Mental Spaces in Discourse and Interaction
Pragmatics & Beyond New Series (P&BNS) Pragmatics & Beyond New Series is a continuation of Pragmatics & Beyond and its Companion Series. The New Series offers a selection of high quality work covering the full richness of Pragmatics as an interdisciplinary field, within language sciences.
Editor Andreas H. Jucker
University of Zurich, English Department Plattenstrasse 47, CH-8032 Zurich, Switzerland e-mail:
[email protected]
Associate Editors Jacob L. Mey
University of Southern Denmark
Herman Parret
Jef Verschueren
Susan C. Herring
Emanuel A. Schegloff
Belgian National Science Foundation, Universities of Louvain and Antwerp
Belgian National Science Foundation, University of Antwerp
Editorial Board Shoshana Blum-Kulka Hebrew University of Jerusalem
Jean Caron
Université de Poitiers
Indiana University
Masako K. Hiraga
St.Paul’s (Rikkyo) University
University of California at Los Angeles
Deborah Schiffrin
David Holdcroft
Georgetown University
Sachiko Ide
Kobe City University of Foreign Studies
Sandra A. Thompson
Thorstein Fretheim
Catherine KerbratOrecchioni
John C. Heritage
Claudia de Lemos
Teun A. van Dijk
Marina Sbisà
Richard J. Watts
Robyn Carston
University College London
Bruce Fraser
Boston University University of Trondheim University of California at Los Angeles
University of Leeds Japan Women’s University
University of Lyon 2 University of Campinas, Brazil University of Trieste
Volume 170 Mental Spaces in Discourse and Interaction Edited by Todd Oakley and Anders Hougaard
Paul Osamu Takahara
University of California at Santa Barbara Pompeu Fabra, Barcelona University of Berne
Mental Spaces in Discourse and Interaction Edited by
Todd Oakley Case Western Reserve University
Anders Hougaard University of Southern Denmark
John Benjamins Publishing Company Amsterdam / Philadelphia
8
TM
The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.
Library of Congress Cataloging-in-Publication Data Mental spaces in discourse and interaction / edited by Todd Oakley and Anders Hougaard. p. cm. (Pragmatics & Beyond New Series, issn 0922-842X ; v. 170) Includes bibliographical references and index. 1. Discourse analysis--Psychological aspects. 2. Social interaction. I. Oakley, Todd. II. Hougaard, Anders. P302.8.M47 2008 401'.41--dc22 isbn 978 90 272 5414 6 (Hb; alk. paper)
2007041026
© 2008 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
TSL[v.20020404] Prn:30/01/2008; 12:12
F: PB170CO.tex / p.1 (v)
Table of contents
introduction Mental spaces and discourse analysis Anders Hougaard and Todd Oakley
1
chapter 1 Connecting the dots: Mental spaces and metaphoric language in discourse Todd Oakley and Seana Coulson
27
chapter 2 The text and the story: Levels of blending in fictional narratives Barbara Dancygier
51
chapter 3 Fictive interaction blends in everyday life and courtroom settings Esther Pascual
79
chapter 4 A semiotic approach to fictive interaction as a representational strategy in communicative meaning construction Line Brandt chapter 5 Designing clinical experiences with words: Three layers of analysis in clinical case studies Todd Oakley and David Kaufer
109
149
chapter 6 Compression in interaction Anders Hougaard
179
chapter 7 Guided conceptualization: Mental spaces in instructional discourse Robert F. Williams
209
TSL[v.20020404] Prn:30/01/2008; 12:12
F: PB170CO.tex / p.2 (vi)
Mental Spaces in Discourse and Interaction
chapter 8 Looking at analyses of mental spaces and blending / Looking at and experiencing discourse in interaction Alan Cienki chapter 9 “Mental spaces” and “blending” in discourse and interaction: A response Gitte R. Hougaard
235
247
chapter 10 Reflections on blends and discourse Paul Chilton
251
Author index
257
Subject index
259
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.1 (1)
Mental spaces and discourse analysis Anders Hougaard and Todd Oakley
Since their inceptions, mental spaces theory (Fauconnier 1994, 1997) and conceptual integration theory, better known as blending theory (Fauconnier & Turner 1994, 1998, 2002), have developed rapidly from being semantic theories of (mainly textual) discourse meaning to general cognitive1 theories of human sense making across diverse domains of human activity. These activities range from text and discourse comprehension to the construction of meaning in religious rituals and meaning in music. This development has been endorsed by the elaboration of Fauconnier’s (1994 [1985]) early work on mental spaces in discourse into a larger theoretical machinery which has the general cognitive capacity of blending (and accompanying processes) as its central object of inquiry. Despite this expansive development, however, discourse, in its many variants (from poetry to face-toface interaction), remains a vibrant area of research within the mental spaces and conceptual integration framework (hereafter MSCI). Two major developments can be observed in the discursive application, elaboration and rethinking of MSCI. One is that increasingly more attention is being paid to the way contextual or situational factors determine mental spaces and blending operations, under which or in accordance with which discourse participants construct meaning. And here we use both terms – “context” and “situation” – in the broadest conceivable fashion as any micro, mezzo or macro, textual or non-textual set of circumstances. The other major development, already indicated in the latter, is an increasing multiplicity of particular theoretic and empirical notions on which researchers study discourse. A great number of major approaches to discourse from the fields of pragmatics, text linguistics, discourse studies and interactional studies are represented in discursive MSCI research. This diversity of approaches presents the joint cognitive enterprise with a great challenge: either 1) to attempt to reconcile these diverse theoretic and empirical bases into a joint enterprise that produces comparable results without violating the the. For a discussion of mental space as a cognitive theory, see below.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.2 (2)
Anders Hougaard and Todd Oakley
oretic and methodological bases of these results or 2) to consider an appropriate division of labor between them, which involves considering which theories and methods do what best and what results are incompatible with which theories and methods. It was in the interest of the two challenges that Hougaard and Oakley organized a panel session at the International Pragmatics Organization’s biannual conference at Riva del Garda, Italy, in 2005. The present volume presents both written and extended versions of papers given on that occasion and additional invited papers by other researchers within the “branch” of discursive MSCI. The volume has two major purposes: One is to stimulate discussions within MSCI as well as cognitive linguistics in general of a joint discursive enterprise or of complementary discursive enterprises, including a thorough exchange of philosophical, theoretical and methodological viewpoints. The other is to provide other researchers who work with discourse and/or interaction and who take an interest in cognitive issues with a palette of different ways in which blending theorists have made discourse and interaction an object of inquiry. Additionally, to initiate the discussion of a general discursive and interactional MSCI enterprise – including both its prospects and problems – we have invited three relevant experts to respond to the contributions made in the present volume: discourse analyst Paul Chilton, cognitive linguist Alan Cienki, and conversation analyst Gitte R. Hougaard. Below we will give a short overview of the contributions to this volume. But first, for readers who are not familiar with MSCI, we offer a short introduction to MSCI and discourse, and then we discuss MSCI in relation to traditions of discourse analysis, with particular focus on some seemingly antithetical major “schools”: Ethnomethodology, Conversation Analysis and Discursive Psychology. We have chosen to focus on these because they are dealt with explicitly in several contributors and because they present novel and very different perspectives from those commonly found in cognitive linguistics. MSCI is commonly accepted as a tool for analysis of individuals’ interpretations of sequences of spoken and written language. Such studies have been a part of the cognitive linguistics endeavor at least since the first version of Fauconnier’s Mental Spaces (1985). Yet, the sociological methods employed by these schools introduce a new concern: meaning construction as a shared or public phenomenon. Many cognitive linguists would perhaps consider this issue as being beyond the concerns of cognitive linguistics proper, as being a sociological not a cognitive or linguistics concern. But in truth cognitive linguistics in general and MSCI in particular are starting to address explicitly the issue of socially shared and constructed meaning. For instance, Croft (to appear) argues explicitly for a “social” extension of cognitive linguistics, and in their 2002 book Fauconnier and Turner provide blending analyses of phenomena that are distinctly social and can only be achieved in and through social processes. These include the concept of ‘money’, Japanese image clubs (where sexual fantasies
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.3 (3)
Introduction: Mental spaces and discourse analysis
are enacted collaboratively by prostitutes and customers) and religious rituals. Furthermore, Turner (2001) dedicates a whole book to arguing for a fusion between cognitive science and social science using blending theory as a platform for bringing the two together. If MSCI makes social aspects of meaning construction a principal concern, then a sociological viewpoint may as a consequence change some of the theoretical premises. As several of this volume’s contributors and commentators make clear, these sociological viewpoints need to be incorporated in future iterations of MSCI theory.
Introducing mental spaces and blending: Achilles sees a tortoise In this section we discuss the technical application of mental spaces and in the following section we address some definitional issues with respect to the status of mental spaces. Consider then first the following piece of text, originally analyzed in Fauconnier (1997: 44): Achilles sees a tortoise. He chases it. He thinks that the tortoise is slow and that he will catch it. But it is fast. If the tortoise had been slow, Achilles would have caught it. Maybe the tortoise is really a hare.
This little slice of discourse consisting mostly of simple sentences requires of its reader an almost acrobatic mastering of information chunks which emerge and get connected in all sorts of ways during the very brief time span it takes to read and understand the text. Fauconnier analyses the reader’s cognitive take on the unfolding text in terms of a proliferating network of interconnected mental spaces. Let us follow Fauconnier’s analysis step-by-step.
Base space According to Fauconnier, all discourse networks of mental spaces grow from a so-called Base space which represents the discourse starting point of a meaning construction. It is important to notice that this space need not be true or real or actual in any way outside the cognizer’s (or cognizers’) understanding. Only as an interpreter’s understanding of something is it claimed to be very real. It thus simply marks the point from which the meaning construction extends. Truth is not an issue – people’s understanding is. People understand the sentence Achilles sees a tortoise whether or not Achilles is real, has been real, could be real or only exists in imagination. MSCI seeks to model how people keep track of currents of information and make sense of them however bizarre, fantastic, imaginary, fictive or real they may be. The Base space is thus a here-and-now space with respect to the unfolding discourse, not with respect to any real or possible world situation. This is also a major reason why mental spaces are cognitive. We need judgement
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.4 (4)
Anders Hougaard and Todd Oakley
a
a name Achilles b tortoise SEE a b
b
Figure 1.
to distinguish between truths, falsities, fictions, imaginations, etc.; but technical cognitive principles of meaning assembly do not make that distinction. Hence one might argue that MSCI follows a trend already set by speech act theory (Austin 1962) in relocating the issue of meaning from being an issue of truth or falsity to being an issue of understanding.
Mental spaces analysis The first sentence in the discourse is then, (1) Achilles sees a tortoise.
In mental space terms, the name Achilles is linked to an element, a, already introduced into the Base space prior to the present excerpt (this must be so due to the definite reference). The element b(tortoise), which is set up by the indefinite noun phrase a tortoise, gets introduced as a new element in the Base space; and finally sees evokes a SEEING frame (a structured domain of knowledge pertaining to seeing, Fillmore 1982, 1985) which takes a and b in the roles of seer and seen respectively (figure 1). Notice here that, mental spaces are only partial conceptual structures. They do not contain entire domains of knowledge, concepts or categories. They “recruit” exactly what is required by the discourse. Hence not all knowledge that is related to seeing is brought to bear on Achilles and the tortoise, only that Achilles is a seer who has caught sight of the tortoise which is a visible object. Yet, frames are considered to be activated and hence available in the background for further mental space construction. Background information telling us that Achilles is a man and that the tortoise is an animal makes it possible in the next sentence, (2) He chases it,
to reference them anaphorically. Otherwise the next elements only add the additional relation CHASE through the verb chases. A frame for chasing is activated and Achilles is cast in the role as chaser and the tortoise in the role as chased (figure 2).
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.5 (5)
Introduction: Mental spaces and discourse analysis
a
a name Achilles b tortoise SEE a b CHASE a b
b
Figure 2.
Notice thus how one space comes to contain a short narrative and thus a temporal and dynamic dimension: Achilles sees a tortoise and then he chases it. Mental spaces are not just frozen images, states or relations; they can have extension in time and change over time. The third sentence, (3) He thinks it is slow and that he will catch it,
sets up new spaces with new structure (figure 3). He thinks is called a ‘space builder’ because it marks a new (or “discontinuous”) chunk, or seam, of information with respect to the previous seams. We are no longer talking about what Achilles and the tortoise do in their shared world but about what Achilles thinks about that shared world and this is not the same – hence the partitioning into new spaces.2 Thus in the diagramming, Fauconnier represents Achilles’ thinking as two new spaces. The first part of sentence (3) sets up a belief space, M, with an internal structure where the tortoise, b’, is slow. Notice that a’, Achilles, appears in this space too even though this element has not been explicitly referred to. This is one way in which mental space theory is different form formal semantics. Language implies more than it explicitly states. It is claimed to be understood that Achilles himself is of course a part of that scenario in which the tortoise is slow since it must be slow for him in his belief. The second . The issue of when a new mental space is needed is a contested issue. Harder (2003), proposes that a new space is needed when chunks of information contrast. However, consider an example like: The sun is shining and that is actually how Achilles sees it too. Here there is a full match between what is the state of affairs in the world and what Achilles thinks is the state of affairs in the world. But we must of course acknowledge the distinction that is made between what is reported to be the objective state of affairs in the world and what is reported to be Achilles’ subjective impression of the world. Hence instead of having “contrast” as a criterion for mental space assignment, one might argue that “discontinuity” is a better candidate. Hence partitioned chunks of information may indeed be compatible although they are discontinues in time, space, ontology, modality, etc. and thus should be treated as separate chunks.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.6 (6)
Anders Hougaard and Todd Oakley
a name Achilles b tortoise SEE a b CHASE a b
.a
.b
Base space B SLOW b’
.a’
.b’
Belief space M
.a’’
Future space
CATCH a’’ b’’
.b’’
Figure 3.
