweitert diesen insbesondere um die Darstellung der wichtigen Auseinandersetzung mit den ‚differenzierten‘ Fundamentalisten Roderick Firth und Roderick Chisholm, die Sellars in einer ganzen Reihe von Aufsätzen bis hin zu seinen letzten Werken in immer differenzierterer Form führte. Eine kurze Bemerkung zu deVries’ Diskussion von Sellars’ Charakterisierung des Mythos in den späten Carus-Lectures, in denen Sellars seine Kritik am Mythos nach Ansicht mancher Autoren abgeschwächt zu haben scheint (deVries 114 ff.): Dort schreibt Sellars, dass der Kern des Mythos in folgendem Prinzip besteht: „If a person is directly aware of an item which has categorial status C, then the person is aware of it as having categorial status C“ (Sellars, „Foundations for a Metaphysics of Pure Process“, Monist 64 (1981), 11). DeVries liest das – wie u. a. auch der von ihm kritisierte William Lycan – nicht als allgemeine Charakterisierung des Mythos, weil es in diesem Zitat scheinbar nur um „facts about categorical status“ (deVries, 115) geht, nicht aber um harmlose Tatsachen, wie sie in Aussagen wie „Dieser Ball ist rot“ thematisiert werden. Dennoch habe Sellars, so deVries, seine Kritik nicht abgeschwächt, sondern nur im Kontext seiner Firth-Kritik spezialisiert (deVries, 116). Doch diese Lesart des Prinzips ist meiner Ansicht nach nicht zwingend: Für Sellars enthält ja jede Aussage über etwas immer auch explizit oder implizit Informationen über den kategorialen Status dieser Sache. Wenn ich sage: „Dieser Ball ist rot.“ verwende ich (unter anderem) bereits den Begriffsrahmen der Farbprädikation. (Dass Sellars dieser Ansicht ist, ergibt sich unter anderem aus seiner Kritik an der Theorie der abstraktiven Begriffsbildung, die nicht mit der Kritik

241

am Mythos des Gegebenen identifiziert werden darf. Sie setzt, mit anderen Worten, eine erfolgreiche Kritik am Mythos nicht voraus.) Vor diesem Hintergrund werden auch solch scheinbar harmlose Aussagen als Aussagen erkennbar, die das Gewahrsein des kategorialen Status bereits voraussetzen. Die richtige sellarsianische Reaktion auf Lycan wäre deshalb: Im erforderlichen Sinne harmlose Aussagen gibt es nicht. Im 6. Kapitel beschäftigt sich deVries mit Sellars’ Wissenschaftstheorie im weitesten Sinne: mit seiner Auffassung von Gesetzen, Theorien, Induktion und Wahrscheinlichkeit. Dieses Kapitel ist vor allem deshalb besonders wichtig, weil deVries sich hier detaillierter mit Fragen beschäftigt, die sonst häufig vernachlässigt werden, obwohl sie für Sellars’ Denken von entscheidender Bedeutung waren. In diesem Kapitel führt deVries eindrucksvoll vor, wie radikal sich Sellars’ Position von anderen unterschiedet. Diese Wirkung erzielt er dadurch, dass er Sellars’ Wissenschaftstheorie konsequent im Kontext von dessen Sprachphilosophie und seiner Zurückweisung des Mythos des Gegebenen verortet. (Das gilt insbesondere für seine Ausführungen zum Problem der Induktion.) Den Übergang zu Sellars’ Philosophie des Geistes vollzieht das 7. Kapitel, in dem deVries sich mit Sellars’ Theorie intentionaler Zustände beschäftigt, wie Sellars sie im Kontext des Jones-Mythos aus EPM entwickelt. DeVries begnügt sich allerdings keineswegs mit einer weiteren Darstellung des Jones-Mythos, sondern widmet einen großen Teil seiner Ausführungen den wichtigen Spezifikationen, die das Verhältnis von Sprechen und Denken in Sellars’ späteren Schriften, insbesondere in „Mental Events“ von 1981, erfährt. Die Überlegungen in diesem wichtigen Aufsatz lassen die Kontinuität zwischen vorsprachlichen und sprachlichen Repräsentationssystemen plausibel werden und sind gleichzeitig – durch eine Unterscheidung von propositionalen und Logik verwendenden Begriffssystemen – dazu geeignet, die spezifischen Unterschiede zwischen unserer menschlichen Repräsentation und anderen Repräsentationssystemen deutlich zu machen. Was deVries an dieser Stelle meines Erachtens nicht hinreichend explizit macht, ist der

Zusammenhang zwischen der Fähigkeit zu repräsentieren und der Fähigkeit zu sprechen: Sellars ist in „Mental Events“ offenbar dazu bereit, nicht-sprachlichen Repräsentationssystemen propositionale Struktur zuzusprechen – nicht allerdings, sie als Repräsentationssysteme von Sprechern zu behandeln. Es handelt sich bei diesen Repräsentationssystemen, wie Sellars betont, um komplexe „map-makers“ (Sellars, „Mental Events“, Philosophical Studies 39 (1981), 336) – und das Erstellen von repräsentationalen ‚Karten‘ hat als solches keine normative Dimension. Damit teilen solche Systeme aber auch nicht die Janus-Gesichtigkeit, die nach Sellars für unsere Sprache und, wie der Jones-Mythos deutlich macht, auch für unsere (in Analogie zu einer Sprache, die unserer zumindest sehr ähnlich ist, konzipierte) Intentionalität wesentlich ist. Insofern scheint mir deVries’ Konklusion („thinking need only be language-like“ (deVries, 191)) problematisch, sofern sie implizieren soll, dass die propositionalen Repräsentationssysteme, die Sellars in „Mental Events“ beschreibt, tatsächlich auch sprachartig sein sollen. Denn dazu fehlt ihnen die normative Dimension, die nur im Kontext von echter Intersubjektivität entstehen kann. Anders gesagt: Hier handelt es sich nicht um einen graduellen Unterschied, wie deVries vorzuschlagen scheint, sondern um einen prinzipiellen. Nach der Theorie begrifflicher Bewusstseinszustände behandelt deVries in Kapitel 8 Sellars’ Konzeption des sensorischen Bewusstseins. Dieser Aspekt von Sellars’ System gehört sicherlich zu denjenigen, die keinesfalls vernachlässigt werden dürfen, weil sie im Mittelpunkt von Sellars’ systematischer Philosophie stehen: Hier laufen zahlreiche Fäden seines Denkens zusammen. In der eingangs angesprochenen Sellars-Renaissance spielen sie allerdings entweder keine Rolle oder werden gar – wie bei McDowell – vehement zurückgewiesen. DeVries bemüht sich in diesem Kapitel erfolgreich, die oftmals verschlungenen Pfade dieser Teile des Systems in einigem Detail nachzuzeichnen und zu entwirren. Seine Analyse ist sorgfältig, erhellend und in ihrer Bewertung des systematischen Wertes teils affirmativ, teils vorsichtig abwägend. Das einzige, was aus meiner Sicht fehlt, ist

242

eine Erörterung der systematischen Funktion, die das sensorische Bewusstsein für Sellars in der Führung unserer Wahrnehmung ‚von außen’ übernimmt (vgl. Sellars, Science and Metaphysics, Kap. 1). Nicht nur deshalb, weil McDowell in seinen ‚Woodbridge-Lectures’ gerade diesen Aspekt von Sellars’ Theorie des sensorischen Bewusstseins zum Gegenstand einer ausführlichen Kritik macht; sondern weil es sich dabei, wie ich glaube, um einen zentralen Baustein für eine Erklärung der Abhängigkeit der begrifflich-normativen Strukturen von dem Einfluss einer unabhängig existierenden Realität handelt, die nicht Gefahr läuft, in den Mythos des Gegebenen zurück zu fallen. Sellars gelingt es auf diese Weise, mit dem Problem des ‚Weltverlusts’ umzugehen, dem sich begriffsinterne Theorien wie etwa die Theorie Robert Brandoms ausgesetzt sehen. DeVries schließt seine beeindruckende Tour de force durch Sellars’ Denken mit

zwei Kapiteln, in denen die praktische Philosophie die Hauptrolle spielt: Dem 9. Kapitel, das Sellars’ Theorie intentionalen Handelns zum Gegenstand hat, und dem abschließenden 10. Kapitel, in dem er eine Sellarssche Vision des synoptischen Blickes auf die Welt entwickelt – zu dem weit in der Zukunft liegenden Zeitpunkt, an dem eine ideale Wissenschaft uns die Mittel für die Erzeugung adäquater Abbildungen der Welt zur Verfügung stellt. DeVries macht deutlich, dass und inwiefern die praktische Dimension des manifesten Weltbildes auch dann ihre Bedeutung für unser Tun nicht verlöre: Das wissenschaftliche Weltbild, so wie Sellars es konzipiert, vermag es nicht, einen Ersatz für die im weitesten Sinne normativen Aspekte unseres Blicks auf die Welt zu schaffen.

243

Johannes HAAG Humboldt-Universität Berlin

Grazer Philosophische Studien 73 (2006), 244–248.

Geert KEIL and Udo TIETZ (eds.): Phänomenologie und Sprachanalyse. Paderborn: Mentis 2006. 277 pp. € 36.00. In 1931, the young philosopher Rudolf Carnap seized upon several sentences of Martin Heidegger’s lecture “What is Metaphysics?” attacking them as exemplary of the emptiness and senselessness of all “metaphysical” philosophy. The basis of Carnap’s confidence in the attack was his allegiance to the project of a newly scientific philosophy of logical analysis that would “overcome” metaphysics through a logical analysis of language. Meanwhile, Heidegger and other inheritors of the legacy of Husserl’s phenomenological return “to the things themselves” grounded in the description of first-person consciousness distanced themselves from the formal analyses of the logical structure of language that typified Carnap’s Vienna Circle. Although Carnap himself was versed in phenomenology and by no means critical of Husserl’s descriptive approach, his criticism of Heidegger, along with other similar instants of methodological dispute and mutual misunderstanding, played a decisive role in consolidating the widespread institutional recognition (especially in America and Britain after World War II) of a tradition of “analytical philosophy” grounded in the analysis of logic and language and distinct from (and often hostile to) phenomenology. The perception of a division persists today, and still exerts determinative effects on philosophical projects on both sides of the Atlantic. But as good practitioners and historians of both approaches have long known, the family quarrel between phenomenology and analytical philosophy has never actually been as deep as the sociological and institutional separation between them would suggest. Prominent analytical philosophers such as Ryle, Sellars, and Austin were well aware of the methods, aims, and results of the phenomenological school of Husserl and Heidegger and, at least sometimes, saw their own projects as congruent with them.1 Even Wittgenstein described his search for clarity regarding the structure of language, for a brief time in

the 1930s, as research into phenomenology. The label “analytical philosophy” itself originated only in the 1950s, and its exclusive use to distinguish its practitioners from those of “continental” philosophy is of even more recent origin. Still, by this time the damage was done, and the labeling of the two philosophical schools combined with the regularization and professionalization of academic philosophy to produce a legacy of misunderstanding, distrust, and self-imposed isolation that largely prevented philosophers who identified with each tradition from exploring, or even understanding, the resources of the other. It is therefore an encouraging sign that philosophers today are increasingly willing to overlook these institutional divisions and re-open the discussion, both methodological and substantive, between the representatives of the two approaches that should have (and, but for contingencies of history and sociological practice, could have) taken place several decades ago. The twelve essays collected in Phänomenologie und Sprachanalyse provide a fascinating collective portrait of this contemporary discussion, particularly as it is now unfolding in the German-speaking world. Even if it remains too early to speak of a fully accomplished “reconciliation” of phenomenology and analytic philosophy, these essays bear ample witness to the new and exciting possibility of an organic practice of philosophical investigation and reflection that draws on the histories of both traditions without favoring either one to the exclusion of the other. The methodological pluralism and rich diversity of methods and results shown here will help to demonstrate the merit of such a practice, in the context of which the arbitrary and destructive division between adherents of the two traditions might simply lapse into irrelevance. Drawn as they are from a variety of influences and specific projects, the twelve essays thematically address a wide diversity of epistemological, semantic, ontological, and metaphysical topics. In the best cases, they approach specific issues of philosophical concern through

244

approaches that are equally “phenomenological” and “analytic” and are informed by the history of both traditions. For instance, three of the essays (“Satz and Sachverhalt” by Herbert Schnädelbach, “Wahrheit and Welterschliessung,” by Cristina Lafont, and “Von Sachverhalt zum Sprachspiel,” by Anke Thyen) take up the interrelated semantical, ontological, and metaphysical issues involved in understanding the reference of sentences and propositions to states of affairs in the world, as this reference is influenced by various elements of social practice and historical situation. As Schnädelbach argues, it is essential to this understanding that we recognize a distinction, most directly drawn from Wittgenstein’s Tractatus but also suggested, at various points, by Husserl in his phenomenological analyses of the logic of language, between objects or particulars as the correlates of the reference of individual terms, and the more articulate complexes (Sachverhalten) that render sentences true or false. And as Thyen argues through an insightful comparison of Wittgenstein’s view of Sachverhalte with that of the phenomenologist Reinach, the recognition of such complexes, in their holistic inferential and evidentiary connections with each other and with everyday behavior and practice, leads directly to the acknowledgment of the more diffuse and widespread background unities of behavior, attitude, and comportment that Wittgenstein characterized as “forms of life” and phenomenologists have discussed under the heading of the “lifeworld.” Along similar lines, by way of an instructive comparison of the views of Heidegger and Putnam, Lafont’s essay considers the revealing convergence of both traditions, as they developed over the course of the twentieth century, with respect to the status of these backgrounds and of various propositions as “a priori” relative to them. Here as elsewhere, recognizing the striking commensurability of results that can be extracted from the two traditions does not preclude the acknowledgment of residual differences of aim and intention between them, but the substantial agreement between prominent representatives of the two traditions on a nontraditional concept of the a priori suggests at least the possibility that both theories can be seen as arising

out of a single, more or less unified, trajectory of philosophical development. Another essay in this genre is Geert Keil’s “Über die deskriptive Unerschöpflichkeit der Einzeldinge.” Considering issues of language, reference, modality and identity, Keil assembles a useful and very detailed review and examination of the various ways in which the description of an individual, or indeed its actual fund of properties or property-instances, might be (and has been) considered to be infinite and inexhaustible. In a similar manner, Dagfinn Føllesdal’s much shorter essay “Indikatoren und der Geist” quickly surveys the fascinating issues surrounding the functioning of indexical terms or “indicators” in natural language and their connection with issues of subjectivity and self-consciousness, showing clearly how the question of indexicality also connects with recent developments in the analytic tradition and pointing out a revealing parallel with Husserl’s own, much earlier treatment. It is unfortunate only that the brevity of Føllesdal’s treatment prevents him from bringing out the significance of this connection more fully or developing the implications for what is still very much an open question for philosophical analysis. As commentators have long noted, many of the most distinctive movements and developments of twentieth-century philosophy, both analytic and phenomenological, have arisen from philosophers’ pursuit of questions about mind and mentality in relation to language, logic, and meaning. Several of the essays collected here take up questions about the relationship of consciousness to language; the most helpful of these, for instance Carl Freidrich Gethmann’s essay on Heidegger, draw on the resources of analysis and phenomenology alike to gain clarity about the status of mental content, pre- or postlinguistic, relative to acts of judgment and communication. Gianfranco Soldati’s essay also provides a helpful historical backdrop to these developments by exploring Franz Brentano’s theory of “inner evidence” and defending it against objections. Since this theory directly influenced both Husserl’s phenomenology and later analytic treatments of intentionality, the clear exposition and defense that Soldati gives should, like many of the essays

245

here, significantly enhance the historical and methodological discussion of the relationship between language and consciousness. Somewhat less helpful, however, is the tendency, evident in several of the essays collected here, to oppose phenomenology to analytic philosophy by reference to their methodological grounding in reflection on consciousness and language, respectively. For instance, Manfred Frank’s essay attempts to settle the question of whether consciousness and thought are “essentially” linguistic, in order to determine whether an analytic methodology grounded in the analysis of propositions or a phenomenological one grounded in the direct experience of consciousness is preferable. And Kathrin Gluer’s contribution takes up the question of whether the phenomenological method of epoche — or the “bracketing” of assumptions of the existence of external objects in order to investigate the contents of immediate, private consciousness — can be reconciled with the externalist position held by many contemporary analytic philosophers, according to which the contents of thought and language are partly determined by external, public states of affairs. One problem with this portrayal of the division between the two approaches as the struggle of “consciousness philosophy” vs. “language philosophy” is historical. By opposing them monolithically as if they each represented only one, single “method,” the juxtaposition fails to acknowledge the great diversity and complexity of methods that actually have emerged internal to each tradition since their beginnings at the turn of the last century. Even Husserl’s own corpus contains a bewildering variety of different modes of analysis, reflection, and description for understanding and clarifying the contents of consciousness and the metaphysics of worlds. Beyond this, as several of the other essays in the collection recognize, the inheritors of the phenomenological tradition, including Heidegger, developed a variety of methods of hermeneutic interpretation and existential reflection that move beyond the Husserlian assumption of the primacy of subjectivity. Meanwhile, the Vienna Circle’s original method of linguistic analysis through the specification of the formal “logic of language”

has few, if any, adherents left today; and the various and diverse descendents of this original project of “language analysis” hardly agree on any single picture of the nature of language, or indeed how its philosophical clarification should be pursued. Moreover, this historical situation makes the attempt to juxtapose “language philosophy” with “consciousness philosophy” also conceptually confused. For as Carl Friedrich Gethmann points out in his contribution, it is today almost impossible to say with any clarity what “language” now means to the many different philosophers, and projects, that appeal to it (and one could certainly say the same about “consciousness”). At the same time, both phenomenologists and analytic philosophers have found good grounds to dispute the assumption that our understanding of logic, meaning, experience, and thought must rely upon any single foundation in “first philosophy,” whether oriented toward consciousness or toward language. Rather than further pursuing the vexed question of the “primacy” of language over consciousness or consciousness over language, therefore, it may be more productive to (as most of the essays collected here indeed do) focus on more specific thematic questions in a pluralistic spirit, applying the best methods and insights that have been won over the twentieth century, whatever their historical source. One set of such questions concerns the nature of persons and the status of the ethical properties we attribute to them. As Sandra Ausborn-Brinker argues in “Phänomenologie als Sprachanalyse,” the conjunction of phenomenology and language analysis seems uniquely well qualified to help develop a theory of the nature of persons vis a vis objects and properties in the natural world. Here, Husserl’s detailed analyses of personhood and the phenomenology of embodiment (Leibhaftigkeit) can join with the results of midcentury ordinary language philosophy (including philosophers like Ryle, Austin, and Strawson) to produce a methodologically supple analysis that recognizes both language and the phenomena of human embodiment as definitive attributes of persons. Somewhat similarly, Thyen’s essay on Reinach and Wittgenstein interrogates the status of “ethical facts” such

246

as the “fact” that a murder (as opposed to a defensible killing) has taken place; as Thyen argues, both linguistic and phenomenological reflection on such “facts” requires ultimately that we situate them within a broader understanding of practices of moral consideration, attribution, and designation, what Wittgenstein discusses as the “language-games” that make up our particular “forms of life.” Further developing this, the final essay of the collection, Christian Beyer’s “Einfühlung and Intersubjektivität,” provides a helpful and suggestive analysis of the nature and possibility of intersubjectivity, drawing equally on phenomenological analysis and on the late Wittgenstein’s considerations about language and privacy. In this essay, it is particularly evident how the synthetic combination of methods drawn from phenomenology and language philosophy can help to resolve some of the deepest traditional problems of philosophy, in this case the “problem of other minds.” Progress with this traditional problem, Beyer argues, requires that we recognize that the relationship between the outward or “external” expression of an emotional state and the state itself is not one of inferential evidence but rather one of direct presentation: in seeing a friend’s sad facial expression, I directly perceive that person herself as sad and can share, to a certain extent at least, her state. This recognition was the cornerstone of the detailed and persuasive theory of “empathy” that the phenomenologist Edith Stein advanced, following and expanding on Husserl, as the constitutive basis for intersubjectivity tout court. This analysis can be enhanced and developed, Beyer argues, by use of Wittgenstein’s distinction between “symptoms” — signs which may or may not accurately represent the presence of an underlying syndrome — and “criteria,” which function more as practical definitions of the underlying phenomena. In this way, drawing on both phenomenological and language-analytic results, we can obtain a practical understanding of intersubjectivity that resolves or overcomes the problem of our knowledge of another’s mental state, as well as the leftover Cartesian picture of subjectivity that suggests this problem to begin with. Several of the essays focus particularly on Heidegger; these essays should help to give a

fuller picture of the development and influences on Heidegger’s project of interrogating the “meaning of Being,” including its relationship to the analysis of language. As noted in Carl Friedrich Gethmann’s “Heidegger und die Wende zur Sprache,” one of the main aims of the early Heidegger’s thought was to overcome what he saw as the Cartesian legacy of Husserl’s description of subjectivity; this led him to take, already in Being and Time, what can be, at least in some ways, described as a “linguistic turn” by recognizing hermeneutic interpretation as a privileged modality of phenomenological description and by theorizing language as at least equiprimordial with the other constitutive structures of human Dasein. Udo Tietz, in “Urteilen and Verstehen,” argues that Heidegger’s account of judgment, despite its development over the years leading up to Being and Time, remained grounded in a neo-Kantian picture of judgment which confused synthesis and predication, a picture which Heidegger might have overcome had he taken greater heed of the innovations of analytic philosophers. Though the interpretive argument does not seem completely convincing, Tietz’s article nevertheless raises important issues for the theory of judgment in relation to truth, understanding, and meaning that should be taken into account in any analysis of Heidegger’s legacy. In any case, given the long history of murky Heidegger exegesis that leaves its subject more obscure than it found him, the advent of clear, descriptive writing on Heidegger that relates his methods and results to those of analytic philosophy is particularly to be lauded. One hopes also that its practitioners will be able to move beyond the position of Being and Time to interrogate some of the very interesting developments of Heidegger’s consideration of language that occur only in his later philosophy, something that none of the essays collected here yet addresses. Paul LIVINGSTON Villanova University, USA 1. For instance, in 1962 Ryle wrote that his masterpiece The Concept of Mind “could be described as a sustained essay in phe-

247

nomenology, if you are at home with that label” (“Phenomenology vs. The Concept of Mind” in his Collected Papers, Vol. 1: Critical Essays. Bristol: Thoemmes, 1971.);

248

similarly, Austin referred to his “ordinary language” method as “linguistic phenomenology” (“A Plea for Excuses,” in his Philosophical Papers. Oxford: 1961.)

Grazer Philosophische Studien 73 (2006), 249–254.

Wolfgang HUEMER: The Constitution of Consciousness: A Study in Analytic Phenomenology. New York: Routledge 2005. viii + 127 pp. It is worthwhile to attend closely to the title of Huemer’s work, as it serves to indicate in considerable detail the concerns and strategy of his inquiry. As the first word after the initial definite article shows, his primary concern is with the notion of constitution as it is understood within phenomenology, but, to move to the third word, the constitution of consciousness has two connotations, both of which are in play for Huemer. In one sense, the “of ” may mean the work of constitution that is done by consciousness, which means, roughly, spelling out how various conscious processes conspire so as to “give” or “present” objects to consciousness. There is a second sense, however, where the “of ” signifies that consciousness is itself constituted, which involves, again roughly, spelling out how a series of processes come to be conscious processes. Huemer devotes considerable effort to articulating this sense of constitution as well. The key phrase in the subtitle of the work, “analytic phenomenology”, may appear to many to be almost a contradiction, as the phenomenological tradition stands squarely within the domain of philosophy rather lazily referred to as “continental philosophy”, whereas “analytic” denotes that kind of philosophy that is anything but continental. Huemer, to his credit, is unimpressed by these sorts of divisions and distinctions, and he moves freely between Husserlian phenomenology and the work of such thinkers as Wittgenstein, John McDowell, Robert Brandom, Wilfrid Sellars, and Hilary Putnam. Indeed, the other key figure in Huemer’s study, John Haugeland, is himself something of a hybrid, a result of crossbreeding Heidegger’s philosophy with contemporary philosophy of mind and cognitive science. As I’ll suggest below, not all of this borrowing, mixing, and matching is entirely satisfying or successful, but the desire to put all of these figures into conversation with one another is certainly to be applauded.