part of sentence (3) sets up a space which is in the future of space B and M. In that space, Achilles, a”, catches the tortoise, b”. One might perhaps argue that what Achilles thinks should rightfully be placed with Achilles in the space already set up (the Base space), because just as Achilles SEES and CHASES, he THINKS in that space. Yet, this would risk confusions and make the modelling unnecessarily complicated. The hierarchical relationship between the spaces is captured in the progression of the network from the Base space and downwards. And placing spaces within spaces might risk confusing a representation of the situation as it is presented with a progressive representation of the chunks of information as they appear, the latter and not the former being the object of Fauconnier’s mental space theory. The fourth sentence, (4) But it is fast,
takes us back to the Base space and sets up the attribute FAST for b, the tortoise. Notice how the word But explicitly marks a contrast between FAST in space B and SLOW in space M. We are now manipulating – simultaneously – incompatible scenarios where the same tortoise is respectively FAST and SLOW (figure 4). The fifth sentence (5) If the tortoise had been slow, Achilles would have caught it,
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.7 (7)
Introduction: Mental spaces and discourse analysis
a name Achilles b tortoise SEE a b CHASE a b FAST b
.a
.b
Base space B SLOW b’
.a’
.b’
Belief space M
.a’’
Future space
CATCH a’ b’
.b’’
Figure 4.
sets up an interesting structure which must clearly belong in a space that is different from the Base space (figure 5). The conjunction If sets up a mental space H with hypothetical content (that is, it has a hypothetical relation to the bases space). And as Fauconnier (Ibid.: 46) observes, the distal past perfect form indicates that the space is counterfactually related to the discourse’s factual state of affairs in space B. In this hypothetical space the novel structures SLOWb1 and CATCHa1b1 appear. For the sake of simplicity and brevity, we skip the more elaborate discussion and detailed analysis of conditionals and matching conditions and instead direct the reader to Chapters 4 and 5 of Fauconnier’s Mappings in Thought and Language (1997). For now we simply observe that the mood of the discourse has now shifted from indicative to conditional. However, below we will elaborate the analysis of sentence 5 in another way, demonstrating how it can be conceived of as a conceptual blend. The last sentence, (6) Maybe the tortoise is really a hare,
sets up an epistemic situation in which the tortoise is a hare (figure 6). The resulting configuration between spaces B and P allows the so-called Access Principle to work: through the naming of an element in one space one may get access to that element’s counterpart in another space. So when talking about the tortoise in this latter case, we are not talking about the Base space tortoise; rather we use the name
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.8 (8)
Anders Hougaard and Todd Oakley Base space B
a name Achilles b tortoise SEE a b CHASE a b FAST b
.a
.b
SLOW b’
.a’
.b’
Belief space M
.a’’
CATCH a’ b’
.b1
Future space
.a1
.b’’
SLOW b1 CATCH a1 b1
Figure 5.
given in the Base space to gain mental access to the new entity, of which we now predicate that it is a hare.
Discourse management: Base, viewpoint, and focus The latter point about mental access shows us that grammar does not just indicate the introduction of new mental spaces and their content; it also guides a variety of other mental tasks in the treatment of the incoming information such as the management of viewpoint, focus and base structure in the constructed network. We have already seen how phrases like He thinks, If and Maybe can partition information away from the Base space, that is the state of affairs in the space that the discourse departs from. Consider now the notion of viewpoint. “At any point in the construction [of mental spaces],” Fauconnier (1997: 49) writes, “one space is distinguished as Viewpoint, the space from which others are accessed and structured or set up.” Consider again sentence (5): (5) If the tortoise had been slow, Achilles would have caught it.
In the Base space the tortoise is fast, but from the point of view of that space we look to a hypothetical space in which the tortoise is slow. Consider next the notion of Focus. As we saw, sentence (5) prompts a hypothetical mental space
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.9 (9)
Introduction: Mental spaces and discourse analysis
Base space B .a
a name Achilles b tortoise SEE a b CHASE a b FAST b
.b2
.b Possibility space P
SLOW b’
.a’
.b’
Belief space M
.a’’
CATCH a’ b’
.b1
Future space
.a1
Counterfactual
space H
.b’’
SLOW b1 CATCH a1 b1
Figure 6.
which is counterfactually related to the Base space. However, attention is on the hypothetical space, not the Base space. Though catching the discursive development of meaning management, the five-space network depicted above is far from the whole story of what is required of the interpreter of sentence (1) through (6). For instance, sentence (5) – If the tortoise had been slow, Achilles would have caught it– not only requires the partitioning of mental space structure to a counterfactual space, it also requires the integration of incompatible mental spaces in order to create a mental space of the counterfactual situation. Even though, the discourse base tortoise is not slow, we can still imagine it being slow, and we can further think of the consequences of it being – contrary to discourse facts – slow. Such discoveries led Fauconnier and Turner to propose the theory of conceptual blending. Above, sentence (5) was analysed in terms of two mental spaces. For the sake of simplicity, we will here treat the counterfactual situation as one blended space. In one input space – which would be the Base space from above, we have the fast tortoise and Achilles. In this space, Achilles cannot catch the tortoise. Yet, under normal circumstance we might expect a tortoise to be slow, so slow at least that a human can easily catch
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.10 (10)
Anders Hougaard and Todd Oakley
Human
Generic space
Tortoise Ac hilles a Input space 1
h Human
Tortoise b FAST b
t Tortoise t SLOW h t CA TC H
Input space 2
a’ Ac hilles b’ Tortoise b’ SL OW a’b’ CA TC H Blended space
Figure 7.
it.3 We have knowledge of such situations which can be used to imagine a situation in which this particular tortoise is slow – though it is not – and consequently that Achilles can catch it – though he cannot. The knowledge we have of normal tortoises constitutes another input space. And the imagined situation in which this particular tortoise is slow and gets caught is technically described as the blended space. Finally, the mapping of the two incompatible inputs gives rise to a generic structure supporting the entire blending work (figure 7). There are mappings between all the spaces in the blending space network. The mappings between the inputs are called vital relations (historically the notion of connectors from mental space theory prior to blending). Achilles and the tortoise in input 1 are connected to the human and the tortoise in input 2 by ROLE relations. There is also DISANALOGY between FAST in input 1 and SLOW in input 2. DISANALOGY is a central relation in the construction of counterfactual spaces. Achilles a and tortoise b map onto Achilles a’ and tortoise b’ in the blended space. This mapping has been established by the projection of Achilles (a’) and tortoise (b’) from input 1. The relations SLOW and CATCH have been projected from input 2 which establishes mappings between these relations in the input and the blend. Finally, Fauconnier and Turner posit a generic space that projects common structure to the other spaces in the network. . Whether Achilles is really human or not is of course debatable if seen from the point of view of mythology (given that only his heels could be wounded), but we abstain from such considerations here and instead observe that whether human or not, one would expect Achilles to be able to catch a tortoise!
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.11 (11)
Introduction: Mental spaces and discourse analysis
The basic elements of a technical description of blending are thus a minimum of four spaces (two inputs, generic space and blended space), mappings and selective projections (only selected parts of the inputs get projected). (Further basic elements are discussed in Fauconnier and Turner (2002)). As a key aspect of the blending, the blended space has emergent structure: The tortoise which is really fast is slow and Achilles who cannot catch the tortoise can now catch it. All blends have emergent structure, not only counterfactuals but also metaphors (He had smoke coming out of his ears), X is the Y of Z constructions (John is the President of the Company), certain caused motion constructions (They clapped the band back on stage), fictive motion (This highway goes to Copenhagen), NN compositions (money problems), AN compositions (false alarm) and many others. Another key component of blending is “compression.” This component may better be understood in the following example (7) At this World Cup, Ronaldo is fighting against his own former performances.
One understanding of this complex utterance has Ronaldo fighting a personification of his earlier performances. Yet, Ronaldo at the present World Cup is separated in time from his earlier performances. But the blending allows the interpreter to put present-day Ronaldo directly – at the same point in time and at the same place – up against his earlier performances. This blend then involves the compression of time that separates Ronaldo today from his earlier performances. This makes blending a phenomenal cognitive tool for bringing together in thought what cannot be brought together in the real world. Fauconnier and Turner (2002) emphasize the great importance and omnipresence of this capacity throughout human activities. Blending, though, has developed into much more than a theory of semantics and pragmatics. Fauconnier (1997) introduces blending as an elaboration of mental spaces theory in connection to discourse and sentence semantics. In Turner (1996), however, blending is further elaborated into a more general theory of imagination. And in Fauconnier and Turner (2002), which sums up nearly a decade of explosively spreading and growing blending research around the globe, blending has become a general theory of the singular nature of human thinking. Such domains as religion, semiotics, art, music, social science, politics, rhetoric, interaction, mathematics, anthropology and much more have been included in applications and elaborations of blending theory. For an overview of work done in blending and mental space theory we direct the reader to Mark Turner’s blending website:http://www.markturner.org/blending.html. Besides containing references and links to papers and events the webpage also links up to further reference sites.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.12 (12)
Anders Hougaard and Todd Oakley
What is a mental space? Above we have discussed technical aspects of the application of mental spaces, including when a new space is needed (at least theoretically speaking). However, thus far we can hardly postulate that we have much more than a useful model. The claim by Fauconnier and Turner that mental spaces are human cognition at work still fails to be supported by definitions and empirical investigations which concern themselves with what mental spaces really are (that is, what they are meant to represent), what their statuses really are. This issue has left the field completely open. Most MSCI researchers do not address the issue explicitly in their work, while others have turned it into a central question (for instance Brandt 2005 and Hougaard 2004). We will not attempt to provide a solution to this issue for a very important reason: the answer to the question of what mental spaces really are seems to go hand-in-hand with the directions that current research, including the papers in the present volume, takes. Often, mental spaces seem to be applied as a sort of “inner” representations of or ideas about the world or possible or imaginary worlds. In such works, mental spaces are not a far cry from earlier representationalism and the disembodied, dualist thinking (Cf. Descartes and Spinoza) that characterises much of cognitive science. However, none of the contributions to the present volume explicitly take on a representationalist view. MSCI’s representationalist relatives, when applied in this vein, include such theories as Johnson Laird’s (1983) Mental Models, which also focus on the way that representations of the world form a basis for reasoning and expectations and Situation Semantics (Barwise & Perry 1981), which also emphasizes the partial nature of represented situations. Others emphasize the embodied nature of the knowledge on which mental space building relies and thus distance themselves from representationalism. But too rarely do MSCI researchers explicitly address the issues of the relation of mental spaces to other cognitive representation theories, asking themselves, in what sense they are cognitive and what are the possible limitations for their application? This volume addresses such questions in connection to discourse.