Huemer’s main strategy in the book is to delineate and combine two senses of “constitution”, which are both anchored in the phenomenological tradition. The first of these is perhaps the original version, from Edmund Husserl. The second is of more recent vintage, from John Haugeland, but Haugeland himself cites Heidegger as a source and inspiration. For Husserl, the notion of constitution is provided as an answer to the question, “How do conscious processes or experiences come to be of or about objects?” Thus, the notion of constitution is bound up with the notion of intentionality. As Huemer repeatedly and correctly warns, constitution does not mean literally creating or producing, as though the objects constituted in consciousness were somehow made of mental stuff (as, for example, some forms of phenomenalism or Berkeleyan idealism might maintain). As an answer to the question of the possibility of intentionality, constitution instead relates to the notions of meaning or sense: what is constituted in consciousness is not literally the object, but the object-as-meant or, even more awkwardly, whatever it is in the experience that directs that experience, and makes it be about, an object. So, for example, a perceptual object may be given via a series of what Husserl calls “adumbrations”, each of which “gives” the object as seen from a particular side. Each such adumbration is of or about the object, but each presentation is only partial, a particular “look” or view of the object. Phenomenologically speaking, the object is a series or system of adumbrations: conscious processes come to be of or about perceptual objects by being adumbrational presentations that are systematically integrated. Were experience to lack such systematic integration, then it would never be of or about perceptual objects. At the same time, this does not mean that the objects we perceive (my desk, the coffee cup, and so on) are literally systems of adumbrations. My desk is made out of wood, not “looks”; my coffee cup is made of clay, not perspectives. Adumbrations or “partial intentions” are only one element in Husserl’s account of con-

249

stitution. More is needed to articulate their systematic or integrated character. The principal form of integration that Huemer considers is the temporal integration of partial intentions. In particular, Huemer stresses Husserl’s idea that partial intentions continually undergo “retentional modification”, which means that each occurring moment of conscious experience does not disappear altogether but recedes into the experiential past. Mental episodes continue to be pushed further and further into the past, eventually fading to a “zero-point of liveliness”. Retention is one way in which adumbrational presentations of objects “hang together”, so as to give, or add up to, the presentation of an object at all, i.e. were consciousness to lack this kind of retentional-modificational structure, then consciousness would never be of or about objects at all. (If, for example, I did not retain my experience of each sounding note, then my auditory experience would never be of or about a melody, which is essentially a temporally extended object.) Retention plays an essential role not just with respect to the constitution of objects, but to the constitution of consciousness itself (recall the dual meaning of the “of ” in Huemer’s title). Huemer argues that for Husserl, the partial intentions that undergo continuous retentional modification ultimately form a kind of “holistic background which is necessary for having mental episodes in the first place” (53). This “holistic background” dynamically interacts with the experienced present, so as to affect the character and content of that experience: how I have experienced things in the past (and so those ever accumulating past experiences) partly determines how I experience things now. (Consider the difference between perceiving something for the first time versus when that thing has been experienced repeatedly: in the latter case, there is far less room for surprise or puzzlement.) The Husserlian account of constitution squarely concerns consciousness, whereas Haugeland’s account, following as it does in the footsteps of Heidegger, emphasizes the being of entities. For Haugeland, to constitute an entity is to “let it be” the entity that it is, i.e. to make it manifest as a particular

kind of being. Again, constitution is not creation or production, but instead a kind of revelatory determination (if that is the right way to construe Haugeland’s talk of letting beings be). Central to Haugeland’s account is the idea of constitutive standards, which are the standards entities have to “live up to” in order to count or qualify as those kinds of entities. The rules of chess is one of Haugeland’s favorite examples: the particular pieces — knights, kings, queens, pawns, and so on — are the pieces they are only in virtue of the rules of chess, along with the practice of playing chess, enforcing the rules, and so on. The rules of chess prescribe what is possible and impossible with respect to those pieces. For example, it is impossible for a bishop in chess to move along straight vertical or horizontal paths on the board. Of course, I can, when playing chess, physically move the bishop piece in such a way, but then I would not be playing chess correctly (and most likely my opponent would quickly inform me of my mistake). “Impossible” thus means something like unacceptable, intolerable, or even impermissible. To violate the constitutive standards of chess means no longer constituting chess pieces as pieces: a bishop piece moved along a straight path is no longer being used as a bishop, and so in some sense no longer is a bishop. For Haugeland, constitution ultimately depends upon commitment. For example, we can only let chess pieces be chess pieces if we remain committed to recognizing and enforcing the rules of chess. If we were to grow indifferent to how pieces are moved, which pieces do what, and so on, then there would no longer be chess pieces, i.e. nothing would any longer be constituted as a chess piece (or a game of chess). For Haugeland, this relation between constitution and commitment is generalizable beyond examples like pieces in a game. The natural sciences likewise have constitutive standards that are sustained by the diligent commitment of working scientists. Given the laws (rules) of physics, it is impossible for an electron to have a positive charge; having a negative charge is part of what it is to be an electron. That means that scientists will not tolerate or permit there to be a positively charged electron. What makes science different from chess are the ways in

250

the sciences that anomalies and disturbances can intrude on the “constituted domain”. If nature refuses to live up to the standards scientists are committed to, then those standards may themselves be called into question. Haugeland’s account of constitution, constitutive standards, and constituted domains does not give any pride of place to consciousness or experience. Instead, as even the chess example shows, the emphasis is on activity and practice: chess pieces are constituted not in any series of “mental episodes”, but in the day-to-day activity of playing chess, with its quotidian observing and enforcing of the rules. Constitution is thus for Haugeland more clearly a social notion, since the day-today playing of chess (and doing of science) is a communal activity involving the interaction, collaboration, and cooperation of many people over an extended period of time. As its Heideggerian roots would suggest, this notion of constitution makes an odd companion to Husserl’s. Huemer, however, wants to combine the insights of each account so as to yield a notion of constitution that bridges the gap between mind and world, or, as he sometimes puts it, to integrate the space of reasons with the space of causes without trying to reduce the one to the other. I will return to this aspiration on Huemer’s part momentarily. The social element of Haugeland’s account of constitution is one reason why Huemer appeals to it to supplement the standard Husserlian notion. Husserl developed his notion of constitution in conjunction with the phenomenological reduction, which involves a fully first-personal, even solipsistic, procedure. Another reason is that Huemer sees Haugeland to be developing an account of constitution that is no longer wedded to any form of idealism, as Husserl, at least in many of his writings, did (in Cartesian Meditations, for example, Husserl argues explicitly that phenomenology should be understood as a form of transcendental idealism). Indeed, Huemer sees Haugeland’s position to be a “realist account of constitution”, which “shows clearly that an account of constitution does not necessarily involve a form of idealism”, (p. 35) and again, that “Haugeland’s realist account of constitution shows clearly that constitution does not amount to an endorsement of

a form of idealism” (p. 38). Despite the clarity that Huemer discerns here, Haugeland’s “realism” is not at all clear to me. This is due in large part to the elusiveness of Haugeland’s talking of “letting be”. If we did not “let” beings be what they are, would they still be what they are? If the answer to this question is “yes”, then Haugeland is a realist, but then it is not entirely clear what work the notion of constitution is doing in his account vis-à-vis the being of entities. If, however, the answer here is “no”, then that would seem to push Haugeland’s position into the realm of idealism. Despite his assertions of clarity, Huemer himself appears to be divided on how to understand Haugeland’s view. Consider the following passage: Haugeland does not explain what objects are constituted from. According to his account, we can conceive objects only relative to constitutive standards. This account does not allow for explaining what a table is in terms of the parts that it is composed of, its atomic structure, for example. Seeing something as a table or as a bunch of atoms means only applying different constitutive standards. According to Haugeland we cannot conceive that there is some raw, unstructured matter, i.e., matter that is not constituted, that every object is composed of. (p. 36) The idea that “we can conceive objects only relative to constitutive standards” strikes me as difficult to reconcile with Huemer’s contention that Haugeland’s account of constitution “shows clearly that an account of constitution does not necessarily involve a form of idealism”. Such relativity to constitutive standards cuts against any clear endorsement of realism. This passage reveals difficulties not just in Huemer’s understanding of Haugeland’s account as a kind of realism, but a more general confusion as well. Here, Huemer complains that “Haugeland does not explain what objects are constituted from”, but given Huemer’s own caveats concerning the difference between constitution and creation, this is not something Haugeland should feel any need to explain. In other words, Huemer here appears

251

to be confusing constitution with construction, creation, or composition, as though what it is to be a particular kind of entity necessarily involved saying what it is made of. Consider again Haugeland’s example of chess: chess pieces can be made of all sorts of things — wood, ivory, plastic, metal, pixels on a computer screen, or symbols on a piece of paper (think about people who play chess with each other by mail) — but they are only constituted as chess pieces by being caught up in the playing of chess. What the pieces are made of is irrelevant to their constitution as chess pieces, and so it is beside the point to say what such pieces are “constituted from”. That is, a proper understanding of what chess pieces are shows the futility of “explaining” what they are “in terms of the parts” of which they are composed. Thus, any worries about “some raw, unstructured matter” are neither here nor there with respect to the notion of constitution Haugeland is interested in (nor is it relevant for Husserl’s notion, for that matter). That Huemer slips here shows how treacherous the notion of constitution is, and perhaps explains why even figures in the phenomenological tradition after Husserl deploy it so infrequently (for example, my edition of Merleau-Ponty’s Phenomenology of Perception does not have an index entry for “constitution”, and Hazel Barnes does not include the term in her Key to Special Terminology at the close of Being and Nothingness; Heidegger, who is of course working in greater proximity to Husserl, occasionally writes of constitution, but most centrally, he writes of “existential” constitution, where what is constituted is not objects, but “the there” of Dasein’s disclosedness). That the notion is slippery is one reason for its displacement, but there is a further reason as well, one that Huemer does not sufficiently attend to, in my view. As Huemer himself acknowledges, Husserl himself deploys the notion of constitution in conjunction with the performance of the phenomenological reduction, the “putting out of play” of all and any beliefs, presuppositions, and commitments concerning the reality of everything one’s intentional states are of or about, as well as any hypotheses about what their underlying causes might be. Such prescinding from any views on the sources and

success of conscious experience is required, according to Husserl, in order to fulfill the transcendental aspirations of phenomenology. That is, Husserl wants to use phenomenology to answer questions concerning the possibility of intentionality, to explain, in other words, how it is possible for consciousness to “reach” or “contact” an object at all. Husserl thought such questions could only be properly posed and answered within the purview of the reduction, since that is the only way to avoid the threat of circularity in one’s explanations. (Thus, Husserl would see Huemer’s claim that “perceptual experiences are caused by the objects they constitute” (p. 48) as either hopelessly circular or philosophically confused.) Husserl appeals to the notion of constitution in order to provide answers to such how-possible questions: consciousness reaches or contacts objects by constituting them within the stream of ongoing experience. Thus, Huemer, I believe, goes wrong in saying that, owing to the reduction, “the relation between mind and world, therefore, is out of reach for the phenomenologist”. (p. 57) For Husserl, the only philosophically satisfying way of making sense of that relation, of founding it, is via the phenomenological reduction, and hence the transcendental conception of constitution. (This is also why Husserl at times thought that transcendental idealism was the ultimate consequence of phenomenology, since there is no intelligible conception of an object apart from what is constituted in the purified stream of consciousness of the transcendental ego.) If one lifts the restrictions of the reduction, then it is not at all clear what philosophical work the notion of constitution is really doing. That is, if one simply helps oneself to a world replete with objects, social groups, and so on, then the transcendental dimensions of the idea of constitution are simply lost. In Huemer’s terms, the problem may be put like this: Huemer wants to make use of the notion of constitution in order to bridge the gap between mind and world, or, as he sometimes puts it, the space of causes and the space of reasons. But in rejecting the phenomenological reduction, and so the transcendental conception of constitution, both mind and world — objects, bodies, communities, lan-