Mental spaces and discourse studies The role cognition and cognitive science can or should play in discourse studies and vice versa is hotly contested, and one’s stand on the issue depends almost entirely on how one conceives of cognition and discourse in the first place. Is cognition something that resides entirely inside individual persons? Is it synonymous with knowledge or epistemic states, and is it antonymous with social action and emotional or passionate states? And if so, are these epistemic states manifestations of innate dispositions of which social action is merely an epiphenomenon? Alter-
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.13 (13)
Introduction: Mental spaces and discourse analysis
natively, is cognition distributed between or emergent among co-participants? Is cognition so dependent on embodied, social, and emotional experience that the very ideas of knowledge, reasoning, understanding, remembering and experiencing have to be understood in terms of how they arise in and from richly perceptual, affective, motor and interpersonal situations? Researchers in MSCI are far more likely to deny the first set of questions and affirm the second, thereby in principle aligning themselves with some of the tenets of major contemporary perspectives in discourse studies. However, MSCI is not itself a theory of discourse; rather, it is a theory of human cognition and conceptualization that is supposed to suggest ways to theorize about and model the “mental work” of discourse in its broadest sense. As pointed out in the commentary by Gitte R. Hougaard, the volume presents two major aims. The first aim is to show how discourse functions as a vehicle of particular cognitive processes of conceptual integration. The second complementary aim is to show how particular cognitive processes of conceptual integration arise from interaction between interlocutors in discourse. Thinking about discourse in terms of cognition is the primary focus of the contributions by Todd Oakley and Seana Coulson, Barbara Dancygier, Todd Oakley and David Kaufer, Line Brandt, and Ester Pascual. Thinking about cognition in terms of interaction is the primary focus of the contributions by Anders Hougaard and Robert Williams in their attempts to reveal what MSCI researchers can learn from microanalytic studies of interaction. Given the “chicken-and-egg” nature of these two aims, it is instructive to take stock of some common divergences and convergences with mainstream discourse studies and see what prospects they leave us with. We will focus in particular on ethnomethodology, conversation analysis and their daughter discipline discursive psychology. Discourse analysis often emphasizes the social and contextual aspects of language use as well as the need to look at actual, naturally occurring language use. As will be addressed in greater detail shortly, arguments about the status of evidence strike a chord among recent researchers in cognitive linguistics, especially those who defend a “usage based approach” to language structure and acquisition (Langacker 1988; Fauconnier 1999; Tomasello 2000). Such approaches take as their starting point descriptions of real situations of language use as and thus avoid isolating “language” as rules for generating sentences. In principle, the areas of cognitive linguistics which go beyond the boundaries of the sentence can be seen as inherently carrying a discourse analysis agenda. Yet, despite this general accord, significant differences and major challenges appear when trying to integrate cognitive linguistics with these other frameworks. The alternative frameworks allow for the complementary mode of investigation in which cognitive processes are to be understood from a social science perspective.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.14 (14)
Anders Hougaard and Todd Oakley
Taking into account the philosophical bases and empirical results of Conversational Analysis (hereafter CA) the question, “what is a mental space?” takes on new urgency. What do researchers assume when positing a mental space? Do they assume that the thing posited represents an individual’s “backstage cognition,” as Fauconnier likes to call it? If so, MSCI as an account of coparticipants’ common (shared, public) sense positions itself in unbridgeable contrast to CA studies, a framework focusing on the ways in which coparticipants make sense in and through ordered practices in a social world. For CA researchers, the world is not seen as made up of individual minds; rather coparticipants are seen as socially embedded from the outset and individual cognition is at best a secondary phenomenon of little interest to the practical sense that coparticipants create. For the notion of mental spaces to mean anything from an CA inspired perspective, it must, first and foremost, be thought of as a modelling of the observable jointly accomplished track of common sense not resident solely in each individual’s mind. CA researchers are, in Robert Hopper’s (2005) words, “cognitive agnostics,” if by “cognitive” one assumes that the most interesting things about discourse are happening within the minds of each individual participant. This is not a denial of individual experience, just a rejection of the idea that human sense making can be studied by taking the individual thinker as explanandum. In this respect, a CA view of mental spaces as co-created and shared bears resemblance to Clark’s (1996) notion of “common ground.” Importantly, still, in principle, neither a cognitive science nor a cognitive sociology is overruled by an EM-inspired approach. In fact, EM with its phenomenological roots (Schutz 1962 and 1964) remains complementary to research into human cognition and a potential source of inspiration or point of departure for such studies. Thus Cicourel (1974) elaborated on Schutzian notions of the everyday assumptions (“attitudes”) that people rely on in making sense of their social world to what may in fact be considered a cognitive science account of social knowledge. MSCI researchers will notice a very close resemblance between Cicourel (1974) and contemporary work by cognitive linguist Charles Fillmore that lead to his influential “frame semantics” (1982, 1985). Others have approached cognitive issues in ways similar to Cicourel’s (including Hougaard 2004). However, these approaches will always be controversial, because they undeniably impose “cognitive” interpretations on EM or CA tenets and because they involve considerations of how far the phenomenological roots of EM and CA can be taken. More recently, several of the most prominent proponents of CA (including Charles Goodwin, Douglas Maynard, and Derek Edwards) probe human cognition through the study of everyday social interaction (see Potter and Molder 2005). They believe this approach can both generate cognitive insights and evoke new ways of considering the issue of cognition altogether. This work mostly avoids any kind of so-called “mentalism” but without, we believe, fully realizing the potential
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.15 (15)
Introduction: Mental spaces and discourse analysis
of CA to cognitive science and with the potential of being agnostic to the empirical that important aspects of social action cannot be captured by action analysis alone and may necessitate a “cognitive” description. For instance, Hougaard and Hougaard (forthcoming) may be seen to be advance an implicit programmatic agenda: only by taking very seriously the attempt to do exhaustive micro-analysis of face-to-face-interaction, aspects of social life which may defy exhaustive description in terms of social “actions” will be revealed. In the long run, CA may turn out to be an exceptionally useful resource for redefining and specifying cognitive science agendas. Influenced by Wittgenstein’s later philosophy, basic tenets of EM and CA, Discursive psychology (hereafter DP) reformulates general topics in cognitive psychology, such as memory, categories, beliefs, attitudes, and emotion, in terms of discourse and interaction (see Edwards and Potter 2005). That is, discursive psychology seeks to employ the analytic frameworks and techniques of discourse analysis and CA as a way of redefining the issue of human cognition. It is tempting to think of MSCI as being synonymous with “the psychology of discourse,” if we take this partitive construction to mean the application of cognitive theories to the study of discourse. However, contributions by Oakley and Coulson, Oakley and Kaufer, Brandt, and Pascual bear witness to the false promise of any such unidirectional application, as each contribution explicitly struggles with how to adapt the mental spaces framework to data under investigation. Most conspicuously, each contribution attests to the need to treat “shared knowledge” between discourse participants explicitly in ways Fauconnier and Turner do not. In these contributions, MSCI to some extent converges with DP (and other mainstream theories of discourse) in making shared knowledge or “common ground” (Clark 1996) an important part of their analytic routines. How one is to treat shared knowledge – as a set of internal representations active in each mind during acts of speaking and hearing, reading and writing, as a set of procedures enacted by the participants at the moment of interaction and hence not internal to the individual participants, or some combination thereof – is perhaps the principal theoretical and empirical problem which has yet to achieve any kind of consensus within the MSCI community. A point at which MSCI and DP show complementary interests is with regards to scripted behavior. In the cognitive science tradition scripts (cf. Schank & Abelson 1977) are abstractions from experiential reality of particular domains of activity. Scripts then are heuristics for guiding people’s thoughts and actions through situations, since each situation can be understood as a manifestation of a kind of situation. DP and MSCI both treat scripts as playing an important role in people’s everyday lives. However, for DP (cf. Edwards 2005) scripts do not reflect an individual’s a priori perceptual and cognitive dispositions toward events in the world. Instead, DP’s interest in scripted behavior concerns how interlocu-
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.16 (16)
Anders Hougaard and Todd Oakley
tors formulate scripts in social conduct and what they achieve by them. This is not just a social application of a cognitive science concept, it is a redefinition of cognitive psychology method. Instead of assuming that mental scripts guide and predetermine people’s behavior, the object of study is how people work up and formulate events as routines. Scripted behavior, in other words, is studied as a member’s issue (that is, in terms of what the interlocutors themselves understand to be scripted) instead of as an analysts’ theoretical issue (that is, in terms of how the analyst makes sense of people’s routine behaviors). But this does not necessarily alienate MSCI. The role moment-by-moment interpersonal interaction and its study should play in the formation of scripts has yet to be comprehensively addressed within MSCI. However, a review of recent work (cf. the collection of papers in Coulson and Oakley 2005) within MSCI (including the papers in the present volume) shows that the ground is being paved for this work.
Embodied cognition is not cognitivism and CA is not behaviorism We would like to propose that a careful dialogue between socially grounded discourse analysis and MSCI can generate much more fruitful common ground than is presently the case. Importantly, such an engagement requires clarification of the ideologies behind both frameworks, namely that MSCI advances a cognitivist agenda and that EM and CA advance a behaviorist agenda. Before we conclude this introduction we would like to respond to these two claims. 1) Present day cognitive science only thinks about cognition in terms of internal disembodied representations: When giving (mostly negative) accounts of the cognitive science enterprise (cf. Coulter 1979), DP, EM and CA researchers in fact typically refer to a tradition of work which MSCI-researchers also distance themselves from, namely what the latter call “First Generation Cognitive Science” (Lakoff & Johnson 1999). This cognitvistic enterprise – spearheaded by people like Noam Chomsky and Jerry Fodor often explicitly adopts 1) a Cartesian, dualist conception of mind and body considering the latter as merely matter that does not transcend the mind, 2) a representationalist view of knowledge of the world as inner interpretations which are completely separated from the outer world, 3) minimalist accounts of for instance meaning as residing “in the words” disregarding contextual and pragmatic circumstances, 4) a modular notion of linguistic cognition as being a separate sort of cognition, and 5) a reductionist notion of language as simply a part of our biological endowment. In contrast to this, MSCI researchers and their fellow “Second Generation Cognitive Scientists” consider mind as inseparable from the body and its world and meaning and knowledge of the world as emerging directly from the body’s interaction with its world. Thus representa-
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.17 (17)
Introduction: Mental spaces and discourse analysis
tionalism is mostly replaced by experientialism (Lakoff & Johnson 1980) in which the mind does not manipulate disembodied ideas of the world but reflects on or turns upon knowledge emerging directly from its experience. In Second Generation Cognitive Science the mind is not just for but also of its social and material world (for accounts of embodiment as a physiological and social condition see for instance Sinha and De Lopez Jensen 2000 and Zlatev et al. 2007). This necessarily entails a maximalist notion of meaning, since the entire embodied experience of and being in the world becomes potential substance of thought just as the world beyond the skin of the individual becomes a locus of thinking (cf. Hutchins 1995, 2005). Language itself is also seen as in great part shaped by the body’s experience of the world and thus as a part of general cognitive capacities. The fact that Second Generation Cognitive Science sees the mind as shaped by the body’s interaction with the world as well as extending beyond the individual to include the surrounding world provides a potential bridge to cognitively interested, socially committed discourse analysis. Therefore, CA focus on meaning as a feature of the shared “outward” world seems all the more relevant to present day cognitive science. 2) EM and CA is a behaviorist discipline: some EM and CA researchers do endorse behaviorism, but CA is not per se a behaviorist enterprise. On the contrary, CA assumes as one of its central tenets that coparticipants rely on common sense assumptions, expectancies, and shared categories in order to make sense of the world. However, CA avoids making assertions about the hypothetical mental “structures” or “processing” underlying them. Instead a central effort of EM and CA is to find out what those assumptions and categories are from careful study of participants’ own orientations. As Cicourel (1974) presaged, this almost offers itself for cognitive analysis. Yet, the challenge is to use EM and CA analyses and findings without resorting to cognitive speculation which in the end offers little or nothing to the understanding of the social world that the coparticipants have in common. In summary, there are many ways in which current trends in MSCI and second generation cognitive science, EM, CA and DP research may cross and develop in mutual exchange. We find that the present rethinking of concepts and methodologies within the cognitive sciences and the growing interest in cognition as an object of discussion and inquiry within discourse analysis and CA provides a historical opportunity for fruitful interdisciplinary exchanges, and we hope that this volume will serve as one among many sources of inspiration or motivation for such a development.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.18 (18)
Anders Hougaard and Todd Oakley
Overview of the volume MSCI is both a theory of sense making in its own right and a central theory within the broader and rapidly developing and growing enterprise of cognitive linguistics. The development of MSCI is at the front line of the development of cognitive linguistics. It both reflects the development of major interests and contributes to this development. Actual studies in MSCI thus play a role in the development of cognitive discourse studies, metaphor theory, metonymy, construction grammar, the social grounding of cognition, the important development of cognitive linguistics outside the realm of linguistics, online meaning construction, rhetoric, the phenomenology of cognitive linguistics and much more. With the collection of papers in the present volume we hope to illustrate how MSCI pushes cognitive linguistics along specifically with respect to discourse and interaction. The papers represent a diversity of approaches within these burgeoning fields and thus reflect that blending theory has witnessed an explosion of studies which apply or discuss blending and mental spaces in connection to discursive and interactional contexts. These studies have been quite diverse in terms of methodology, discursive and interactional theory and specific focus and hence we see a need to bring these diverse studies together in the hope of initiating a fruitful debate about prospective common ideas, results and goals as well as to identity specific but commonly important contributions from each type of study. The seven chapters of this volume comprise three sections, each of which is focusing on a specific dominion of inquiry related to the larger domain of discourse studies. The three chapters of the first section focus attention on discourse in its spoken and written manifestations that conform to well-defined and distinct speech genres, from a radio interview program, to narrative fiction, to articles published in a medical journal. The two chapters of the second section focus attention on spoken interaction as a mode of conceptualization and representation within discourse in highly constrained and formalized courtroom settings and, alternatively, in the more spontaneous and less formalized but nevertheless structured context of a classroom discussion. The two chapters of the third section focus on interaction in semi-formal situations of a call-in talk radio program and an instance of classroom instruction from the micro-social perspective of conversational analysis. Before describing each chapter, we would like to say a few words about terminology and diagrammatic conventions. One thing that will become clear to the readers is that issues about what to call something in MSCI are not completely settled and that some researchers use different names for similar entities. Nowhere is this more apparent than in the chapters by Oakley and Coulson, Oakley and Kaufer, Pascual, and Brandt. The notion of Base space, outlined above, achieves a different status and function in the modes of analysis and graphic representa-
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.19 (19)
Introduction: Mental spaces and discourse analysis
tions in the Chapters 1, 3, and 5, particularly as it pertains to Base space, variously referred to as Semiotic space in Chapter 5 and Grounding in Chapters 1 and 3, and the HERE AND NOW space in Chapter 4. As with Fauconnier’s treatment of Base space (1997), these entities specify the point from which discourse extends. Unlike Fauconnier’s treatment, however, these entities function as conventionalized placeholders for local discourse context to include discourse participants (i.e., the speaker and hearer) rather than as characters (i.e., Achilles) and the relevant setting and situations in which the discourse unfolds. In this respect, the mode of analysis seeks to specify certain real world discourse conditions researchers consider indispensable for online meaning construction to occur. Other divergences in nomenclature will be apparent to readers, particularly in the names given to input spaces (i.e., presentation space, reference space, narration space, etc.). The contributors also exhibit different aesthetic preferences for graphic conventions: diagrams appearing in Chapter 1, 2, 3, 5, and 7 use solid arrows to signify selective projection of conceptual structure among the spaces in the network, while diagrams in Chapters 4 and 6 use dotted and solid lines, respectively. Diagrams in Chapter 4, 5, and 7 highlight the cross-space mappings between input spaces with solid lines, while diagrams in Chapters 1, 2, and 3 do not. The divergent nomenclature and diagram aesthetics, however, should not be taken as an indication that MSCI researchers are working in different theoretical frameworks or that the original analytic model developed by Fauconnier and Turner is radically different. The basic tenets of mental spaces theory – that meaning construction proceeds by a proliferating unfolding of situations and scenes that can be activated, elaborated, integrated and blended to satisfy any number of expressive purposes – remain intact. What these differences do suggest is that the precise way of graphically representing these meaning construction processes will vary significantly across researchers working in this framework, not to mention the fact that many feel no need whatsoever to create diagrams at all. Rightly or wrongly, the larger MSCI community seems to interpret these variations as indicators of differences in temperament and aesthetic sensitivity than as indicators of contrasting theoretical and methodological approaches. In the opening chapter, Todd Oakley and Seana Coulson examine the rise of the expression “connect the dots” in the 21st century and its relationship to other conventional metaphors in English that refer to the interpretation of information. They examine three particular cases of this metaphor in the speech of counter-terrorism expert Richard Clarke in a radio interview. Since its inception, the authors argue, MSCI was designed to capture spontaneous online processes that can yield short-lived and novel conceptualizations (Grady, Oakley, & Coulson 1999), thus necessitating its extension to the study of intonation contours in spoken discourse. Doing so will enable researchers to assess whether intonational cues for indicating the status of information in the on-going discourse can
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.20 (20)
Anders Hougaard and Todd Oakley
be used by listeners to determine whether structure from an existing mental space or whether a new space must be constructed. Such an examination will enable the assessment of the compatibility between mental spaces theory and Chafe’s theory of discourse (1994). Moving from the realm of spoken conversation to narrative fiction, Barbara Dancygier’s chapter shows how MSCI can account for processes enabling readers of a fictional narrative to construct a complete sequential story out of the incomplete set of narrated events. Using the novels of Margaret Atwood, F. Scott Fitzgerald, Ian McEwen, and Jon Potocki as her objects of analysis, Dancygier shows how specific textual choices prompt the construction and development of cross-mapping links between narrative spaces in a constant give-and-take of inferences from which an emergent story develops. Dancygier argues that the MSCI framework can be used successfully to explain not only the processes of narrative construction at the level of specific sentences, or even paragraphs but whole texts. Dancygier, therefore, examines the ways in which sentence-level textual features are used as prompts for the construction of higher levels of narrative structure. Esther Pascual’s chapter takes us from the world of texts to the world of criminal courts of the American judiciary to examine a particular kind of conceptual blend: fictive interaction. This is a class of conceptual blending in which discourse participants set up mental scenes in which they talk or otherwise interact with inanimate objects or dead persons as if those objects or persons were alive and in person. The examples of fictive interaction analyzed in this chapter are shown to perform three rhetorical tactics used in criminal trials, as revealed in official court transcripts, audio-visual material, and ethnographic notes from three highprofile murder trials. These Fictive interactional scenarios can be used to portray subsequent communicative performances of attorneys at trial as simultaneous turn-taking; they can be used in the introduction of material evidence – or the lack thereof – as the deceased victim testifying in open court; or they can be used to present of the final (accusatory) verdict as the jury’s verbal scolding of the defendant. In addition, the author argues that the tactical uses of fictive interaction mental spaces are dependent on contextual factors, such as the nature of the communicative event, the cognitive tasks involved, the precise legal issues at stake, and the discursive goals of the participants. It is also suggested that the study of blending types (see Brandt, this volume) – rather than anecdotal blending tokenscan throw much light on the reasons why some blends are more frequent and successful than others. Line Brandt’s chapter extends Pascual’s theory of fictive interaction and presents a framework for distinguishing different kinds of fictive interaction in online meaning construction. Two general uses of fictive interaction are outlined: semiotic blends and embedded metonymic enunciation. One type makes use of conceptual blending in the construction of a sign, and the other embeds a fictive
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.21 (21)
Introduction: Mental spaces and discourse analysis
enunciation in the current discourse context. These two rhetorical strategies afford a number of possibilities for making the discourse more engaging to the participants. Taking as a starting point the case of the Debate with Kant, an example given in Fauconnier and Turner 2002 to illustrate the fictivity involved in certain complex blends. Brandt analyzes additional examples of blended integrations of fictive interaction in complex sign structures occurring in situated communication. The generalizations from the proposed semantic analysis of the Debate with Kant example are extended to the rest of the examples. This perspective offers a method for analyzing semiotic blends, integrating pragmatics and semantics in cognitive analysis of natural language phenomena. The style of analysis presented by Brandt adds a semiotic dimension to blending theory, which is intended to address the pragmatic aspect of linguistic meaning. Todd Oakley and David Kaufer take up Dancygier’s multilayer analysis and extend it to place into alignment three layers of text analysis: the genre-layer of features and expectations; the artifact-layer of whole texts; and the grammar-layer of lexical and grammatical constructions. The authors argue that MSCI has a proper role in all three layers of analysis but that each layer brings with it its own theoretical and methodological challenges. Using a corpus of thirty-four articles published in the journal Hospital Practice the authors attempt to place in productive alignment an approach to rhetorical analysis that combines the genre analysis using the computer program Docuscope and factor analysis with the analytic machinery of mental spaces. Factor analysis reveals three distinct rhetorical strategies within the corpus at large and Oakley and Kaufer examine six with the corresponding highest and lowest concentrations of factors for each strategy and argue that each strategy calls for a slightly different arrangement of elements and relations in mental spaces representing similar scenes and situations. Analysis at the genre- and artifact-layers set the stage for a grammar-layer analysis of specific lexical and grammatical constructions, some of which count as unique features of particular texts others of which count as formulae present across texts. The authors argue that the threelayered analysis addresses two methodological problems in MSCI; the granularity of analysis problem and the problem of selectional bias. Anders Hougaard’s chapter studies the MSCI notion of “compression” (Fauconnier and Turner 2002) from the point of view of an CA study of “turn-packing utterances” in conversations on an American radio call-in show. His interactional approach consists in conducting full-blown interactional analyses of naturally occurring talk in interaction according to the standards of conversation analysis (CA) and then establishing interactional cognitive processes on the basis of directly observable interactional behavior. In this way, Hougaard seeks to explore the possible reality of the analytic notion of compression to coparticipants in everyday talk-ininteraction. Hougaard analyses a batch of data which shows how members close topics by summing up and construing previous turns at talk in a single utterance.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.22 (22)
Anders Hougaard and Todd Oakley
His findings lead him to conclude that the notions of blending and compression may capture cognitive work that the coparticipants themselves orient to. However, an important point in Hougaard’s application of the MSCI notions is that they are redefined according to the premises of social interaction. This means that the MSCI notions are thought of as “onstage,” interactionally enacted cognitive processes which are part and parcel of the interactional action, not as “hidden,” “backstage” processes inside the minds of individuals. Hougaard further proposes that the interactional cognitive processes are not ends in themselves but features of an interactionally achieved creation of a chunk of interactional memory. In this way interaction is not dictated by “inner” pre-existing psychological or cognitive processes; the psychology is created by the interaction. An ambitious inherent goal in Hougaard’s chapter is thus to think cognition inspired by CA. Robert Williams’s chapter presents readers with a variant of the CA tradition inspired by Charles Goodwin’s work on gesture and embodiment through a careful study of instructional discourse on the topic of time-telling. Williams seeks to understand just how a teacher’s actions-speaking, gesturing, manipulating objects, and so on-shape the learner’s construction of meaning, guiding the learner toward a coherent and, we hope, “correct” interpretation of some state of the world as it relates to the activity at hand. In order to achieve such an understanding, discourse participants must be able to relate discourse acts, which we can observe directly, to conceptual operations involved in meaning-making, which we must infer from various forms of linguistic and non-linguistic evidence, as has been done in the field of cognitive semantics. Williams argues that key to applying this theory to the analysis of instructional discourse requires Hutchins’ (2005) insight into how the material world can be used to anchor a blended mental space; such “material anchors” maintain sets of conceptual relations as they are acted upon to generate task-relevant inferences. On this occasion, Williams shows how material anchors for conceptual blends are used to analyze the discourse involved in teaching young children to read the time from an analog clock. Time-telling is an interesting case study because it involves constructing meaningful interpretations of states of a physical artifact – a clock face-and relating these interpretations to domains of human activity. The analyses illustrate two necessary steps toward arriving at an understanding of how instructional discourse works. Williams advances two arguments. First, explanations of institutional discourse must include detailed accounts of bodily actions and relevant aspects of the material setting in the unit of analysis. Second, a semantic theory supported by independent evidence (in this case, the theory of mental spaces and conceptual integration networks) to relate observable actions to underlying meaning-making operations, producing a coherent account with some explanatory power.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.23 (23)
Introduction: Mental spaces and discourse analysis
We are very fortunate to have three thoughtful commentaries from scholars representing the relevant constituencies of cognitive linguistics, conversational analysis, and discourse analysis. All three bring to the fore basic questions of interest to anyone working at the intersection of discourse and cognition. Alan Cienki, representing the cognitive linguistics constituency, addresses two questions of critical interest. The first question is: whose conceptual operations are being modeled in a mental spaces analysis? Is it an ideal speaker-hearer, is it the researcher’s interpretation, or is it the discourse participants themselves? The second, related question is this: is conceptual blending an “invisible process” and thus in principle closed to direct investigation? Cienki takes the ecumenical attitude that MSCI legitimately models all three, depending on the purpose of the study and its empirical status of the data but suggests that attention to data produced in real time by real speakers affords researchers the best opportunity for capturing conceptual operations of the participants themselves. Cienki’s stated preference for real-time data that includes prosodic vocal features aligned with gestures, postures, and facial expressions as the best evidence that interlocutors are themselves performing blending operations, and thus are not mere artifacts of analysis. Cienki offers readers additional mental spaces analysis based on this type of data as a means of answering no to the second question. Blending and mental spaces operations can be made directly accessible to researchers. Gitte Hougaard, representing the conversational analysis constituency, asks a very similar set of questions in her commentary, namely are mental spaces and blending operations processes that individuals do and then manage to calibrate with one another or are these operations processes individuals do together? In other words, are mental spaces and blends individual or communal achievements? Hougaard sees MSCI research as falling along the axis of individuality and communality. The first end takes the individual mind as the primary unit of analysis; the second end starts with the coordinated actions of multiple participants as the primary unit of analysis. Hougaard concludes that the individualist starting point leads to the an empirical impasse in which MSCI researchers model their own understanding of the discourse, whereas the communality starting point offers the most promising framework because it focuses on “how participants themselves understand their actions.” While Hougaard and Cienki share the view that micro-analysis of real-time discourse opens the window onto the relationship of discourse and cognition, they do not draw the same conclusions from extant research. For Cienki, micro-analysis of interaction data “provide further grounds for justifying blending analysis,” whereas for Hougaard, existent micro-analysis may indeed prove “fatal” to the MSCI hypothesis that conceptual blending is the way human beings think, for the data reveal that blending turns out to be “a way of thinking” but not “the way we think.”
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.24 (24)
Anders Hougaard and Todd Oakley
Paul Chilton, representing the broader discourse analysis constituency, takes up several issues in his commentary, from the role of context of as a cognitive operation to genre as a contextual operation. The thrust of Chilton’s commentary focuses on the psychology of joint attention as the most fruitful avenue of research in MSCI. This facet of cognitive theory leads Chilton to wonder if the study of discourse among clinical populations, namely autists and schizophrenics, might lead to better blending theories. If mental spaces and blending are not indiscriminate and haphazard processes, then studying the verbal behavior of populations that lack the general means of producing consistently coherent discourse may provide researchers with the crucial “photographic negatives” of the ways we think.
Acknowledgements We are grateful to the authors of the first seven chapters, for their stellar presentations at the 2005 International Pragmatics Association conference in Riva Del Garda and for their subsequent hard work, patience and forbearance in revising and preparing their chapters. We are also grateful to Chuck Goodwin, Gitte R. Hougaard, and Esther Pascual for their insightful comments as discussants. In addition, we would like to thank Andreas Jucker, editor of Pragmatics and Beyond New Series, for his very positive response to our proposal and commitment to the project and to Isja Conen, Acquisitions Editor at John Benjamins, for her efficient directions and assistance. We are likewise indebted to the anonymous referee whose constructive criticisms and comments have improved the volume. We will forever be in Gitte R. Hougaard’s, Paul Chilton’s, and Alan Cienki’s debt for reading the contributions and for sharing their thoughts with us all. Finally, neither of us could have brought this project to its conclusion without the love and patience of our respective families.
References Austin, J. 1962. How to Do Things with Words. Oxford: Oxford University Press. Barwise, J. and Perry, J. 1981. “Semantic Innocence and Uncompromising Situations,” in P. French, T. Uehling, and H. Wettstein, eds, Midwest Studies in the Philosophy of Language, vol. VI. Brandt, PAa. 2005. “Mental spaces and cognitive semantics: a critical comment.” Journal of Pragmatics, vol. 37, pp. 1578–1594. Chafe, W. 1994. Discourse, Consciousness, and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: The University of Chicago Press. Cicourel, A. V. 1974. Cognitive Sociology. New York: Free Press.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.25 (25)
Introduction: Mental spaces and discourse analysis
Coulson, S. and Oakley, T. (eds.). 2005. Conceptual Blending. Special issue of the Journal of Pragmatics, Volume 37, Issue 10. Coulter, J. 1979. The Social Construction of Mind. U.K.: Macmillan, London, 1979: U.S.A.: Rowman and Littlefield, N.J. Clark, H. 1996. Using Language. Cambridge: Cambridge University Press. Croft, W. “Toward a social cognitive linguistics.” New Directions in Cognitive Linguistics, ed. V. Evans and S.Pourcel. Amsterdam: John Benjamins. Descartes, R. 1641. Meditationes de prima philosophia. Edwards, D. 1995. “Two to tango: Script formulations, dispositions, and rhetorical symmetry in relationship troubles talk.” Research on Language and Social Interaction, 28 (4), 319–350. Edwards, D., and Potter, J. 2005. “Discursive Psychology, Mental States and Descriptions.” In Conversation and Cognition, H. te Molder and J. Potter (eds), 241–259. Cambridge: Cambridge University Press, pp. 241–259. Fauconnier, G. 1994 (1985). Mental Spaces: Aspects of Meaning Construction in Natural Language. New York: Cambridge University Press. Fauconnier, G. 1997. Mappings in Thought and Language. New York: Cambridge University Press. Fauconnier, G. 1999. “Methods and Generalizations.” In Cognitive Linguistics: Foundations, Scope, and Methodology, T. Janssen and G. Redeker (eds.), 95–120. The Hague: Mouton de Gruyter, pp, 95–127. Fauconnier, G. and Turner, M. 1994. “Conceptual Projection and Middle Spaces.” San Diego: University of California, Department of Cognitive Science Technical Report 9401. Fauconnier, G. and Turner, M. 1998. “Conceptual Integration Networks.” Cognitive Science, 22.2:133–187. Fauconnier, G. and Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Fillmore, C. 1982. “Frame Semantics.” Linguistic Meaning in the Morning Calm. The Linguistic Society of Korea (eds.), 111–137. Seoul: Hanshing Publishing Co. Fillmore, C. 1985. “Frames and the Semantics of Understanding.”Quaderni di Semantica 6.2: 222–253. Garfinkel, H. 1967. Studies in Ethnomethodology. Englewood Cliffs, N.J.: Prentice Hall. Grady, J, Oakley, T. and Coulson, S. 1999. “Conceptual Blending and Metaphor.” In Metaphor in Cognitive Linguistics, R. Gibbs and G. Steen (eds.), 100–124. Amsterdam: John Benjamins. Harder, P. 2003. “Mental Spaces: Exactly When Do We Need Them?” Cognitive Linguistics 14.1: 91–96. Hopper, R. 2005. “A Cognitive Agnostic in Conversation Analysis: When Do Strategies Affect Spoken Interaction?” In Conversation and Cognition, Jonathan Potter and H. te Molder (eds.), 134–158.Cambridge: Cambridge University Press. Hougaard, A. 2004. How’re we doing: An Interactional Approach to Cognitive Processes of Online Meaning Construction. Unpublished PhD dissertation. University of Southern Denmark. Hutchins, E. 2005. “Material Anchors for Conceptual Blends.” Journal of Pragmatics 37.10:1555– 1577. Hutchins, E. (1995). Cognition in the Wild. Cambridge, MA: MIT Press. Johnson-Laird, P.N. 1983. Mental models: towards a cognitive science of language, inference and consciousness. Cambridge, UK: Cambridge University Press. Lakoff, G. and Johnson, M. 1999. Philosophy in the Flesh. New York: Basic Books. Langacker, R. W. 1988. ”A Usage-Based Model.” In Topics in Cognitive Linguistics, B. RudzkaOstyn (ed.), 127–161. Amsterdam: John Benjamins.