252

guages, and so on — are already in view from the start, already constituted, and so not really constituted at all. It is not clear what gap there is to bridge, what relation there is to establish, if one adopts the starting point of Huemer’s revised conception of constitution. In particular, it is not clear what philosophical question can be appropriately asked and answered from this perspective. The situation is, for me, reminiscent of Quine’s rejection of the project of rational reconstruction once one has eliminated any kind of experientially basic unit, along with its corresponding atom of meaning. Once one has moved beyond those notions, to a more “naturalized” perspective, the whole idea of rationally constructing, and so justifying in some foundational way, our knowledge of the “external” world goes by the board. As Quine famously complains: “But why all this creative reconstruction, all this make-believe?” Without the phenomenological reduction, constitution may itself just be more such make-believe. Huemer’s skittishness about idealism is clearly an important factor in his attempt to “de-transcendentalize” the notion of constitution. My main worry is that without the context of transcendental philosophy, the notion of constitution cannot do the work it was meant to do. Though I do not have an argument to establish this, my hunch is that any serious, transcendentally charged notion of constitution is going to lead to some form of idealism. That said, I also find Huemer’s commitment to realism itself rather puzzling and unclear. The final chapter of the book tries to carve out a “Wittgensteinian” kind of realism that is neither metaphysical realism nor idealism. Given what Huemer says, however, it is not clear what this neither-nor position really is. For the most part, the position Huemer lays out looks an awful lot like unqualified realism, or at least it is not clear just why his view does not just collapse into it. Consider the following passage: The very fact that we are playing languagegames, thus, forces us to accept the existence of (at least some) physical objects. Hence, the general question of whether physical objects exist undermines the basis of the ground that determines its mean-

ing. In consequence, this general question is meaningless, and so are all positions that try to answer, rather than reject it. Therefore, both idealism and metaphysical realism are positions that cannot be formulated meaningfully. (p. 93) There are a number of inferences drawn in this passage, none of which strike me as particularly well-founded. For example, I do not see how the fact that we are “forced” to accept that there are physical objects renders the general question, which presumably is something like “Are there physical objects?”, meaningless. If we are so “forced”, that would suggest that we are likewise “forced” to answer “Yes” to any such question. But this is not a rejection of the question, which is what Huemer claims to want. Perhaps his point is that there can be no general doubt about the existence of physical objects, since we are always “forced” to accept them, but that again would not show that the question concerning their existence was itself meaningless. But notice that if we do not reject the question and instead acknowledge the inevitability of just one answer, that would make realists of us all, indeed metaphysical realists. We can see this same tension two pages later, when Huemer claims, first, that “if there were no physical objects, we could not have mental episodes and consequently could not come to wonder whether physical objects exist”. (p. 95) He then concludes: “Mimicking Wittgenstein’s way of speaking, we can state that the question whether physical objects exist cannot be meaningfully thought”. Notice that the first sentence includes, as the antecedent to a conditional, the subjunctive appeal to the nonexistence of physical objects. According to the second sentence, such a situation is unthinkable, and so Huemer, by his own lights, is not in a position to assert any such conditional. However, since Huemer does assert such a conditional, which presupposes that the antecedent is at least meaningful, then again it would seem that a question that asks after the existence of physical objects is itself meaningful. At most, Huemer has shown that there is only one answer we can ever want or hope to give, but that does not yet rule out the question.

253

Rather than eliminating the debate, Huemer’s arguments instead seem to show that one side will always win. David R. CERBONE West Virginia University

254

Grazer Philosophische Studien 73 (2006), 255–258.

Fabrice CORREIA: Existential Dependence and Cognate Notions. München: Philosophia Verlag 2005. 171 pp. ISBN 3-88405088-5. € 68.00. Fabrice Correia’s excellent book presents a systematic account of certain modal and ontological relationships in terms of which various subtly different notions of existential dependence may be explicated. These notions are central to many projects in analytical metaphysics and the philosophy of mind and as such need to be understood as clearly as possible. For example, substance ontologists typically attempt to characterize the category of substance in terms of some suitable notion of existential independence. Again, physicalists in the philosophy of mind frequently try to explain the relationship between mental and physical properties in terms of the notion of supervenience, which is usually taken to imply some kind of existential dependence of the former properties upon the latter. Both of these examples, along with many others, are taken up and explored by Correia in the course of his book. Although there have been other attempts in recent years to provide a rigorous account of the chief varieties of existential dependence, Correia’s treatment of the topic is by far the most thoroughgoing and systematic to date. Correia critically examines some of these previous approaches in Chapter 2 before introducing his own in Chapter 3. In particular, he raises objections to three main alternative accounts, which he calls the “modal-existential” approach, the “purely essentialist” approach, and the “essentialist-existential” approach. The first of these approaches has had numerous advocates. The second Correia attributes to me and the third he attributes, independently, to Kit Fine and to Kevin Mulligan, Peter Simons and Barry Smith. In each case, Correia argues that the approach in question fails to capture certain strong pre-theoretical intuitions concerning relationships of existential dependence between putative entities of various ontological categories. Correia’s own “foundational” approach,

as he calls it, appeals to a generalized notion of grounding — which he takes to be primitive — along with various more specific notions of grounding. He introduces, in Chapter 3, some relatively simple formal apparatus which enables him to represent perspicuously the basic principles that he takes to govern grounding relations. In Chapter 4, Correia elucidates these notions and their use in defining what he calls the “simple foundation” of one entity in another, with the help of detailed illustrative examples. Here is his proposed definition: “x is [simply] founded in y iff it is impossible that x exists without being based on y — i.e. (Ex Š x B y)”, where “An object x is based on an object y … when the fact that x exists is partly grounded in some fact about y” (p. 66). Later, in Chapter 5, Correia provides an analysis of “generic” dependence and other species of existential dependence, illustrating the importance of the former by discussing its role in debates between “Aristotelians” and “Platonists” concerning the nature of universals. He also examines in some detail Husserl’s conception of foundation, comparing it (unfavourably) with his own. Subsequently, Correia develops the distinction between non-temporal existential dependence and various temporalized forms of dependence, applying it to the definition of substance. In a final chapter, Chapter 6, he analyses the cognate notion of supervenience with the aid of his foundational approach to existential dependence and compares this analysis with others to be found in the voluminous literature on supervenience. Correia’s formal treatment of the various themes dealt with in the book exploits a logic of modality and essence which is provided with a philosophical motivation in Chapter 1 and is systematically formulated in an appendix. The book is extremely clearly written and meticulous in its presentation and organization. A consistently high standard of precision and rigour is maintained throughout, both in formal and in philosophical analysis and argumentation. Correia’s critiques of alternative approaches are well-informed and generally judicious and the theory that he defends is

255

cogently argued for. While building constructively on the work of others in many places, he develops an original and powerful system of ideas which makes an important contribution to a fundamental area of metaphysics and ontology — one that has, until recently, been inadequately explored. Since Correia clearly thinks it important for the success of his own project that rival accounts such as my own can be seen to be flawed, I shall conclude this note by briefly examining the objection that he raises against my so-called “purely essentialist” approach. My approach appeals to the notion of what I call identity-dependence, which I define as follows: x is identity-dependent on y iff necessarily, there is a function f such that it is part of the essence of x that x is the f of y.1 I claim, crucially, that identity-dependence entails what Correia calls “m-necessitation”, i.e., that if x is identity-dependent on y, then, necessarily, if x exists, then y exists. In support of this entailment thesis, I argue thus: “[S]ince for x to exist is for there to be something identical with x …, x cannot exist unless everything upon which x’s identity depends also exists”,2 and I illustrate this with examples — such as that of an assassination and the person assassinated. Thus, for an assassination to exist, there must be some particular assassination that it is (i.e., that it is identical with), and which assassination it is will depend on which person is assassinated. But if there is no such person, then it is undetermined which assassination it supposedly is, and so no such assassination exists. Correia, however, thinks that there are credible counter-examples to my entailment thesis. By contrast, he maintains that, according to his own definition of simple foundation, “unlike identity-dependence foundation has the desired property of entailing m-necessitation” (p. 67). Here is one such putative counter-example.3 Correia suggests that the property IS of being identical with Socrates is identitydependent on Socrates and yet can exist even though Socrates does not exist, given that IS is a necessary existent and Socrates is a contingent existent. However, in my view, properties are not, in general, necessary existents (I hold that they exist only in worlds in which they are exemplified) and, certainly, I see no

reason to suppose that IS is. In any case, it may be doubted whether IS, if indeed it is a genuine property, is identity-dependent on Socrates: for, plausibly, if it is a genuine property at all, it is a haecceity and as such an individuator of Socrates and so not something to which Socrates is ontologically prior.4 Correia has other putative counter-examples to my entailment thesis involving disjunctive sums and disjunctive facts. Since I don’t believe in either disjunctive sums or disjunctive facts (at least as Correia seems to conceive of the latter), these “counterexamples” are, in my view, unconvincing. With regard to disjunctive sums, these seem to be in no better shape than disjunctive sets. Is there a “disjunctive set” whose members are either Mars or Venus? Certainly, there is such a thing as the set of all and only those things x such that either x = Mars or x = Venus, i.e. {x: x = Mars › x = Venus}, but this is just the set {Mars, Venus}. The disjunctive sum of Socrates and Plato would be something that is indistinguishable from Socrates in some worlds, indistinguishable from Plato in others, and indistinguishable from the (conjunctive) sum of Socrates and Plato in yet others — and which exists in no other worlds. So there is no world in which it is not indistinguishable from something else. I don’t believe that there can be any such thing. As for facts, to the extent that I am in favour of them at all, I incline towards the Plantingan view that facts are states of affairs which obtain and that states of affairs are necessary existents. I could then happily accept that there are disjunctive states of affairs and that such a state of affairs is identity-dependent on each of its disjuncts: but this won’t provide a counterexample to the thesis that identity-dependence entails mnecessitation, precisely because states of affairs thus conceived are necessary existents. A final putative counter-example offered by Correia is that of my present thought about Pegasus: Correia maintains that “we can hold with some plausibility that [this] thought is identity-dependent on Pegasus” (p. 49), even though Pegasus does not exist. However, I would say that, if there really is such a thought, then in fact it is either a purely descriptive thought or else is one containing as a constituent Pegasus’s individual concept, where the latter is tak-

256

en not to be identity-dependent on Pegasus. If it really were an object-dependent thought, then it couldn’t exist unless Pegasus did. (Of course, it may seem to me that I am having such an object-dependent thought when really I am not, as in the case of apparent demonstrative thoughts about hallucinatory objects.) All of Correia’s alleged counterexamples to my entailment thesis are indeed, as he concedes, “somewhat baroque”.5 He takes these examples to defeat my “premise that where F is a function, for every x, the F of x cannot exist unless x exists” (p. 48, n. 13). But this premise simply appeals to the standard conception of a function, according to which a function is undefined for values for which it lacks an argument. (What is the length of Santa Claus’s beard, given that Santa Claus, and hence his beard, doesn’t exist?) Of course, Correia has already made it clear that his quantifiers are possibilist, not actualist,6 which rather stacks the cards against my position from the outset. On my (actualist) view, something doesn’t exist at all unless it actually exists. For me, a singular term which lacks a referent in this, the actual world, cannot have one in some other, merely possible, world. For me, to say that something exists at all is just to say that there is actually something with which it is identical. I should mention that Correia thinks that the force of his alleged counter-examples doesn’t depend on our accepting that they really are genuine counter-examples, but only on our accepting that some metaphysician could reasonably take them to be counter-examples. This is in line with his general aim to explicate the notion(s) of existential dependence in a metaphysical neutral fashion (though how his preference for possibilist quantifiers can be taken to be consistent with such neutrality is not entirely clear to me). In his view, “[A]ny account of the various notions of existential dependence should be compatible with any viable, i.e. non-absurd, i.e. dialectically possible metaphysical view” (p. 10). And he acknowledges that my own account of existential dependence does not aim at such neutrality and hence that he “will not take [his] objections to establish that Lowe’s characterization is inadequate given his particular purposes” (p. 46, n. 12). In defence of my own non-neutrality in this matter, I would only

say that, unlike Correia,7 I take the explication of existential dependence not to be a mere exercise in conceptual analysis, but a substantive contribution to fundamental metaphysics: and, concerning the truths of fundamental metaphysics we cannot, if we are realists, adopt a neutral stance. In any case, how does Correia’s own entailment thesis fare, given his policy of metaphysical neutrality? As I mentioned earlier, he contends that, “unlike identity-dependence foundation has the desired property of entailing m-necessitation” (p. 67). Correia says that this “follows from the fact that grounding is factive … and [that] … for something to be a base of x, it must have an ‘existence-entailing’ feature which explains the existence of x” (p. 67). But I confess that I don’t really see how, if Correia’s putative counter-examples work against my own entailment thesis, they or similar ones won’t work equally against his. Why, for instance, shouldn’t some metaphysician claim that the property of being identical with Pegasus is “based” on Pegasus, i.e. that “the fact that the property of being identical with Pegasus exists is partly grounded in some fact about Pegasus” — for example, that it is partly grounded in the fact that Pegasus could exist — even though Pegasus does not exist? Is this an incoherent or absurd claim, according to Correia, and if so, why? But if it is not an incoherent or absurd claim, then it would seem that such a metaphysician could coherently maintain that the property of being identical with Pegasus is, in Correia’s sense, founded in Pegasus even though that property exists and Pegasus does not — thus contradicting Correia’s claim that foundation entails m-necessitation. Concerning Correia’s positive proposals, I would further ask only this: is it really satisfactory to take the notion of “metaphysical grounding” as primitive, as Correia does?8 The reason why I abandoned my own earlier attempts along such lines is precisely that leaving such a notion undefined seems intolerably obscure, even after all of the various elucidatory points have been made and illustrative examples have been provided.9