TSL[v.20020404] Prn:25/01/2008; 8:03
F: PB170IN.tex / p.26 (26)
Anders Hougaard and Todd Oakley
Potter, J. and Molder, H. 2005. “Talking Cognition: Mapping and Making the Terrain.” In Conversation and Cognition, J. Potter and H. Molder (eds.), 1–54.Cambridge: Cambridge University Press. Sacks, H. 1992. Lectures on Conversation, vol. 1 and 2. Oxford: Blackwell Publishers. Schank, R.C. and Abelson, R.P. 1977. Scripts, Plans, Goals, and Understanding: An Inquiry into Human Knowledge Structures. Hillsdale, New Jersey: Lawrence Erlbaum Associates. Schutz, A. 1962. Collected Papers I. The Problem of Social Reality. The Hague: Martinus Mijhoff. Schutz, A. 1964. Collected Papers II. Studies in Social Theory. The Hague: Martinus Mijhoff. Sinha, C. and Jensen de López, K. 2000. “Language, Culture and the Embodiment of Spatial Cognition.” Cognitive Linguistics 11.1: 17–41. Spinoza, B. 1663. Principia Philosophiae Cartesianae. Amsterdam. Tomasello, M. 2000. “First Steps in a Usage Based Theory of First Language Acquisition.” Cognitive Linguistics, 11, 61–82. Turner, M. 1996. The Literary Mind. Oxford: Oxford University Press. Turner, M. 2001. Cognitive Dimensions of Social Science: The Way We Think About Politics, Economics, Law, and Society. Oxford: Oxford University Press. van Dijk, T. 2000. Cognitive Discourse Analysis. Webpage. http://www.discourses.org/ Unpublished/cogn-dis-anal.htm van Dijk, T. (ed.). 2006. Discourse Studies 8.1 Zlatev, J., Racine, T., Sinha, C. and Itkonen, E. (eds.). 2007. The Shared Mind: Perspectives on Intersubjectivity. Amsterdam: Benjamins.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.1 (27)
chapter
Connecting the dots1 Mental spaces and metaphoric language in discourse Todd Oakley and Seana Coulson How are metaphors understood in local context? We examine the rise of the expression “connect the dots” as it appears in the speech of counter-terrorism expert Richard Clarke in a radio interview. Although conceptual metaphor theory (Lakoff & Johnson 1980, 1999) provides important insights into the congruence of the connect-the-dots-metaphor with conventional metaphoric mappings between seeing and knowing, it cannot account for novel uses of the metaphor in the expressions used by these discourse participants. These novel metaphors are better described by mental spaces and conceptual integration theory, as involving conceptual structure integrated from multiple mental spaces. We examine how specific aspects of background, contextual, and linguistic knowledge contribute to their meaning in this particular instance. Examining prosodic features in the speech will help us account for the role vocalization plays in the structuring of ongoing discourse. This examination enables researchers to assess the compatibility between mental spaces theory and Wallace Chafe’s theory of discourse.
Introduction Consider the following excerpt from a radio interview with former White House Special Advisor on Counterterrorism, Richard A. Clarke: (34) no I don’t think so. (35) there’s uh some, (36) .. uh dots, (37) which are meaningless unless you put them together with lots of other dots. (38) an=d, . The authors owe a special debt to Alan Cienki for his patient reading and valuable comments and suggestions.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.2 (28)
Todd Oakley and Seana Coulson
(39) I understand what he’s sayi=ng. (40) but there are some dots that come out screa=ming at you. (41) uh do something now about me.
Out of context, this excerpt is a bit baffling, to say the least. One might wonder, for example, what sort of dots the speaker is referring to, how they can scream, and how they can refer to themselves using the first person pronoun. More importantly, what communicative goal is achieved by the speaker’s invocation of sentient screaming dots? This particular example involves creative elaboration of the conventionalized metaphoric expression “connect the dots,” but like many such uses goes considerably beyond its conventional meaning. We will argue that understanding the excerpt above relies upon three interrelated kinds of knowledge that we call elements of understanding. The first element is linguistic knowledge, such as knowledge of word meaning, grammar, knowledge of entrenched conceptual metaphors, as well as knowledge of particular metaphoric constructions. Closely related to linguistic knowledge, the second element is cultural knowledge of the relevant conceptual domains. In this case, it involves having background knowledge about the particular domains that are being discussed, including intelligence, national security, and international terrorism. In addition, since the example involves a metaphoric blend, it is necessary to have cultural knowledge of the source domain of the metaphor, the game of connect the dots, which will become clear when more of the co-text is provided. The third element of understanding is situational knowledge, and cannot be fully disentangled from the other two. A point emphasized by various stripes of discourse, rhetorical, and social theorists, understanding language utterances requires consideration of who is talking, who is being talked to, and what social relationships exist between them. Is the speaker representing him- or herself or speaking for another? What was said just prior to the current utterance, and how does it fit into the on-going conversation? To date, most blending analyses have focused on the first two elements of understanding. Our goal here, however, is to show how all three contribute to the comprehension of this metaphoric blend, with a special emphasis on its momentby-moment construction. This chapter represents our preliminary efforts to integrate mental spaces and conceptual integration theory (MSCI), as developed by Fauconnier ([1985]1994, 1997) and Fauconnier and Turner (2002), with Chafe’s (1994) theory of discourse based on conscious experience. In doing so we seek to place mental space theory on firmer analytic ground as both a method of discourse analysis and as a theory of discourse. The next section on Mental Spaces and Integration Networks offers a brief introduction to our framework. The section called To Connect the Dots presents an analysis of discourse that precedes the excerpt above (lines 26–31). The excerpt itself is discussed in the section called Scream-
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.3 (29)
Chapter 1. Metaphoric language in discourse
ing Dots, along with subsequent discourse that elaborates its meaning. Overall, the analysis shows how creative meanings draw heavily on cultural, linguistic, and situational knowledge.
Mental spaces and integration networks In this section, we present in broad outline the relevant features of the mental spaces and conceptual integration framework, beginning with a definition of “mental spaces,” their relationship to the “discourse ground”, and the networks in which they operate as discourse unfolds.2
Mental spaces In his seminal book on the subject, Fauconnier defines mental spaces as “constructs distinct from linguistic structures built up in any discourse according to guidelines provided by the linguistic expressions” (1994: 16). In this definition, Fauconnier stresses two things: first, that mental spaces are not themselves linguistic, and second, that they are the products of on-going discourse. For the purpose of this chapter, we define mental spaces as representations of the scenes and situations in a given discourse scenario as perceived, imagined, remembered or otherwise understood by the speaker. Mental spaces are used to package information about an interlocutor’s center of interest within an interactive context (c.f. Coulson & Oakley 2003). This definition is intended to capture the extent to which language users formulate and understand concepts by focusing not on individual properties but on simulating physical, social, and introspective scenes and situations (see e.g. Barsalou & Wiemar-Hastings 2005). We assume that mental spaces represent distinct physical, social, and/or introspective scenes and situations where attention is focused on a few salient elements therein. Along these lines, we assume that a mental space is governed by a semantic domain or domain matrices and that the particular characteristics of a mental space are determined by semantic frames for structuring the micro-features of a scene, such as role assignments, action and event sequences. We assume that as cognitive constructs, mental spaces are built up in real time. Moreover, unlike some prior discussions of mental spaces, we assume a given mental space can rep. For those familiar with other recent models of language processing, it may seem at first blush that MSCI resembles Pickering & Garrod’s (2004) Interactive Alignment model of discourse. While we find their model intriguing and agree with the central premise that interlocutors come to align linguistic representations at many levels (i.e., phonology, syntax, and semantics) automatically and without much conscious effort, the precise relationship between these two models are beyond the scope of the present discussion.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.4 (30)
Todd Oakley and Seana Coulson
resent events unfolding in time without having to posit a different space as each change occurs. A final assumption is that all mental spaces in a network are accessible to consciousness. Although discourse participants are not necessarily conscious of mental spaces per se (that is the provenance of analysts), they do have conscious access to the content of each mental space in the network. Following Mandler (2004: 59–89), we suggest that mental spaces comprise implicit and explicit declarative knowledge, with different facets of declarative knowledge made explicit and left implicit as discourse proceeds. This is why a mental space analysis can only claim to capture so much of the meaning potential in discourse, as one is never sure exactly what kinds of declarative knowledge a linguistic structure is going to activate.
Grounding Perhaps the most critical element in our mental space model is grounding. Although largely absent in the mental spaces and blending theory developed by Fauconnier and Turner, this notational device has gained increasing prominence among cognitive linguists working in the area of discourse analysis, conversation analysis, and semiotics. The term grounding comes from Langacker (1999) who uses it to discuss grammatical elements that evoke the speaker’s situation, elements such as tense markers, quantifiers, and determiners. In cognitive grammar, grounding refers “to the actual speech event, its participants, and its immediate circumstances” (1999: 79). On our model, however, grounding more closely resembles the semiotic space proposed by Brandt & Brandt (2005: 19–22). In this capacity, grounding represents the discourse participants’ acts of engaging in discourse, of constructing meaning. Grounding allows the theory of mental spaces to consider explicitly how situational knowledge contributes to the understanding and management of discourse. Grounding involves specifying (1) the discourse participants and their roles, (2) the rhetorical situation that serves as the immediate local context for the current communicative act, (3) the situational and (4) argumentative relevance of the mental spaces network. The grounding of networks is such that a given mental space network functions as a discourse scenario, or a rhetorical or discourse move that can be characterized in terms of its illocutionary force and perlocutionary effects. Integration networks Mental space networks are comprised of various different sorts of spaces that operate in “dialogue” with the grounding. In the excerpt analyzed here, the networks involve a reference space, a presentation space, and a blended space (see Brandt & Brandt 2002 for a somewhat similar approach). The reference space relates to
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.5 (31)
Chapter 1. Metaphoric language in discourse
the discourse topic active at the moment of speaking and listening. The presentation space is a predicating space and can be highly figurative. In metaphoric examples, the presentation space involves conceptual structure from the source domain in conceptual metaphor theory. The blended space involves the integration of select elements from the reference and presentation spaces. In many cases, the blended space functions much as a “flight simulator” in that it sets up the mental conditions for thinking about something in a particular way. In such cases, the referential structure in the blended space is far less important than its pragmatic implications for the reference space. Mental space networks can function both iteratively and recursively. They can function iteratively in virtue of the fact that the same array of spaces (i.e., spaces representing the same set of scenes and situations) can be run repeatedly at different times in the discourse, and can serve different pragmatic and rhetorical functions depending on the distribution of attention within the array at any given moment in the discourse. They can function recursively in virtue of the fact that the initial network of spaces serve as catalysts for the construction of a different mental space network serving different pragmatic and rhetorical functions, oftentimes by embedding a mental space from the initial network into a subsequent network. Indeed, this is precisely what happens in the example under discussion in this chapter, as one of the participants appropriates a mental space from his intelocutor’s network and embeds it into a new network to serve his own rhetorical goals.
To connect the dots Recall the excerpt of the focal conversation, reprinted here to facilitate analysis: (34) no I don’t think so. (35) there’s uh some, (36) .. uh dots, (37) which are meaningless unless you put them together with lots of other dots. (38) an=d, (40) I understand what he’s sayi=ng. (41) but there are some dots that come out screa=ming at you. (42) uh do something now about me.3
. The transcript conventions used in this study adhere to Chafe’s notation (1994) with these notable exceptions. We conflate loudness and pitch change using underlines, whereas Chafe treats them separately with boldface (loudness) and acute and grave accents (pitch change). We
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.6 (32)
Todd Oakley and Seana Coulson
Our analysis of this excerpt begins with an account of the situational knowledge needed to understand the excerpt, including knowledge about participants, their circumstances, and setting. The excerpt under analysis has three participants. Two of the participants are present in the immediate grounding, and one is virtually present as the discourse unfolds. The first participant is Richard Clarke, former White House Coordinator of counter-terrorism for the Clinton Administration and the George H.W. Bush Administration. Clarke is the author of Against all Enemies, a trenchant critique of the George W. Bush Administration’s handling of the “war on terrorism” as well as their handling of intelligence leading up to the 9/11 terrorist attacks. The second participant is Dave Davies, a journalist who is filling in for the regular host (Terry Gross) of Fresh Air, a popular interview program on National Public Radio. Davies has read Clarke’s book and is interviewing him for the radio audience. The third “virtual” participant in the conversation is Richard Posner, a federal judge, legal scholar, and conservative columnist. The exchange takes place within the genre constraints of an interview radio program in which Davies asks open-ended and leading questions designed to prompt a response. Such interview strategies may include the interviewer playing the role of devil’s advocate, usually in the form of paraphrasing or reporting the argument of someone with a position diametrically opposed to that of the guest. On the other side, the interviewee comes prepared with his own talking points and agenda. The setting is a taped radio broadcast, meaning that Davies and Clarke are the only two “talking” participants who can effectively influence the course of the conversation, although the editors and producers can decide which parts of the conversation to air. In contrast to spontaneous conversation, little overlap or interruption occurs, due to the format of this particular program. The rhetorical context of the exchange is the differing assessments of Clarke (the actual guest) and Posner (the virtual participant) of the Congressional report by the 9/11 Commission. The report placed considerable blame on the intelligence community and made policy recommendations for improving the intelligence gathering and interpretation processes. Clarke agreed with the report’s overall findings on intelligence, but did not think it covered the most critical issue: how the Bush Administration ignored intelligence gathered on terrorist activities in the United States and Europe prior to the attacks. The virtual Posner, in contrast, disagreed with the 9/11 Commission’s assessment that the intelligence community was to blame for the attacks.
added special symbols for acceleration (« ») and deceleration (» «) not employed by Chafe. (See appendix for a full description of these symbols.)
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.7 (33)
Chapter 1. Metaphoric language in discourse
Davies’s contribution The exchange begins as Davies says: (1) there’s a clearly (2) a perceived need here for, (3) . . .uh sensitive information to be shared, (4) even among investigators among different agencies. (5) uh, (6) ..who are working, (7) on..issues that that «information about terrorism»might arise. (8) in– (9) in effect to connect the dots.
Chafe’s (1994) principal contention is that spoken discourse such as that in the above excerpt is produced and comprehended in prosodic “spurts” he calls Intonation Units (IUs). These units guide meaning construction because they possess prosodic instructions for understanding what information is prominent in the speaker’s consciousness at any given moment. Hearers then unpack that information according to the prosodic guidelines of the perceived speech. In our theory, then, an IU helps discourse participants to construct mental spaces and mental space networks that are sufficiently similar in their semantic and pragmatic facets to facilitate interaction. Our analysis focuses primarily on one of the three types of IUs (substantive) leaving the other two, (regulative and fragmentary), in the background. While substantive units present ideas, states, and referents, regulatory units perform discourse functions such as taking the floor (e.g., “uh” and “well”4 ), holding the floor (e.g., “um”), signaling agreement and permitting continuance (e.g., “Mhm”), (see Chafe 1994: 63–65). On the other hand, fragmentary units are essentially “false starts” or aborted attempts by the speaker to create a substantive IU. Both regulatory (e.g. line 14) and fragmentary (e.g. line 21) units are evident in this exchange, and, in many instances the regulatory unit comprises a sub-unit of a substantive IU as a floor holding device. However, the interview format already provides a regulatory structure so that instances of these are far less frequent than in casual conversation, and, the editors and producers ensure that fragmentary IUs be held to a minimum. . We mark “uh” as a single intonation unit only when prosodically “bounded” by pauses or continuative or terminal contours. Most of the instances in this exchange do not appear to be so bounded and are thus treated as onsets of an intonation unit.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.8 (34)
Todd Oakley and Seana Coulson
The transcription above (a complete version of which can be found in the appendix) breaks up the spoken discourse into intonation units by focusing on pauses based on Chafe’s criteria for a coherent IU. A coherent IU includes perceptible pauses preceding and following a string of words, a detectable pattern of deceleration and acceleration within a string of words, an overall decline in pitch level, the falling pitch contour at the end of a string of words, or creaky voice at the end of a string (Chafe 1994: 60). There is no set length to a substantive IU. According to Chafe, however, the mean length is around five words but can be lengthened extensively with acceleration. Our analysis will focus primarily on vocal prominence established through loudness and vowel and consonant lengthening. Examining the content of the intonation units in our example, it is clear that there is no one-to-one mapping between material in an IU and material in a single mental space. For example, in this excerpt the first IU serves to open a mental space that is lexically cued by “perceived need” (line 2). This space will represent normative information about how the intelligence world should be, and stands in disanalogy to present reality. In other words, the situation depicted in the reference space is being construed in two ways in order to profile the contrast between the real situation and the preferable one. The second IU specifies the structure that goes in this normative construal: sharing information (line 3). The fourth IU further specifies the sharing information structure by noting that investigators in different agencies will be the ones sharing the information. The sixth and seventh IUs add still more information about the investigators: investigators working on terrorism issues. Finally, the ninth IU signals a new space with the words “in effect,” a hedge that rephrase the conception prompted by the previous utterance. Thus, lines 3, 4, 6 & 7 successively refine information in the normative version of the reference space, and then 9 compresses the information set up by the others with the metaphoric expression “connect the dots.”
Cultural and linguistic knowledge As noted earlier, understanding the meaning of an expression such as “connect the dots” requires cultural knowledge, a good deal of linguistic knowledge, as well as situational knowledge evolving over the course of the interaction. In this section, we focus on the cultural and linguistic knowledge our interlocutors might conceivably bring to bear in their construction of the meaning of “connect the dots.” Generically, “to connect the dots” means to understand the relationship between apparently isolated bits of information. However, the expression has a particular conventionalized meaning in the context of the 9/11 terrorist attacks on the World Trade Center in New York city and the Pentagon in Washington, DC, and relates to the American government’s failure to predict those attacks. One popular explanation of how 19 hijackers were able to outsmart the collective resources of
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.9 (35)
Chapter 1. Metaphoric language in discourse
the FBI (Federal Bureau of Investigation), the CIA (Central Intelligence Agency), and the NSA (National Security Agency) is that while a considerable amount of information relevant to the planned attacks was known to various members of these agencies, it was distributed amongst them such that no single agency had enough information to take preventive action. In this context, “failure to connect the dots” refers to the failure by US intelligence agents to understand the relationship between different facts about individuals with links to terrorist groups. Although metaphoric uses of “connect the dots” occurred long before September 11, 2001, their frequency increased dramatically in the latter part of 2001, as this expression was used almost exclusively to refer to the failure to predict and prevent the 9/11 terrorist attacks. In fact, during the first part of 2002 this particular use of “connect the dots” was so common it actually became a cliché, provoking self-conscious rumination on the part of linguistically minded journalists. For instance, here is New York Times columnist William Safire’s (2002: 22) explanation of the expression: The origin appears to be in a game run in newspapers early in the 20th century, based on a children’s game. A field of apparently random dots is displayed and numbered, and the dots are joined by the player in numerical order. Sometimes, clues to running lines between the dots are given below. As the game is played, the short lines begin to take the shape of an object, and lo! A pattern or image leaps up from the page.
Given the restricted use of the phrase “connect the dots” to refer to intelligence gathering relevant to 9/11 it would be possible to understand its meaning simply by memorizing its conventional meaning in this context. However, we suggest that the success of this expression as a meme partly reflects the fact that it is supported by linguistic knowledge, including knowledge of the meaning of “connect.” The verb “connect” has an entrenched literal meaning that is relevant for the game of Connect the Dots (described by Safire), as well as an entrenched metaphoric meaning relevant to the comprehension of abstract relationships. Even if one knew nothing about the Connect the Dots game, it might be possible to infer the contextually relevant meaning given knowledge of the meaning of “connect” and cultural knowledge of the prevailing explanation of the 9/11 attacks as described above. Moreover, “connect the dots” is even more meaningful to someone with knowledge of the Connect the Dots game due to its potential interpretation via the entrenched conceptual metaphor KNOWING IS SEEING (Lakoff & Johnson 1980, 1998; Sweetser 1990). This conceptual metaphor is motivated by the existence of metaphoric expressions such as the following in (1)–(7). (1) I see what you mean. (2) Oakley’s theory is murky. (3) His motives are transparent.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.10 (36)
Todd Oakley and Seana Coulson
(4) Her lecture was opaque. (5) That’s a myopic approach. (6) This book opened my eyes to new ideas. (7) These data shed new light on an old hypothesis.
All of these expressions involve positing particular mappings between elements and relations in the domain of seeing and the domain of knowing. For example, statements about the seeing agent are understood to pertain to the knowing agent; statements about the viewed object are understood to pertain to the object of knowledge; and, the act of seeing corresponds to the act of understanding. Consequently, statements about the visual acuity of the seeing agent have implications for the agent’s ability to understand, and the clarity of the seeing agent’s percept corresponds to the quality of the knowing agent’s understanding. Seeing Agent (subject) Viewed Object Percept Act of Perceiving Agent’s Visual Acuity Clarity of Scene Percept
→ → → → → →
Knowing Agent (subject) Object of Knowledge Knowledge Act of Understanding Agent’s Ability to Understand Quality of Understanding
The “connect the dots” blend co-opts the KNOWING IS SEEING metaphor and applies it to the game of connect the dots by adding a few new mappings. The most important of these is the mapping between dots and information that is quite simple to establish because a unit of information is often construed as a point. The metaphoric significance of the expression “connect the dots” is bolstered by the fact that there is one entrenched meaning of connect that suits the game, and another that suits the epistemic domain. Because the percept corresponds to the knowledge in the KNOWING IS SEEING metaphor, connected dots correspond to known relationships between different pieces of information. In the game, connected dots afford pattern recognition; in the epistemic domain, knowing relationships between different pieces of information allows the inference of new information. Further, because the clarity of the percept maps onto the quality of the knowledge, unconnected dots that yield an unclear percept map onto poor knowledge. Moreover, the unseen pattern in the game maps onto the fact that important information is unknown to the subject of knowing. The integration network for “connect the dots” involves a presentation space that pertains to the Connect the Dots game, while the reference space pertains to national security and intelligence. In the blended space, information gleaned by intelligence agents map onto dots on a two-dimensional plane, and the pattern implicit in these dots (structure projected from the presentation space) maps onto terrorist plots in the reference space. In the blend, it is possible for the intelligence
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.11 (37)
Chapter 1. Metaphoric language in discourse Grounding space
Presentation space
Reference space
Richard Posner (speaker)
Agents gather, collate, and interpret “bits” of information about suspected terrorist activities
Connect-the-Dots
Richard Clarke (hearer) Situational relevance
Base
Speaker makes a claim
Situation Setting
Argumentative relevance
Invitation to respond to the claim
Terrorist Plots Blended space
Dave Davies Richard Clarke
Real participants
Illocutionary Force: Speaker makes a controversial assertion
Unconnected “Dots” of information about terrorist activities
Elaboration space
Pragmatic Implication: The Bush administration did everything to prevent the Al Qaeda attacks
Figure 1. Connect-the-dots metaphor
officer to see terrorist activities represented in the dots. The intelligence officer in the blend draws lines between the dots just as the child does in the presentation space. However, while the child sees a pleasing picture in her drawing, the intelligence officer gains a growing understanding of impending terrorist attacks. Integration networks such as that in Figure 1, however, must be understood as operating in dialogue with grounding. The particular structure activated by a linguistic utterance will vary with the participants, the setting, the co-text, along with other aspects of the situational context. As noted above, lines first 7 lines successively refine structure in the reference space characterizing intelligence agencies, and appeal to a culturally shared explanation of the 9/11 attacks as being due to the failure of intelligence agencies to combine disparate pieces of information. Interpretation of the metaphoric utterance in line 9 will depend upon the extent of a given listener’s cultural and linguistic knowledge, her willingness to engage with the material, and the way in which various contextual factors affect the structuring of mental spaces. Indeed, even prosodic aspects of an utterance’s production may influence the construction of meaning. For example, line 9 in which “connect the dots” initially
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.12 (38)
Todd Oakley and Seana Coulson
appears after a noticeable terminal contour in line 7 and false start in line 8. The terminal contour may signal the end of the structuring of the reference space and allow the listener to anticipate the need to activate structure from a novel domain. Further, there is no pause between the phrases “in effect” and “to connect the dots,” suggesting they are part of one substantive intonation unit and signaling to the hearer that “connect the dots” is a summary expression epitomizing the sentiments expressed in the previous four intonation units. The relevance of line 9 is thus established by the first four intonation units describing the need for intelligence agencies to share information. Although Davies’s language alone does not support the mapping between sharing information and connecting dots, he can rely on his listeners having at least some subset of the cultural and linguistic knowledge described above. We speculate that if “connect the dots” did not already have a packaged, compressed meaning in the domain of intelligence gathering, it would be less suitable for this summarizing function. Moreover, once Davies utters it, the blended space becomes activated and available for creative elaboration.
Posner’s argument Davies follows by articulating an argument by Richard Posner to absolve the White House and the wider intelligence community from responsibility for failing to prevent the terrorist attacks on the World Trade Center and the Pentagon. (10) ..and Richard Posner, (11) in a piece in The New York Times recently (12) ..um.. (13) in reviewing the 9/11 Commission Report, (14) criticized the idea that, (15) as he put it, (16) the failure to prevent the attacks was due to, (17) ..a failure to collate the «bits of information possessed by different security agencies», (18) especially the CIA and the FBI. (19) ..um, (20) he says, (21) the best bits of information were not obtained until, (22) ..the month or so before the attack, (23) and he says, (24) its unrealistic to suppose they could’ve been »inter=grated«,
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.13 (39)
Chapter 1. Metaphoric language in discourse
(25) »and understoo=d« in time to detect the plot. (26) ..um in effect he’s saying, (27) «there’s all this information flowing around there», (28) and it’s, .. (29) w— (30) we weren’t just pick this pick those dots out, (31) and connect them in time. (32) is he right?