257

E. Jonathan LOWE University of Durham

1. See my The Possibility of Metaphysics: Substance, Identity, and Time (Oxford: Clarendon Press, 1998), p. 149. 2. Ibid., p. 150. 3. See Correia, Existential Dependence, p. 48. 4. On this point, compare Gary S. Rosenkrantz, Haecceity: An Ontological Essay (Dordrecht: Kluwer, 1993) and see further my “Individuation”, in M. J. Loux and D. W. Zimmerman (eds), The Oxford Handbook of Metaphysics (Oxford: Clar-

endon Press, 2003). 5. Correia, Existential Dependence, p. 49. 6. Ibid., p. 19. 7. Correia proclaims at the very outset of his book, that his “investigation is a conceptual one … not one of metaphysics … [but] of meta-metaphysics”: ibid., p. 9. 8. Ibid., p. 57. 9. For one of those attempts, see my “Primitive Substances”, Philosophy and Phenomenological Research 54 (1994), pp. 531–52.

258

Grazer Philosophische Studien 73 (2006), 259–264.

Albert CASULLO: A Priori Justification. New York: Oxford University Press 2003. 249 pp. ISBN 0–19–511505–8. In contemporary epistemology there has been a resurging interest in a priori knowledge. Over the past ten years or so innovative contributions to this topic have been put forth by philosophers like George Bealer, Paul Boghossian, Laurence BonJour, Tyler Burge, Albert Casullo, Hartry Field, Alvin Goldman, Christopher Peacocke, Alvin Plantinga, Ernest Sosa and others. In his excellent book, A Priori Justification, Albert Casullo provides a comprehensive treatment of the main issues concerning the a priori which is sensitive to the main recent contributions to the topic. Casullo’s book is organized into three parts devoted to the following questions: (i) What is a priori knowledge? (ii) Is there a priori knowledge? (iii) What are the relationships between the a priori and semantic and metaphysical notions? In this review I want to focus on the first two parts of Casullo’s book. The first part of Casullo’s study consists of three chapters and is (as already mentioned) concerned with the question how the notion of a priori knowledge is to be explicated. This question is of utmost importance, since the notion has often been conflated with unrevisability, certainty, necessity or analyticity. In the first chapter, Casullo provides a survey and critique of different analyses. Casullo begins his discussion with the question what the primary target of an analysis of the concept of a priori knowledge should be. He argues convincingly that the notion of a priori knowledge should be explicated “reductively” via the notion of a priori justification. The primary target of an analysis should accordingly be the notion of a priori justification, and not the notion of a priori knowledge. Casullo’s first rationale for this order of explication is that a “non-reductive” analysis, which aims at analyzing the notion of a priori knowledge, would have to address the Gettier-problem. But the Gettier-problem has, as Casullo rightly contends, no obvious bearing on the question of what distinguishes empirical from

non-empirical knowledge. An analysis of a priori knowledge via the notion of a priori justification circumvents the question of how the Gettier-problem should be solved and should therefore be preferred. Furthermore, it is uncontroversial, as Casullo maintains, that knowledge entails truth, whereas it is controversial whether justification entails truth. Since this distinction can be easily overlooked when one pursues the “non-reductive” approach to analyzing the notion of a priori knowledge, Casullo pleads for the “reductive” approach. Next, Casullo presents an illuminating taxonomy of analyses and corresponding conditions. There are, according to Casullo, two main types of analyses, namely epistemic and non-epistemic ones. Epistemic analyses consist of epistemic conditions, like strength, source and defeasibility-conditions, whereas non-epistemic analyses contain semantic or metaphysical conditions, like analyticity or necessity. Since the notion of a priori justification is an epistemic notion, Casullo is convinced that it should be explicated via epistemic concepts and not semantic or metaphysical ones. The latter concepts do not highlight what is distinctive about this kind of justification. Although Casullo dismisses non-epistemic analyses, he discusses epistemic analyses that contain necessity as a necessary, though not sufficient condition for a priori knowledge. He considers the views of BonJour and Plantinga who treat necessity as part of an epistemic condition, and of Chisholm, who thinks of necessity as an independent non-epistemic condition for a priori knowledge. Casullo rejects all of these analyses. He then examines strength and defeasibility conditions and finally source conditions. Casullo argues that strength and defeasibility-conditions are not essential for the a priori and that an analysis of a priori justification in terms of the sources by which a given belief is justified is the correct one. According to Casullo, a proposition is justified a priori if it is justified by some non-experiential source. I think that this analysis is correct, but Casullo fails to spell out what a source of justification exactly is. Furthermore, he doesn’t point out how

259

sources of justification are related to reasons or grounds for a given belief. Although these questions are general questions of epistemology and are not confined to the domain of the a priori, I think that a defendant of an analysis of a priori knowledge in terms of its sources should be explicit on these points. The second chapter is devoted to a defense of this “minimal” analysis, according to which a proposition is justified a priori if it is justified by some non-experiential source. Casullo presents several considerations that favor this minimal analysis over one that combines this source condition with a “weak” defeasibility condition. According to this richer analysis, a proposition that is justified by a non-experiential source cannot be defeated by experience. Casullo argues inter alia that this second analysis violates a plausible condition of adequacy of an analysis of a priori knowledge, namely that an a priori justified proposition should also be capable of empirical justification. The third chapter is devoted to the question whether a priori justified beliefs are fallible. Casullo presents here different concepts of infallibility and (correspondingly) fallibility. He argues that there are two main senses of infallibility (and fallibility), namely Cartesian infallibility and Peircean infallibility. Cartesian infallibility concerns the relation between justification and truth, whereas Peircean infallibility concerns the relation between justification and defeasibility. He then distinguishes between different versions of each type of infallibility and investigates the relationships between them. Finally, he critizes the views of Donna Summerfield and Bob Hale, who proposed analyses of a priori knowledge that are supposed to be fallible. The second part of Casullo’s book is centered on the question whether there exists a priori knowledge. This part of Casullos study consists of two chapters in which positive and negative arguments for the existence of a priori knowledge are evaluated and a chapter in which Casullo’s own approach is articulated. Casullo aims to establish the conclusion that neither the positive nor the negative arguments adequately support their conclusion. He discusses in the fourth chapter three types of arguments that purport to show that

there exists a priori knowledge. These arguments fall into three categories: (i) “conceptual” arguments, (ii) “criterial” arguments and (iii) “deficiency” arguments. The first type of argument consists of the idea that there are items of knowledge that satisfy the conditions that were laid down in the analysis of the concept of a priori knowledge. Casullo discusses two proposals along these lines. First, he examines Hilary Putnam’s argument from his papers “There Is at Least One A Priori Truth” and “Analyticity and Necessity: Beyond Wittgenstein and Quine”. Putnam maintains that there are propositions that can never be rationally given up, namely the minimal principle of contradiction (i. e. that not every statement is true and false). Since Putnam thinks that an a priori justified proposition is one that cannot be rationally given up, there are items of knowledge that are justified a priori. Casullo convincingly shows that Putnam’s argument fails. He objects that the analysis of a priori knowledge that Putnam endorses does not entail that a proposition is justified a priori in the sense of Casullo’s explication, i. e. justified independently of experience. Being rationally unrevisable does not preclude that a proposition is justified by empirical evidence. Hence, Putnam’s argument is not sound. Hartry Field puts forth the second conceptual argument that Casullo discusses. Field maintains that the principles of classical logic are “strongly” justified a priori, i. e. they are not revisable on empirical grounds. Field argues that the principles of classical logic are so deeply embedded in our evidential system that it makes no sense to revise them without giving up the evidential system itself. Casullo objects that Field’s overall position leads to the consequence that the principles of classical logic cannot be justified nor defeated by any type of evidence. But this does not show, as Casullo observes, that these principles are justified a priori. Hence, both “conceptual” arguments fail. Casullo then tries to show that “criterial” arguments fare no better than “conceptual” arguments. “Criterial” arguments proceed to identify certain features of propositions (e. g. necessity or certainty) that cannot be accounted for by empirical modes of justification. This in turn shows that these prop-

260

ositions must be justified a priori. Casullo argues convincingly that these arguments fail. The fourth chapter closes with a discussion of BonJour’s arguments against “radical” empiricism. BonJour’s “deficiency” arguments purport to make the case for the a priori by showing that radical empiricist accounts of justification lead to a certain form of skepticism. BonJour maintains, for example, that certain empirical propositions that are general in their content cannot be justified by experience alone. These propositions can, according to BonJour, only be justified inferentially by experience and certain principles of inference. Since the radical empiricist accepts only direct experience as a source of justification, he cannot account for the justification of these principles of inference and hence these general kinds of propositions. Casullo rebuts BonJour’s arguments by showing that BonJour’s own position faces similar problems. Casullo insists first of all that there are propositions whose content goes beyond the content of what can be justified by rational insight and that accordingly, these propositions cannot be justified by rational insight alone. But what do these kinds of propositions look like? I think that this objection is not intelligible at all, since Casullo fails to provide an example of these kinds of propositions. Casullo further complains that BonJour is not offering a convincing explanation of rational intuition. BonJour advocates the view that a thinker who employs rational intuition must have some grasp of universals. But according to Casullo, BonJour’s explanation of this grasp is only metaphorical and cannot provide a basis for accepting rational intuition as a genuine source of justification. I think this objection is convincing but it seems that BonJour could modify his position to circumvent it. BonJour could, for instance, drop this kind of explanation of rational intuition and leave it at an intuitive level. Furthermore, there exist other versions of rationalism (e.g. Bealer’s) in which claims like BonJour’s are not endorsed. Thus, Casullo’s attack only hits BonJour’s position but not all rationalist conceptions of the a priori. If BonJour’s attack on radical empiricism were coupled with a different theory of a priori justification, Casullo’s argument would not survive.