In mental spaces analysis, we posit three distinct versions of interactive grounding that develop during Davies’s turn, as depicted in Figure 2. The first is the ontological ground applicable from lines 10–15, 20, and 23. In the ontological ground, Davies and Clarke are the participants, the setting is a taped interview program broadcast to millions of listeners, and the situational relevance constraining their interaction is the 9/11 Commission Report. The argumentative relevance of this grounding is that the speaker is setting up a context for a controversial claim to which he wishes Clarke to respond. To this end, Davies engages in what Chafe calls direct speech in lines 16–18 and 21–24 as he produces a verbatim reproduction of Posner’s published thoughts in a new interactive context involving Richard Clarke, a displaced grounding scenario with Posner as the absent discourse participant: Davies is speaking for Posner. A third interactive grounding is similarly displaced, this time as an indirect enactment of Posner’s words evidenced in lines 21, 22, 24, 25, and 27–31. In this third grounding, Davies speaks as Posner, with stresses and prominences of his words taken as Posner’s own consciousness, thereby inviting Clarke to interact fictively with Posner (cf. Pascual 2002). The situational relevance remains the same, but the argumentative relevance of this grounding is Posner’s disagreement with the 9/11 Commission’s findings. By introducing Posner’s words with the present tense verb “says,” Davies brings the past speech event from the ‘there-and-then’ of a past speech event to the ‘here-and-now’ of the present conversation. Schiffrin (1981: 60) has observed that speakers more frequently invoke the historical present and other forms of immediate modes in the climax of a narrative. Davies’s tendency to speak as Posner can be understood in a similar light, as he prompts Clarke to respond directly and immediately to the words and ideas of Richard Posner by making them the subtopic of discussion. This third grounding space is in fact an integration of the first two, as it involves a hybrid identity of Davies speaking in his own voice, but as Posner. This discourse tactic allows Clarke and the hearers to experience the dynamic unfolding of Posner’s ideas as performed by Davies. The second and the third grounding can be differentiated by the use of the “connect the dots” blend, as Posner himself does not use it, while Davies does when
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.14 (40)
Todd Oakley and Seana Coulson Grounding
Grounding space 1
Dave Davies
Richard Posner
(speaker)
(speaker)
Richard Clarke
Richard Clarke
(hearer)
(hearer) Situational relevance
Base
Base
Responses to 9/11 Commission Report
Situation
Situation
Setting
Situational relevance
Hearer agrees with 9/11 Report
Setting Argumentative relevance
Lines 10-15; 20 &23
Argumentative relevance
Speaker sets up context for the response
Lines 16-18; 21 -24
Speaker disagrees with the 9/11 report
Grounding space 2 Dave Davies Richard Clarke Real participants
Davies = Posner (speaker)
Richard Clarke (hearer)
Base Situation
Speaker justifies the claim
Setting
Dave Davies Richard Clarke Real participants
Situational relevance
Argumentative relevance
Lines 21, 22, 24, 25; 27-31
Speaker specifies disagreement
Figure 2. Varieties of grounding
acting as Posner in lines 30, “we couldn’t just pick this pick those dots out.” Therefore, the metaphoric expression itself issues from ontological grounding as in the initial segment (lines 1–9) where Davies introduces the blend only after detailed structuring of the reference space, in the excerpt below, Davies presents Posner’s fairly abstract characterization of the issue in lines 16–18 and then paraphrases it in terms of the connect the dots blend in lines 30–31 when speaking as Posner. Differences in grounding correlate with focus on either the reference space of intelligence agencies or the blended space in which intelligence agents attempt to connect dots. Prosodic emphasis on “best bits” in line 21 focuses attention on the reference space, where bits of information can be hierarchically ranked for their usefulness. Similarly, the stress on “unrealistic” in conjunction with the vowel lengthening in “integrated” and “understood,” focus on the reference space and evoke the disanalogy between Posner’s reference space models and the conventional implications of the connect-the-dots blend. Davies helps highlight the disanalogy when he refers explicitly to “dots” in line 28 and uses the verb “connect” in line 29 to shift the focus to the blended space and its presentation input.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.15 (41)
Chapter 1. Metaphoric language in discourse
On the standard reading of the blend, failure by intelligence agents at a children’s game is seen as a sign of their incompetence. Posner’s point, however, is that the agent’s task was far more complex than that in the children’s game and that the epistemic connections between intelligence data were less obvious than the ones children must draw between dots. The dots in the blend are not numbered in the way that they are in the children’s game. Further, the dots in the blend are not all there at the outset, but emerge in real time. Moreover, they are not all equally relevant for solving the puzzle as the intelligence officers in the blend have to “pick” which dots to connect, and have to do so under time pressure. Although the mappings in 21–29 are the same as in the conventional “connect the dots” blend, the implications are not. That is, the dots still correspond to information, and connecting the dots corresponds to understanding the relationship between the different pieces of information. However, failure to make connections does not necessarily stem from failure to share information. Instead it could be due to insufficient time to consider the relevant relationships. Interestingly, this point is initially made in abstract terms in Posner’s own words, and subsequently expressed by Davies via structure in the blended space. Novel structure in the blended space – such as agents picking which dots to connect and doing so under time pressure – reflects information recently added to the intelligence-gathering space about the timely integration of information. This excerpt shows how in context meanings can differ substantively from their standard meanings as speakers adapt them to suit their rhetorical goals. We speculate that the intelligibility of this creative extension of “connect the dots” is related to the frequent use of this phrase to discuss intelligence gathering, the entrenched nature of the underlying mappings, and the fact that the standard connect-the-dots blend has been established in the immediately preceding utterances. One possibility is that because Davies introduces the blend in its most standard verbiage and in a very standard context, it sets it up as active and supportive of further blending activity.
Screaming dots Indeed, the “connect the dots” blend seems to provide the conceptual scaffolding upon which Clarke builds his response. It therefore sets the stage for Davies to ask the question, “Is he right?” With this question, the attention shifts back to the initial interactive ground. It is in this context that Clarke refers to “screaming dots.” (34) no I don’t think so. (35) there’s uh some, (36) .. uh dots,
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.16 (42)
Todd Oakley and Seana Coulson
(37) which are meaningless unless you put them together with lots of other dots. (38) an=d, (40) I understand what he’s sayi=ng. (41) but there are some dots that come out screa=ming at you. (42) uh do something now about me.
In fact, the full meaning of his remarks is not apparent until Clarke completes his turn. (42) uh, (43) and.. when we, (44) uh, (45) knew..that there were.. Al Qaeda people (46) uh, (47) going to Malaysia, (48) going to Kuala Lampur, (49) ..uh and meeting there, (50) and plotting there. (51) ..we became very concern=ed (52) ..uh because it looked like the »kind of« meeting, (53) .. uh where Al Qaeda people, (54) operational people, (55) get together, (56) . . .and go over the details, (57) .. of some impen=ding attack. (58) so if we had then learned, (59) ..that some of those people in that meeting, (60) were in Southern California, (61) ..and ent@ered the United States, (62) ..that would have been the kind of dot, (63) uh that didn’t need a lot of connecting. (64) that would have screamed out at you, (65) do something about me now. (66) and I think, (67) y’know it’s,
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.17 (43)
Chapter 1. Metaphoric language in discourse
(68) twenty-twenty hindsight and all that, (69) but, (70) ..I really do belie=ve, (71) that if I were sitting in my old job in the White House, (72) and I had see=n a report that said that, (73) ..uh I would have ma=de the FBI, (74) uh a– (75) and frankly FBI headquarters would have wa=nted to, (76) go=all out, (77) to find those two guys.
Like Posner, Clarke is arguing that the “connect-the-dots” game is a bad analogy for understanding the situation. However, he does so by constructing yet another blend that acknowledges the presentation (Connect the Dots game) and reference (intelligence-gathering) spaces already set up. He starts by reiterating the standard version of the blend in line 36 and 37, showing that he understands it, and even acknowledging that there are times when it is apt. It is relevant here that “dots” itself is not emphasized. This may be because it is already semantically active in the discourse. Alternatively, it may signal Clarke’s discomfort with the standard analogy, if we attend closely to lines 34–42. (34) no I don’t think so. (35) there’s uh some, (36) .. uh dots, (37) which are meaningless unless you put them together with lots of other dots. (38) an=d, (40) I understand what he’s sayi=ng. (41) but there are some dots that come out screa=ming at you. (42) uh do something now about me.
We might conjecture that the disfluency before “dots” in 36 results because of a last minute change in word choice. The plural count noun “dots” does not agree in number with the verb in line 36 (cf. “There’s some dots,” “There’s some information,” and “There are some dots”), but does agree with the verbs in line 37. In fact, one could replace “dots” with “facts” in lines 36 and 37 and leave Clarke’s meaning intact, raising the issue of why he frames his reply in terms of dots. We can only speculate that some combination of cognitive and social factors, such as the active nature of the blend and the demands of conversational relevance, are at play.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.18 (44)
Todd Oakley and Seana Coulson
Beginning with the active structure evoked by Davies, Clarke sets up a novel dots blend in lines 41–42. Clarke distinguishes the new dots blend from those evoked by Davies via the use of “but” and contrastive prosody on the quantifier “some” in line 41. In Clarke’s blend, the reference space is the same as that in the prior blend, intelligence information pertaining to terrorist activities. However, Clarke introduces a novel presentation space that involves a screaming person. In this space, the person’s screams could only go ‘unnoticed’ by others unless they were purposefully ignoring them. Thus in Clarke’s blend, the dots are not pieces of a larger picture, but sentient beings in their own right. In the blended space, the dots are bits of information that shout instructions at the intelligence agents. Just as the “connect the dots” blend recruits existing metaphoric mappings, Clarke’s blend, too, recruits an existing blending schema, that of fictive interaction (Pascual 2002). Fictive interaction involves the use of frames for the structure of ordinary communicative acts to animate all sorts of processes. Fictive interaction is particularly common when the reference domain involves interpreting information, as for example when a lawyer says that the facts in a case “tell a story.” Because the reference space in this example also involves interpreting facts, fictive interaction is particularly apt. In Clarke’s blend, the dots do not simply speak, they scream. The implication from this screaming dots blend, then, is that the only conclusion to draw is that the responsible government agencies were ignoring the warning signs. The vowel lengthening on “screaming” in line 41, and the contrastive higher pitch on ‘now’ in line 42, both signal blending operations. The first dynamically recruits a new presentation space for screaming. The second reinforces the time-sensitive nature of activities in the reference space. The active space is the blended space in which intelligence officers interact verbally with the dots. In line 41, Clarke places himself in this space with a screaming dot. Then in 41 he quotes the dot’s verbal utterance in his fictive interaction. Having set up the screaming dot blend, Clarke can then return to it several moments later (in line 64) after adding structure (in lines 42– 63) to the reference space with which it is linked. Figure 3 presents both Clarke’s mental spaces network for Clarke as building up recursively from the network constructed by Davies. One salient difference is in the grounding, as Clarke is the speaker and Davies the hearer. With line 58, “so if we would have learned,” Clarke shifts into a counterfactual mode in which he places himself and like-minded agents in the same situation the real agents were in at the time. Listeners know they are in a counterfactual situation by Clarke’s emphasis on “if ” at the beginning of the IU and the past tense of “learn” at the end of the unit followed by a slight pause. In line 61, Clarke is laughing as he says the verb “entered.” Such vocal prominences do not just signal what ideas are of keen interest to the speaker, they can also be used to convey a speaker’s attitude or perspective toward the ideas presented. The presence of laughter in this instance highlights
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.19 (45)
Chapter 1. Metaphoric language in discourse Grounding space
Presentation space
Reference space
Richard Clarke (speaker)
Random “Dots” of information about terrorist activities
Attention!
Dave Davies (hearer) Situational relevance
Base
The anticipated response
Situation Setting
Argumentative relevance
Speaker makes a counter claim Do something about me now!
Blended space
Dave Davies Richard Clarke
Real participants
Illocutionary Force: Speaker accuses administration of negligence
The White House and National Security Council ignores these screaming dots
Elaboration space
Pragmatic Implication: The sitting president should not be reelected
Figure 3. Screaming dots
Clarke’s own derisive stance toward the way intelligence was being handled by the Administration in the months leading up to the September eleventh attacks. With line 62, we get prosodic emphasis at the beginning of the distal demonstrative “that” (which signals Clarke’s own position as an outsider looking in) and on the noun “dot” at the very end. Line 62, then, shifts attention back to Clarke’s blend, and lines 63–65 reiterate the screaming dot scene, this time with heavy prosodic emphasis on the temporal adverb “now.” Lines 35–42 and 62–65 are repeated below in order to show the similarities in Clarke’s wording, as well as similarities in the information in the intonation units in which the screaming dots blend is evoked. (35) there’s uh some, (36) .. uh dots, (37) which are meaningless unless you put them together with lots of other dots. (38) an=d, (40) I understand what he’s sayi=ng.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.20 (46)
Todd Oakley and Seana Coulson
(41) but there are some dots that come out screa=ming at you. (42) uh do something now about me. (62) ..that would have been the kind of dot, (63) uh that didn’t need a lot of connecting. (64) that would have screamed out at you, (65) do something about me now.
The meat of Clarke’s argument, however, is the information he presents in lines 42–61 – facts about the activity of known terrorists in early 2001 and facts that were known to various intelligence agents in the United States. Moreover, the interpretability of the screaming dots blend is much greater in 62–63, i.e. after the intervening information has been presented. Just as Davies was able to alter the standard connect-the-dots blend after evoking structure in the reference space, Clarke’s novel blend becomes much more compelling after the addition of information to the reference space. How knowledge of the actual topic changes discourse participant’s likely understanding of a conventional metaphor as a result of its being applied to that domain is not easily handled by cognitive accounts predicated on the asymmetrical flow of inferences from a concrete to abstract domain. In the cases we have examined, understanding the more abstract domain appears to facilitate comprehension of the figurative blends.