The fifth chapter examines arguments that purport to show that there is no a priori knowledge. These arguments fall also into three types: (i) “conceptual” arguments, (ii) “radical empiricist” accounts, and (iii) arguments from philosophical naturalism. Conceptual arguments against the existence of a priori knowledge purport to show that there are no items of knowledge that satisfy the putative conditions for a priori knowledge. Putnam and Philip Kitcher proposed arguments along these lines. Both philosophers allege that the concept of a priori knowledge includes an unrevisability condition, according to which an a priori justified proposition is one that cannot be revised on any type of evidence (Putnam) or on empirical evidence (Kitcher). Casullo has argued in the first part of his book that unrevisability-conditions are not constitutive for the concept of a priori knowledge, he now shows that a priori justified beliefs do not entail such conditions. He constructs several thought experiments that establish this conclusion. Hence, these kinds of arguments are not successful. Next, he considers radical empiricist accounts that aim to show that propositions alleged to be known only a priori are known empirically. Casullo discusses the views of Mill, Kitcher and Quine who have put forth considerations of this sort (especially for the propositions of arithmetic). Mill tried to show, for instance, that the propositions of arithmetic or geometry can be justified by observation and the employment of the rule of enumerative induction. Casullo responds to this maneuver by arguing that the fact that these propositions can be justified empirically does not rule out the possibility that they can also be justified a priori. There exists the phenomenon of epistemic overdetermination, i. e. there can be evidence for a given proposition from different sources. Mill could rule out this response on grounds of simplicity. But to this Casullo replies, in my eyes convincingly, that the goal of an epistemological theory is not to offer the simplest account but an accurate one. The final set of arguments that Casullo discusses against the existence of a priori knowledge purport to show that a priori knowledge is not compatible with philosoph-

261

ical naturalism. The general idea of these types of arguments is due to Paul Benacerraf. Benacerraf has argued in his influential paper “Mathematical Truth” that there exists a problem as of how to account for our knowledge of abstract entities like numbers, when knowledge requires some kind of causal contact with the entities a proposition is about. Since mathematical entities are causally inert, a thinker cannot stand in a causal relation to these entities and hence cannot have knowledge of them. Casullo discusses two main versions of naturalism in relation to the a priori. The first version of naturalism tries to accommodate knowledge of abstract entities within a naturalist framework. The second version of naturalism challenges the existence of the a priori by arguing for some kind of causal constraint on knowledge or justification. Casullo first discusses the view of Mark Steiner who advocates a modified version of a causal theory of knowledge that purports to be compatible with knowledge of abstract entities. On Steiner’s theory, for a subject to know that a sentence s is true, the sentence s must be used in a causal explanation of why the subject knows that s is true. Since, on Steiner’s view, no causal relations between a subject and abstract entities are needed, he can account for the knowledge of abstract entities without giving up a causal requirement. Casullo argues that Steiner’s attempt to reconcile a causal theory of knowledge with Platonism fails, since Steiner’s causal constraint on knowledge is too liberal to be of any epistemological significance. He objects that Steiner’s proposal is subject to counterexamples, because any ascription of a belief in a mathematical statement s involves the statement s in some causal explanation of why the belief is held. He then discusses two versions of reliabilism that also aim to be compatible with knowledge of abstract entities, namely the thermometer model of knowledge put forth by David Armstrong and process reliabilism mainly advocated by Alvin Goldman. Casullo argues that both theories are plagued with difficulties. Armstrong’s theory can for instance not accommodate knowledge of abstract entities. Process reliabilism is confronted with the problem of how to explain the reliability of psychological processes that justify

beliefs about abstract entities. This latter fact generates suspicion concerning the reliability of these processes and hence introduces potential defeaters. Casullo finally considers proposals that aim to provide support for a causal constraint on knowledge. He examines the view of Penelope Maddy who argues that a causal constraint is underwritten by the practice of mathematicians. Casullo dismisses her argument because she cannot provide adequate support for the thesis that the practice of mathematicians supports a causal constraint. In chapter six Casullo aims to provide a solution to the dispute between those who embrace and those who reject the a priori. Since both arguments for and against the a priori fail, as Casullo has shown in the previous two chapters, Casullo argues that an alternative approach is needed. Before he lays down his positive proposal, he addresses a question that was postponed in the first part of his book, namely what experience is. He aims to show that the ordinary approach to analyzing the concept of experience fails. Among the proposals that he rebuts are Plantinga’s, Chisholm’s, McGinn’s and BonJour’s. I think that Casullo’s case against BonJour’s explication of the notion of experience is not convincing. BonJour advocates an analysis of the notion of experience according to which it “should be understood to include any sort of process that is perceptual in the broad sense of (a) being a causally conditioned response to particular, contingent features of the world and (b) yielding doxastic states that have as their content putative information concerning such particular, contingent features of the actual world as contrasted with other possible worlds”. Casullo objects to this analysis that it is vulnerable to counterexamples. He distinguishes between two readings of BonJour’s condition (a). On the first reading, the “contingent features” that are causally responsible for the belief in a certain proposition are meant to be any entities that are to be taken into account for explaining the belief in a certain proposition. But this has the consequence that an act of rational insight would be an instance of an experiential process, since the proposition itself and the act of considering it (which are essential for rational insight

262

to work) are also causally responsible for the belief-state. I think that this is a spurious objection to BonJour, since propositions are abstract entities and are not accessible via perception. Hence, in the case of rational insight, condition (a) is, contra Casullo, on this reading not fulfilled. On the second reading, the particular features of BonJour’s condition (a) are meant to be the entities that are referred to in condition (b). But this also has, as Casullo claims, the consequence that a process of rational insight satisfies condition (a). According to BonJour, if I am to be justified, for instance, in the proposition that nothing can be red and green all over at the same time, the properties redness and greenness must be capable of influencing my state of mind. Since properties are abstract entities and are accordingly causally inert, the obvious alternative is that concrete objects which instantiate these properties must be involved in some causal chain of events that is responsible for the content of my state of mind. I think the fact that in BonJour’s construal rational insight involves some kind of causal contact with “particular contingent features” does not show that rational insight satisfies the conditions laid down in the analysis of the concept of experience. Because on this reading, clause (b) is not satisfied, since the proposition that nothing can be red and green all over is not about these particular contingent features. The particular contingent features that are referred to in clause (a) of BonJour’s explication, according to Casullo, are not identical to those referred to in clause (b). Additionally, the causal contact to these “contingent features” of the world has only the status of an enabling condition for rational insight (analogously to having experiences in order to acquire a language and being able to understand a proposition that is justified a priori). Causal contact to objects that instantiate certain universals does not play some justificatory role. Finally, even if on BonJour’s construal of rational insight there would be some kind of causal contact with contingent “features” of the world involved, this would not go against BonJour’s explication of the concept of experience, but rather against his explanation of rational insight. If one would adopt an alternative construal of

rational insight, then Casullo’s counterexample would not stand. Although I do not think that Casullo’s case against BonJour is convincing, Casullo takes the failure of these analyses as evidence that an alternative approach is needed. His approach to analyzing the notion of experience consists of the idea that “experience” is a term for a natural kind. The notion of experience is according to this view introduced by local paradigms, but the nature of experience that determines the reference of the term “experience” can only be discovered by empirical investigations. If such empirical investigations are successful then the question as to which processes are non-experiential, if there should be any, and why they are non-experiential can be answered. According to Casullo’s novel proposal, empirical considerations are also salient in determining whether the sources of a priori knowledge are truth-conducive and whether their truthconduciveness can be explained in some way. Finally, Casullo considers objections to his view, which he rebuts successfully. In sum, Casullo’s book is a must read for anyone who is interested in a priori knowledge. The first part contains the most comprehensive treatment of the present literature questioning how the notion of a priori knowledge is to be explicated. Casullo discusses all the major proposals and argues very thoroughly for his own position. It should therefore be carefully read by anyone who is seriously interested in the question what a priori knowledge is. The second part of his book, in which positive and negative arguments for the existence of a priori knowledge are evaluated and Casullo’s own proposals are presented, is also a very important contribution to the present debate. Although Casullo discusses all the main proposals, he has unfortunately omitted a discussion of the views of Peacocke and Boghossian. Peacocke and Boghossian have tried to revive the idea that the phenomenon of a priori justification can be explained in terms of meanings or concepts. Boghossian has, for instance, tried to explain how a belief can be justified a priori via a rehabilitation of a Fregean conception of analyticity, whereas Peacocke has tried to give an account of a priori justification in terms of concepts. Despite this shortcoming, Casul-

263

lo’s book is one of the most important contributions to this topic. Martin GRAJNER Technische Universität Dresden

264

Grazer Philosophische Studien 73 (2006), 265–267.

R. Jay WALLACE, Philip PETTIT, Samuel SCHEFFLER, Michael SMITH eds.: Reason and Value. Themes from the Moral Philosophy of Joseph Raz. Oxford: Clarendon Press 2004. ix + 429 pp. ISBN 0-19926188-1. $ 35,00. Joseph Raz is undoubtedly one of the main thinkers in contemporary practical philosophy. His work on freedom, authority, equality, practical reason, value, and the relations between law, legality, and morality had a substantial impact on the subject and shaped the subsequent debates. The volume Reason and Value. Themes from the Moral Philosophy of Joseph Raz is a collection of essays by prominent philosophers presented originally at a conference on Joseph Raz’s work at Columbia University in 2002. The papers in the volume not only elucidate important aspects of the work of Raz, the authors present their own theories on several of the ideas and topics discussed by Raz. It is beyond the scope of this review to discuss all the papers; all papers are highly interesting and in combination they sum up to an impressive collection of practical philosophy. A main focus of several essays is Raz’s conception of reasons. The essays by Michael Bratman, John Broome, Jonathan Dancy, Ulrike Heuer, Harry Frankfurt, Peter Railton, Thomas Scanlon and Samuel Scheffler deal with this topic. For Raz practical reasons are embedded in evaluative structures. Reasons are connected with ends and activities that do have value. Values are shaped by social practices. Raz concedes that the reasons an individual has do not always prescribe a specific course of action as the uniquely rational one. This feature of Raz’s approach is criticized by Michael Bratman (Shared Valuing and Frameworks). Bratman objects that Raz’s approach is underdetermined. Judgements of value, Bratman argues, are not a sufficient normative guideline for leading one’s life. The criterion of cross-temporality applied to our background frameworks (intentions, plans,

and policies) offers according to Bratman a more promising route: the demand of crosstemporality amounts to a stronger normative control of deliberation than simply anchoring reasons in value judgments. This lack of criteria in Raz’s conception for singling out the normatively decisive reasons is also the theme in the essays of Ulrike Heuer and R. Jay Wallace. Heuer (Values and Reasons) argues that, though Raz offers a description of the kinds of reasons values are (i.e. reasons to preserve, not destroy, engage with, and acknowledge whatever is of value), he does not answer the problem that his account allows too many values and too many reasons as having potential strength in rational decision making. Raz, she criticizes, fails to develop criteria to put restrictions on the reasons we consider that also allow us to identify the strongest reason. R. Jay Wallace (The rightness of acts and the goodness of lives) poses the question whether Raz’s pluralist account of reasons can live up to the idea that moral considerations do have normative force, are ethically significant requirements. Wallace criticizes that Raz’s scheme — that what should be done is dependent on whether it is good for the agent — is not adequate. Morality, Wallace emphasises, structures practical deliberation and puts requirements on practical reflection; Raz does not meet this feature of morality and downplays the differences between moral requirements and the activities that are good for an agent like personal values, projects etc. Raz defends a value-based conception of reasons. Mere desires or wants cannot be reasons. Desires are directed to an object, and this specific directedness can according to Raz only be justified by value-based reasons. Once this value-based justification is in place, the fact that one wants or desires the object, does not add an additional reason. Ruth Chang challenges this thesis in her paper (Can desires provide reasons for action? ) She argues for a middle position between a desire-based and a value-based account of reasoning by trying to show that affective desires of the form ‘I feel like it’ can provide rea-

265

sons. Affective desires, as Chang points out, amount to reasons in cases where the value-based reasons pro and con are balanced and do not support a choice. She does not want to defend the thesis that all reasons are desire-based, but she wants to correct the established picture of desire-based and value-based accounts by adding a category of desires that can be reasons. The question raised by Chang’s paper is obviously in which respect her account is different from a value-based conception. If the choices based on affective desires are not completely arbitrary and idiosyncratic, then they must be connected to value-based considerations and, one might object, a value-based perspective is implicitly presupposed. Something similar, namely pointing out certain categories of reasons that Raz neglects, is pursued by Jonathan Dancy (Enticing Reasons). Dancy objects that Raz’s skepticism towards enticing reasons is misplaced. Dancy suggests to associate enticing reasons with promoting the most valuable, and he argues that this category should have a distinctive place besides those reasons that are tied to ‘ought’ and requirement. Enticing reasons, Dancy thinks, can be sufficient to make the action they recommend worth doing. John Broome (Reasons) criticizes that Raz understands normativity exclusively in terms of reasons. This way, Broome points out, we get a wrong and distorted picture of normativity. For Broome (a thesis he has defended in several other essays) the distinction between normative requirements and reasons is important. Normative requirements which are essential for normativity do have, Broome points out, a different logical structure from reasons. They are wide-scope oughts: the ‘ought’ refers to the whole expression (to phi, in case a fact X normatively requires you to phi). Reasons often do have narrow scope: certain facts (e.g. that x is a means for some end) do give you a reason, but the ‘ought’ is only referring to this aspect. Broome’s point is that practical reasoning must be directed to satisfying normative requirements, it cannot be reduced to weighing reasons that are narrow scope oughts. As valuable and important as Broome’s point is it would have been good if he had shown more clearly why this distinc-

tion cannot be integrated into Raz’s account and why the notion of a reason as such cannot be extended to wide-scope oughts. The essays by Harry Frankfurt and Philip Pettit / Michael Smith focus more on the deontological aspects in Raz’s practical philosophy. By drawing on his own work on wholeheartedness and being open to values like care or love Frankfurt criticizes in his contribution, (Disengaging Reason) that Raz assigns a greater significance to reason than it actually possesses. Frankfurt’s essay nicely illuminates the differences between a rational will conception and an account that holds that practical reflection is open to broader values. Philip Pettit and Michael Smith (The Truth in Deontology) argue that deontological constraints (Raz assumes that some normative constraints do have deontological status) do not rule out consequentialism. They show this by analogy with the rules of a game: some rules do have deontological status for the players; but within the space defined by these rules there is room for consequentialist considerations. The papers of Peter Railton, Samuel Scheffler and Thomas M. Scanlon analyze against the background of Raz’s account of reasons the difficulties of a reason-approach in general. Railton (How to Engage Reason: The Problem of Regress) discusses the problem that due to the regresses in practical justifications it becomes difficult to judge that someone is acting for a reason — it could also be a sequence of desiring, believing, recognizing conceptual connections etc. Reasoning cannot therefore be the sole ground of normativity. He then suggests a functional account (based on the notion of the self as a functional whole) as a solution of this problem. Scanlon in his article (Reasons: A Puzzling Duality?) deals with the problem how we can confer normative status on certain considerations by our choice of ends if reasons just have a normative status by favouring a certain way of acting. Scheffler (Projects, Relationships, and Reasons) discusses the relationship between valuing certain kinds of things and the recognition of reasons for action. He argues that the distinction between respect and engagement, as Raz draws it, does not suffice to distinguish wrong-making reasons from others. Project-dependent reasons

266

are reasons we can neglect whereas relationship-dependent reasons (Scheffler cites as an example our relation to our children) we cannot neglect so easily. The other essays, those of Seana Shiffrin, Michael Stocker, Michael Thompson and Donald H. Regan focus on topics different from the problem of reasons and reasoning. Stocker (Raz on the Intelligbility of Bad Acts) deals with the question whether Raz’s account can deal with the problem of bad action. Thompson (What Is It To Wrong Someone? A Puzzle about Justice) discusses the problem which reasons the just person is really drawing on in her action towards the other person, are these considerations of justice or just reasons due to being drawn to the other person with a certain attitude. Seana Shiffrin (Egalitarianism, ChoiceSensitivity, and Accomodation) takes up the question of egalitarianism and choice-sensitivity, i.e. whether egalitarians can really demand that people do have to bear the consequences of their voluntary choices. She does

not want to reject choice-sensitivity altogether, but, as she points out, there are areas where we accommodate choice-sensitive norms, e.g. the medical sphere, personal relationships, decisions in regard to one’s work, religion. Donald H. Regan (Why am I my Brother’s Keeper?) concentrates on Raz’s statement that morality is concerned with advancing people’s well-being. Regan discusses what sort of obligation this is, i.e. wether it is an obligation addressed to promote people’s well-being directly or an obligation to act in a principled way bringing about an increase in well-being as an indirect effect. The volume is not restricted to an exegesis of Raz’s theories. For the most part the authors present their own ideas on problems and topics that Raz has influentially discussed in his own work. This way the significance of Raz’s thinking is acknowledged in a highly respectful way.

267

Herlinde PAUER-STUDER University of Vienna

LISTE EINGELANGTER BÜCHER

(November 2005–Dezember 2006) A. Bottani, R. Davies (Hg.): Modes of Existence. Papers in Ontology and Philosophical Logic. Philosophical Research 5. Frankfurt/Main: Ontos, 2006, 237 pp. M. Erler: Platon. Beck’sche Reihe „Denker“ 573. München: C. H. Beck, 2006, 253 S. J. H. Y. Fehige: Die Geschlechtererosion des semantischen Realismus. Eine logischsemantische Untersuchung zum Begriff des biologischen Geschlechts. Paderborn: Mentis, 2006, 296 S. M. Forschner: Thomas von Aquin. Beck’sche Reihe „Denker“ 572. München: C. H. Beck, 2006, 238 S. R. Geiger: Dialektische Tugenden. Untersuchungen zur Gesprächsform in den Platonischen Dialogen. Mentis: Paderborn, 2006, 172 S. C. Illies: Philosophische Anthropologie im biologischen Zeitalter. Zur Konvergenz von Moral und Natur. stw 1743. Frankfurt/Main: Suhrkamp, 2006, 361 S. P. König: Giambattista Vico. Beck’sche Reihe „Denker“ 571. München: C. H. Beck, 2005, 156 S. F. Modenato: La conoscenza e l’oggetto in Alexius Meinong. Padova: Il Poligrafo, 2006, 427 S. C. U. Moulines: La philosophie des sciences. L’invention d’une discipline (fin XIXe –début XXIe siècle). Paris: Éditions rue d’Ulm/Presses de l’école normale supérieure, 2006, xii + 171 S. V. A. Munz: Satz und Sinn. Bemerkungen zur Sprachphilosophie Wittgensteins. Studien zur österreichischen Philosophie 39. Amsterdam–New York: Rodopi, 2005, 302 S. I. Persson: The Retreat of Reason. A Dilemma in the Philosophy of Life. Oxford: Clarendon, 2005, viii + 494 pp. G. Reuter: Bedeutungen und soziale Praktiken. Probleme des Sozialexternalismus und Perspektiven einer individualistischen Theorie. Paderborn: Mentis, 2006, 406 S. G. E. Rosado Haddock: A Critical Introduction to the Philosophy of Gottlob Frege. Aldershot: Ashgate, 2006, x + 157 pp. K. Sander/Jan St. Werner: Vorgemischte Welt. edition suhrkamp 2391. Frankfurt/ Main: Suhrkamp, 2005, 232 S. C. Strub: Vom freien Umgang mit Gepflogenheiten. Eine Perspektive auf die Praktische Philosophie nach Wittgenstein. Paderborn: Mentis, 2005, 114 S. Z. Torbov: Erinnerungen an Leonard Nelson (1925–1927). Herausgegeben, neu übersetzt und mit einer Einleitung versehen von Nikolay Milkov. Studien

und Materialien zur Geschichte der Philosophie 70. Hildesheim: Olms, 2005, LVIII + 206 S. C. Travis: Thought’s Footing. A Theme in Wittgenstein’s Philosophical Investigations. Oxford: Clarendon, 2006, 217 pp. J. Westerhoff: Ontological Categories. Oxford: Clarendon, 2005, xiv + 261 pp. M. Wolff: Einführung in die Logik. Beck’sche Reihe 1728. München, 2006, 195 S.

Philosophical Analysis, Volume 18

Jean-Maurice Monnoyer (Ed.)

Metaphysics and Truthmakers The essays collected in this volume concern the general question of truthmaking. Most of them also bear upon the metaphysical nature of truthmakers (moments, tropes, property-instances, Aristotelian substances, states of affairs, meanings or essences ? ). Taking as their starting point a famous seminal paper by K. Mulligan, P. Simons and B. Smith, as well as D. Armstrong’s outstanding contribution to the subject, they offer a fresh assay of the main concepts involved, in order to assess the explanatory value of truthmakers and truthmaker necessitarianism, and explore such delicate issues as contingent truth, bare possibility, tensed propositions, the ontological irreducibility of relations, the subsistence of facts and the epistemic role of negative truths. The collection as a whole provides a comprehensive and thought-provoking survey of the current debate about truthmaking theory and deserves to be read carefully by anyone interested in the relationship between language, thought and reality. Jean-Maurice MONNOYER : “Vingt ans après” Kevin MULLIGAN, Peter SIMONS & Barry SMITH : Truth-Makers (1984) Kevin MULLIGAN : Two Dogmas of Truthmaking Peter SIMONS : Truth in virtue of Meaning Barry SMITH & Jonathan SIMON : Truthmaker Explanations David ARMSTRONG : Truthmakers for Negative Truths and for truths of mere possibility Philipp KELLER : A World of Truthmakers David ARMSTRONG : Reply to Keller François CLEMENTZ : Relational Truthmakers Herbert HOCHBERG : “Is True” and “Makes true”, Two predicates without properties Frederic NEF : Senex erit puer. Truthmakers for tensed sentences Jonathan LOWE : Truthmaking as Essential Dependance Pascal ENGEL : Identity Makers

ontos verlag Frankfurt • Paris • Lancaster • New Brunswick 2007. 337 pages. Format 14,8 x 21 cm Hardcover EUR 98,00 ISBN 13: 978-3-938793-32-9 Due April 2007

P.O. Box 1541 • D-63133 Heusenstamm bei Frankfurt www.ontosverlag.com • [email protected] Tel. ++49-6104-66 57 33 • Fax ++49-6104-66 57 34

r o d o p i [email protected]–www.rodopi.nl

Spiritual and Political Dimensions of Nonviolence and Peace Edited by David Boersema and Katy Gray Brown

This book is a collection of philosophical papers that explores theoretical and practical aspects and implications of nonviolence as a means of establishing peace. The papers range from spiritual and political dimensions of nonviolence to issues of justice and values and proposals for action and change.

Amsterdam/New York, NY, 2006 VII-266 pp. (Value Inquiry Book Series 182) Paper € 55 / US$ 72 ISBN-10: 904202061X ISBN-13: 9789042020610

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Understanding Moral Weakness Daniel P. Thero

This book considers the common human predicament that we often choose an action other than the one we perceive to be best. Philosophers know this problem as akrasia. The author develops a nuanced understanding of the nature and causes of akrasia by integrating the best insights of Socrates, Aristotle, Augustine, and Aquinas, and several contemporary philosophers.

Amsterdam/New York, NY, 2006 IX-166 pp. (Value Inquiry Book Series 183) Paper € 36 / US$ 47 ISBN-10: 9042020784 ISBN-13: 9789042020788

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Bioethics and Vulnerability A Latin American View Florencia Luna

This book presents some of the challenges bioethics in Latin America faces today. It considers them through the lenses of vulnerable populations, those incapable of protecting their own interests, such as the illiterate, women in societies disrespectful of their reproductive rights, and research subjects in contexts where resources are scarce.

Amsterdam/New York, NY, 2006 XI-177 pp. (Value Inquiry Book Series 180) Paper € 40 / US$ 52 ISBN-10: 9042020733 ISBN-13: 9789042020733

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Faith in the Enlightenment? The Critique of the Enlightenment Revisited Edited by Lieven Boeve, Joeri Schrijvers, Wessel Stoker & Hendrik M. Vroom

One of the urgent tasks of modern philosophy is to find a path between the rationalism of the Enlightenment and the relativism of postmodernism. Rationalism alone cannot suffice to solve today’s problems, but neither can we dispense with reasonable critique. The task is to find ways to broaden the scope of rational thought without losing its critical power. The first part of this volume explores the ideas of Enlightenment philosophers and shows nuances often absent from the common view of the Enlightenment.

Amsterdam/New York, NY, 2006 380 pp. (Currents of Encounter 30) Bound € 76 / US$ 99 ISBN-10: 9042020679 ISBN-13: 9789042020672

The second part deals with some of the modern heirs of Enlightenment, such as Durkheim, Habermas, and Derrida. In the third part this volume looks at alternatives to Enlightenment thought in West European, Russian and Buddhist philosophy. Part four provides, over against the Enlightenment, a new starting point for the philosophy of religion in thinking about human beings, God, and the description of phenomena.

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Metaphoricity and the Politics of Mobility

Amsterdam/New York, NY, 2006 188 pp. (Thamyris 12) Paper € 40 / US$ 52 ISBN-10: 9042020342 ISBN-13: 9789042020344

Edited by Maria Margaroni and Effie Yiannopoulou

This collection of essays investigates the convergence between the postmodern politics of mobility and a politics of metaphor, a politics, in other words, in the context of which the production and displacement of meaning(s) constitute the major stakes. Ranging from discussions of re-territorialization, multiculturalism, “digisporas” and transnational politics and ethics, to September 11th, the Pentagon’s New Map, American legislation on Chinese immigration, Gianni Amelio’s film Lamerica, Keith Piper’s online installations and Doris Salcedo’s Atrabiliarios, the collection aims to follow three different theoretical trajectories. First, it seeks to rethink our concepts of mobility in order to open them up to the complexity that structures the thoughts and practices of a global order. Second, it critically examines the privileged position of concepts and metaphors of mobility within postmodern theory. In juxtaposing conflictual theoretical formulations, the book sets out to present the competing responses that fuel academic debates around this issue. Finally, it evaluates the influence of our increasingly mobile conceptual frameworks and everyday experience on the redefinition of politics that is currently under way, especially in the context of Post-Marxist theory. Its hope is to contribute to the production of alternative political positions and practices that will address the conflicting desires for attachment and movement marking postmodernity.

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Ectogenesis Artificial Womb Technology and the Future of Human Reproduction Edited by Scott Gelfand and John R. Shook

This book raises many moral, legal, social, and political, questions related to possible development, in the near future, of an artificial womb for human use. Is ectogenesis ever morally permissible? If so, under what circumstances? Will ectogenesis enhance or diminish women’s reproductive rights and/or their economic opportunities? These are some of the difficult and crucial questions this anthology addresses and attempts to answer.

Amsterdam/New York, NY, 2006 XII-197 pp. (Value Inquiry Book Series 184) Paper € 42 / US$ 55 ISBN-10: 9042020814 ISBN-13: 9789042020818

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl Amsterdam/New York, NY, 2006 XIII-133 pp. (On the Boundary of Two Worlds: Identity, Freedom, and Moral Imagination in the Baltics 7) Paper € 30 / US$ 38 ISBN-10: 904202092X ISBN-13: 9789042020924

Vasily Sesemann Experience, Formalism, and the Question of Being Thorsten Botz-Bornstein

Born in Vyborg in 1884 by parents of German descent, Vasily (Wilhelm) Sesemann grew up and studied in St. Petersburg. A close friend of Viktor Zhirmunsky and Lev P. Karsavin, Sesemann taught from the early 1920s until his death in 1963 at the universities of Kaunas and Vilnius in Lithuania (interrupted only by his internment in a Siberian labor camp from 1950 to 1956). Botz-Bornstein’s study takes up Sesemann’s idea of “experience” as a dynamic, constantly self-reflective, “ungraspable” phenomenon that cannot be objectified. Through various studies, the author shows how Sesemann develops an outstanding idea of experience by reflecting it against empathy, Erkenntnistheorie (theory of knowledge), Formalism, Neo-Kantianism, Freudian psychoanalysis, and Bergson’s philosophy. Sesemann’s thought establishes a link between Formalist thoughts about “dynamics” and a concept of Being reminiscent of Heidegger. The book contains also translations of two essays by Sesemann as well as of an essay by Karsavin.

USA / Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Contemporary Pragmatism 3:1 Edited by John R. Shook and Paulo Ghiraldelli, Jr.

Contemporary Pragmatism is an interdisciplinary, international journal for discussions of applying pragmatism, broadly understood, to today’s issues. This journal will consider articles about pragmatism written from the standpoint of any tradition and perspective, but it will concentrate on original explorations of pragmatism and pragmatism’s relations with humanism, naturalism, and analytic philosophy. The journal welcomes both pragmatism-inspired research and criticisms of pragmatism. We cannot consider submissions that principally interpret or critique historical figures of American philosophy, although applications of past thought to contemporary issues are sought. Contributions may deal with current issues in any field of philosophical inquiry. CP encourages interdisciplinary efforts, establishing bridges between pragmatic philosophy and, for example, theology, psychology, pedagogy, sociology, medicine, economics, political science, or international relations.

Amsterdam/New York, NY, 2006 182 pp. (Contemporary Pragmatism Vol. 3, Issue 1) Paper € 37 / US$ 46 ISBN-10: 9042021225 ISBN-13: 9789042021228

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

www.rodopi.nl

Fichtes Spätwerk im Vergleich Beiträge zum Fünften Internationalen Fichte-Kongre␤ »Johann Gottlieb Fichte. Das Spätwerk (1810-1814) und das Lebenswerk« in München vom 14. bis 21 Oktober 2003. Teil III Edited by Zöller, Günter und Hans Georg von Manz (Hrsg.) Amsterdam/New York, NY, 2006 XI-242 pp. (Fichte-Studien 30) Paper € 50 / US$ 63 ISBN-10: 9042021144 ISBN-13: 9789042021143 Vol. 1 to 5 ISBN-10: 9042020458 ISBN-13: 9789042020450

Jürgen STOLZENBERG: Fichtes Deduktionen des Ich 1804 und 1794 Ulrich SCHLÖSSER: Worum geht es in der späteren Wissenschaftslehre und inwiefern unterscheiden sich die verschiedenen Darstellungen von ihr dem Ansatz nach? Enrico GIORGIO: Der Begriff »absolutes Wissen« in der WL-1801/02 aus der Perspektive der Spätlehre Faustino FABBIANELLI: Ist die späte Wissenschaftslehre ein »Aktualer Idealismus«? Ein spekulativer Vergleich zwischen Fichtes und Gentiles Denken Vadim V. MURSKIY: Fichtes Spätwerk in Bezug auf das Problem der Einheit und der Veränderung seiner Lehre Johannes BRACHTENDORF: Substanz, Subjekt, Sein – die Spinoza-Rezeption der frühen und der späten Wissenschaftslehre Birgit SANDKAULEN: Spinoza zur Einführung. Fichtes Wissenschaftslehre von 1812 Ewa NOWAK-JUCHACZ: Philosophie als vox pacis. J. G. Fichtes Pragmatik als Gegenstück des regulativen Friedensideals I. Kants Vladimir ALEKSEEVIC ABASCHNIK: Die ersten Fichteaner über die Schwierigkeiten des Verständnisses der Wissenschaftslehre Marina PUSCHKAREWA: Fichte und Schelling. Das Problem des »Trägers des Wissens« Robert MARSZALEK: Fichtes Religionstheorie im Licht der Schellingschen Gedanken zur Mythologie Salvatore PATRIARCA: Gesetz und Selbstbestimmung des Absoluten. Ein Vergleich zwischen der späten Philosophie Fichtes und der mittleren Philosophie Schellings Paul ZICHE: Systemgrundriß und blitzartige Einsichten. Zum Verhältnis von Propädeutik und systematischer Philosophie bei Fichte und Schelling Giacomo RINALDI: Method and Speculation in Fichte’s Later Philosophy Angelica NUZZO: Fichte’s 1812 Transcendental Logic – Between Kant and Hegel Diogo FERRER: Hegels Fichte-Kritik und die späte Wissenschaftslehre Rolf AHLERS: Der späte Fichte und Hegel über das Absolute und Systematizität Matteo Vincenzo D’ALFONSO: Schopenhauer als Schüler Fichtes Xabier Insausti UGARRIZA: José Manzanas Rezeption des späten Fichte Ibon Uribarri ZENEKORTA: Manzana zwischen Kant und Fichte. Das Absolute als entscheidende Frage Hiroshi KIMURA: Fichte und Tekirei Edo – Bild und Feld

USA/Canada: 295 North Michigan Avenue - Suite 1B Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Perspektiven der Philosophie

Amsterdam/New York, NY, 2006 391 pp. Paper € 80 / US$ 104 ISBN-10: 9042021314 ISBN-13: 9789042021310

Band 32 – 2006. Begründet von Rudolph Berlinger Edited by Herausgegeben von Wiebke Schrader, Georges Goedert und Martina Scherbel

Inhalt I. Erschließung von Sinnräumen Jürgen GROßE: Langeweile. Zur Metaphysik einer Stimmung Sigbert GEBERT: Welt, Sinn, Gefühle und “das” Nichts. Blinde Flecken der Systemtheorie Heinz-Gerd SCHMITZ: Onto-Semiotik. Zur Grundlegung der Zeichentheorie bei Saussure und Heidegger Vítezslav HORÁK: Das Bild als Werkzeug II. Fluchtpunkte der Freiheit Dagmar FENNER : Ist die “negative Freiheit” ein Irrtum? Berlins Konzept “negativer Freiheit” im Kontrast zu Taylors Gegenentwurf “positiver Freiheit” Reinhard PLATZEK: Moderne Hirnforschung oder das vermeintliche Ende des freien Willens Kurt MAGER: Wissen als Verrat an der Freiheit der Existenz? Zum Problem der Subjektivität bei Karl Jaspers Georges GOEDERT: Dankbarkeit als Dialogizität Andreas LISCHEWSKI: Über die Erziehung zum Patriotismus. Geschichtlicher Streifzug zu einem aktuellen Thema III. Perspektiven des Sinngrundes Edgar FRÜCHTEL: Inneres Auge und göttliche Schau. Reflexionen zum antiken Horizont des Begriffs “Vision” Helke PANKNIN-SCHAPPERT: Ein Denker zwischen Mittelalter und Neuzeit. Zum Selbstverständnis des Nikolaus von Kues in seiner Spätschrift De apice theoriae Peter BÖHM: Vom Wesen des Menschseins. Überlegungen zur politischen Ästhetik bei Karl Philipp Moritz Harald HOLZ: All-Wesen und Unendlichkeit: Chinesische und europäische Landschaftsmalerei im Vergleich IV. Buchbesprechung Redaktionsnotiz

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Essays in Logic and Ontology Edited by Jacek Malinowski and Andrzej Pietruszczak

The aim of this book is to present essays centered upon the subjects of Formal Ontology and Logical Philosophy. The idea of investigating philosophical problems by means of logical methods was intensively promoted in Torun by the Department of Logic of Nicolaus Copernicus University during last decade. Another aim of this book is to present to the philosophical and logical audience the activities of the Torunian Department of Logic during this decade. The papers in this volume contain the results concerning Logic and Logical Philosophy, obtained within the confines of the projects initiated by the Department of Logic and other research projects in which the Torunian Department of Logic took part.

Amsterdam/New York, NY, 2007 400 pp. (Poznan´ Studies in the Philosophy of the Sciences and the Humanities 91) Bound € 80 / US$ 104 ISBN-10: 9042021306 ISBN-13: 9789042021303

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Education for a Democratic Society Central European Pragmatist Forum, Volume Three Edited by John Ryder and Gert Rüdiger Wegmarshaus

This book is the third volume of selected papers from the Central European Pragmatist Forum (CEPF). It deals with the general question of education, and the papers are organized into sections on Education and Democracy, Education and Values, Education and Social Reconstruction, and Education and the Self. The authors are among the leading specialists in American philosophy from universities across the U.S. and in Central and Eastern Europe. Studies in Pragmatism and Values (SPV) promotes the study of pragmatism’s traditions and figures, and the explorations of pragmatic inquiries in all areas of philosophical thought.

Amsterdam/New York, NY, 2007 XVII-194 pp. (Value Inquiry Book Series 179) Paper € 43 / US$ 58 ISBN-13: 9789042021532

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations

r o d o p i [email protected]–www.rodopi.nl

Science in Culture Piotr Jaroszyn´ ski

This book tries to uncover science’s discoverer and explain why the conception of science has been changing during the centuries, and why science can be beneficial and dangerous for humanity. Far from being hermetic, this research can be interesting for all who want to understand deeper what really conditions the place of science in culture.

Amsterdam/New York, NY, 2007 XXI-314 pp. (Value Inquiry Book Series 185) Paper € 68 / US$ 88 ISBN-10: 9042021365 ISBN-13: 9789042021365

USA/Canada: 295 North Michigan Avenue - Suite 1B, Kenilworth, NJ 07033, USA. Call Toll-free (US only): 1-800-225-3998 All other countries: Tijnmuiden 7, 1046 AK Amsterdam, The Netherlands Tel. +31-20-611 48 21 Fax +31-20-447 29 79 Please note that the exchange rate is subject to fluctuations