Conclusion We argue that creative metaphoric blends rely heavily on conventional linguistic knowledge, such as the metaphoric meaning of “connect” and conventional mapping schemes such as that between seeing and knowing. Moreover, such metaphoric expressions are often given particularized cultural meanings. In the case of “failure to connect the dots,” for example, it refers to intelligence agencies’ failure to prevent the terrorist attacks on September 11, 2001 and particular mappings between intelligence information with dots, and terrorist plots with patterns in the dots. Even in creative extensions of connect-the-dots as in Richard Clarke’s screaming-dots blend, people seem to appeal to established blending schemes such as fictive interaction. Our exploration of the importance of situational knowledge in understanding blends in context suggests that diagrams common in blending analyses are somewhat deceptive. Indeed, a common reaction to such diagrams is that they require the listener to activate an unrealistically large amount of conceptual structure given working memory limitations. However, the totality of the information represented in the diagrams is never simultaneously present in the observer’s
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.21 (47)
Chapter 1. Metaphoric language in discourse
mind. Rather, the diagram is an atemporal representation of an understanding that evolves incrementally over the course of several minutes. We suggest that Chafe’s notion of active, semi-active, and inactive information may provide a valuable framework for a more refined model of how these processes actually occur. With each intonation unit, the speaker activates a small amount of conceptual structure in mental spaces. Successive IUs have been seen to provide additional information about various elements in a single mental space. One function of metaphoric expressions such as connect-the-dots is to compress a complicated scenario that takes several IUs to build up into an easily conceptualized scene. One function of prosodic information is to cue perspective shifts, and to indicate which mental space is currently in focus. Once a blend is established into the context, or in Chafe’s terminology becomes active, speakers are free to elaborate it, or to alter it creatively by introducing a new presentation space to change the emergent inferences. In fact, if the example examined here is at all representative, it is quite difficult to abandon a blending network that has been established into the context. The standard connect-the-dots blend was not compatible with either of the argumentative positions expressed by Davies and Clarke, yet they both used terms “connect” and “dots” to summarize their points. Davies (qua Posner) objects to the idea that intelligence data includes only relevant information, that it is all available simultaneously, and that the relationship between different facts is easy to establish. Clarke, on the other hand, objects to the very idea that predicting the 9/11 terrorist attack required understanding the relationship between seemingly unrelated facts, yet he, too, begins with the connect-the-dots blend. As argued above, the culturally shared connect-the-dots blend is supported by a particular set of mappings between the game Connect the Dots and a culturally acknowledged framing of the 9/11 intelligence failures, as well as by more generally entrenched mappings between, for example, connecting and understanding. Although the former does not suit Davies’s argumentative purpose, the latter does. Consequently, his blend does not correspond to the game, but does recruit key mappings that support the original blend. However, because Clarke objects to the very premise of the mappings in the connect-the-dots blend, he explicitly disputes the applicability of the term “connect,” recruiting instead the conventionalized fictive interaction blend. While fictive interaction is compatible with Clarke’s construal of certain facts as being independently meaningful, it is less compatible with the mapping between facts and dots. Indeed, the creative nature of Clarke’s screaming-dots blend may be attributable to his maintenance of this mapping in the face of a larger (perceived) disanalogy. Innovation here emerges from the intersection of cultural, linguistic, and situational knowledge in the service of the interconnected demands of discourse coherence and argumentative goals.
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.22 (48)
Todd Oakley and Seana Coulson
References Barsalou, L and Wiemar-Hastings, K. 2005. “Situating abstract concepts.” In D. Pecher and R. Zwaan (Eds.), Grounding Cognition: The Role of Perception and Action in Memory, Language, and Thought, 129–163.New York: Cambridge University Press. Brandt, L. and Brandt, PA. 2005. “Making Sense of a Blend.” In Annual Review of Cognitive Linguistics. R. Mendoza Ibáñez &F. José (eds.), 216–249. Amsterdam: John Benjamins Press. Chafe, W. 1994. Discourse, Consciousness, and Time. Chicago: University of Chicago Press. Coulson, S. & Oakley, T. 2003. “Metonymy and Blending.” In K. Panther and L. Thornburg (eds.), Metonymy and Pragmatic Inferencing, 51–81. Amsterdam and Philadelphia: John Benjamins Press. Fauconnier, G. 1994. Mental Spaces: Aspects of Meaning Construction in Natural Language. Cambridge & New York: Cambridge University Press. Fauconnier, G. and Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Lakoff, G and Johnson, M. 1980. Metaphors We Live By. Chicago: University of Chicago Press. Langacker, R. W. 1999. Grammar and Conceptualization. Berlin: Mouton de Gruyter. Mandler, J. (2004). Foundations of Thought. Cambridge & New York: Cambridge University Press. Pascual, E. 2002. Imaginary Trialogues: Conceptual Blending and Fictive Interaction in Criminal Courts. Utrecht: LOT. Pickering, M. and Garrod, S. 2004. “Toward a Mechanistic Psychology of Dialogue.” Behavioral and Brain Sciences 27: 169–226. Safire, William. 2002. On Language: Blame Game. In New York Times. Late Edition, 22.
Appendix Transcript [Interview with Richard Clarke, former White House Advisor on Counterterrorism. Interview aired on NPR’s Fresh Air: Wednesday, September 22, 2004 program. Beginning 4:52 minutes and ending 7:10 minutes into the interview.] Available online at: http://www.npr.org/templates/story/story.php?storyId=3931123.
Transcript Conventions Underline: contrastive change in pitch or intensity (loudness) .. a brief pause ... longer pause , a continuative contour . a terminal contour ? a yes-no question contour –truncated or fragmentary unit = lengthening of a preceding vowel or consonant @ laughter « » accelerated speech » « decelerated speech
TSL[v.20020404] Prn:11/12/2007; 14:05
F: PB17001.tex / p.23 (49)
Chapter 1. Metaphoric language in discourse
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
Dave Davies: there’s a clearly a perceived need here for, . . . uh sensitive information to be shared, even among investigators among different agencies. uh, ..who are working, on.. issues that that «information about terrorism»might arise. in– in effect to connect the dots. ..and Richard Posner, in a piece in The New York Times recently ..um.. in reviewing the 9/11 Commission Report, criticized the idea that, as he put it, the failure to prevent the attacks was due to, ..a failure to collate the «bits of information possessed by different security agencies», especially the CIA and the FBI. ..um, he says, the best bits of information were not obtained until, ..the month or so before the attack, and he says, its unrealistic to suppose they could’ve been »inter=grated«, »and understoo=d« in time to detect the plot. ..um in effect he’s saying, «there’s all this information flowing around there», and it’s, ..w— we weren’t just pick this pick those dots out, and connect them in time. is he right? Richard Clarke: no I don’t think so. there’s uh some, .. uh dots, which are meaningless unless you put them together with lots of other dots. an=d, I understand what he’s sayi=ng. but there are some dots that come out screa=ming at you. uh do something now about me. uh,
TSL[v.20020404] Prn:11/12/2007; 14:05
Todd Oakley and Seana Coulson
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77
and.. when we, uh, knew..that there were.. Al Qaeda people uh, going to Malaysia, going to Kuala Lampur, ..uh and meeting there, and plotting there. ..we became very concern=ed ..uh because it looked like the »kind of« meeting, .. uh where Al Qaeda people, operational people, get together, . . .and go over the details, .. of some impen=ding attack. so if we had then learned, .. that some of those people in that meeting, were in Southern California, .. and ent@ered the United States, .. that would have been the kind of dot, uh that didn’t need a lot of connecting. that would have screamed out at you, do something about me now. and I think, y’know it’s, twenty-twenty hindsight and all that, but, .. I really do belie=ve, that if I were sitting in my old job in the White House, and I had see=n a report that said that, .. uh I would have ma=de the FBI, uh a– and frankly FBI headquarters would have wa=nted to, go=all out, to find those two guys.
F: PB17001.tex / p.24 (50)
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.1 (51)
chapter
The text and the story1 Levels of blending in fictional narratives Barbara Dancygier The paper investigates applications of the mental spaces approach to the study of narrative discourse. It is postulated that fiction narratives are structured as blends, emerging out of a number of major narrative constructs called narrative spaces, with their rich topology and specific structure. It is argued further that two mechanisms are responsible for emergent story structure: narrative anchors and cross-input projections. The former are narrative equivalents of space builders, while the latter account for the constant enrichment of narrative spaces as a result of emerging cross-space links. Finally, narrative viewpoint is defined as a feature of space topology, rather than of a specific ‘consciousness’. The paper argues further that blending opens new perspectives in narratological research.
Among discourse types, narrative discourse is perhaps the category which attracts the attention of the broadest array of disciplines. Fictional narratives, in particular, pose questions for linguists, psychologists, narratologists, literary critics, and stylisticians alike. The question of how a linguistic form, even of considerable complexity, can so fully absorb readers, transport them into fictional worlds, and affect them emotionally, is certainly a question which is difficult to answer from within one discipline alone. The growing interest in the cognitive approaches to various areas of the humanities brought the cognitive turn to narrative studies as well. The influences are coming from different directions, mainly cognitive psychology and cognitive science (cf. Herman 2003a), but also cognitive linguistics, in its various forms. The frameworks of cognitive grammar, mental spaces theory, and blending all afford new possibilities for representing the ways in which texts construct meaning, whether in poetry or in fictional prose (see Stockwell 2002; Semino & Culpeper 2002), but blending appears to be particularly well suited to the analysis of the . I want to thank David Herman, Sean McAlister, Tina Lupton, and Jennifer Schnepf for their helpful comments and great suggestions. I will go on learning from their advice in further work. Needless to say, the flaws of the present paper are all my own doing.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.2 (52)
Barbara Dancygier
narrative. Beginning with an early study done by Oakley (1998), through Semino’s work on mind style (2002) and Turner’s work on double-scope stories (2003), to recent analyses by Dancygier (2004a, b, 2005), Rohrer (2005), Schnepf (2006) and Semino (2006), blending is emerging as an important framework in the study of narrative fiction. This paper intends to open the discussion of how numerous levels of narrative structure are integrated to yield a coherent story. The approach to narrative structure which is proposed below relies on the concept of a narrative space, defined as a type of mental space, and outlines the most important characteristics of such spaces and their participation in conceptual integration networks.
Story versus text: The role of narrative sequence The specific goal of this paper is to propose a blending analysis of some of the processes which enable the reader of a fictional narrative to construct a complete sequential story out of an incomplete set of narrated events. As any discussion of fiction observes, events in the text may be presented non-sequentially, they may be distorted by the selection of one specific viewpoint, and there are usually gaps in the story line (to mention just a few such observations). All fictional narratives are thus fragmented, but all (or most) are read as coherent and complete. Some novels move the fragmentation to a still higher level and act on the pretense of telling more than one story. The particular choices vary, but what calls for an explanation is how such narrative structure is understood, and what integration mechanisms are responsible for the coherence and continuity of the underlying story, which is never told as such. I will suggest that blending provides the tools needed for explaining those mechanisms. Most fictional narratives rely on the basic narratological distinction between the text (and discourse or narration) and the story (Chatman 1978, 1990; RimmonKenan 1983; Toolan 2001). The story is essentially a sequence of events, as they occurred in the fictional world, leading from the beginnings of the situation described, through all the developments, to the conclusion. While narratologists differ in the weight they give to other aspects of the story, such as characterization or causal chains, the temporal chronology is generally agreed upon as the most definitive feature. 2 The text itself, however, typically deviates from the sequence, elaborates some parts of the story, while barely signaling other parts, shifts the viewpoint as the need arises, etc. In other words, various aspects of narrative . An extensive discussion of the story/text (or story/discourse/text, or syuzhet/fabula) distinction is beyond the scope of this paper. It is generally accepted in one form or another, although some narratologists, e.g. Fludernik (1996) question the need for maintaining it, while others (Herman 2002, 2003b) attempt to define it in more specific cognitive and linguistic terms.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.3 (53)
Chapter 2. The text and the story
structure make it possible for the reader to be engaged with the story on many levels, beyond simply grasping what happened and in what sequence. This common feature of fictional narratives has many functions: it may put more emphasis on the characters’ inner thoughts, create suspense, pose questions about a character’s motivation, etc.3 The pervasiveness of non-sequentiality in fictional texts may seem puzzling in view of the fact that lower level language constructions rely on the sequential interpretation to a high degree. It has often been noted, for example, that coordinate constructions are typically interpreted sequentially (and causally), even though other interpretations are equally plausible (cf., Haiman 1980; Sweetser 1990; Dancygier 1998). A sentence such as I checked my e-mail and had a cup of coffee could describe a situation where coffee-drinking preceded mail-checking, or where they were occurring simultaneously, but is typically understood in a way where the sequence of clauses iconically represents the sequence of events. The nonsequential modes of narration found in many fictional texts thus cut against the grain of what seems to be a general cognitive and communicative preference, and some explanation of fictional narrative techniques is thus required. Naturally, syntactic constructions can also express a non-sequential order, but this is typically the case when another order, not the order of events, is imposed (cf. Carston 1993). If the speaker wants to describe an event first and offer an explanation for the state of affairs later, the sequence Our team lost again – see, the goalie had injured his elbow is acceptable, and it is clear that the injury preceded/caused the defeat. But the “story” being told is now presented in two dimensions – the events are represented, but their occurrence is also explained via a mention of other, earlier events. It is also worth noting that oral narratives, as described by Labov (Labov & Waletzky 1967; Labov 1972), also contain elements which interrupt the sequential stretches of complicating action. These elements of the narrative, called evaluation, may bring in observations from outside of the story’s line of events. It appears, then, that even those narratives which are essentially faithful to the sequential presentation of events do require that the addressee have the ability to look at events as contributing to different levels of story construction and that different mental spaces are supported throughout, while reinforcing each other’s contribution to the whole. In fiction, some narrative spaces may play both roles – of filling the gaps in the event sequence, but, perhaps primarily, of offering a different view of the character’s motives. The reader of The Great Gatsby, for instance, is primarily engaged in following the events which are narrated by the eye witness, Nick Carraway, and which lead directly to Gatsby’s death, but the actions of the . See Herman (2002) for an interesting discussion of the issues of temporality in the narrative.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.4 (54)
Barbara Dancygier
characters which contribute to the ending would not be justified, or even comprehensible, without the knowledge of events which occurred outside of Nick’s narrative, in the past. It appears, then, that the function of explanation plays a role both in constructional sequencing of an utterance and in textual sequencing of a fictional narrative. In fiction, some narrative spaces may play both roles – of filling the gaps in the event sequence, and, perhaps primarily, of offering a different view of the characters’ motives. In other words, both oral and fictional narratives break the chronological sequence in order to highlight the relevance of background information or to allow specific story events to be seen from more than one perspective. The question, then, is whether the comprehension of a novel like Gatsby consists to any significant degree in reconstructing the full sequence of events, or whether the flashback is only a prompt leading to a different kind of understanding of the events being narrated, and thus has to be introduced into the story when such explanations are needed. If we assume that the latter is the case, the conclusion might be that the sequential story itself does not ensure the kind of comprehension intended. However, the text still requires that the links between the main story line and its flashbacks be provided by the reader herself. For example, the text of Gatsby does describe Gatsby’s past poverty and current fortune, but the reader has to provide the understanding of how the change in his situation relates to the events narrated by Nick. One way to describe the process of reading would then be to focus on the text as a prompt which triggers narrative comprehension.4 It does so by setting up mental spaces which require more elaboration and structure, as well as extended maintenance, and which participate in the process of story construction. I will call such spaces narrative spaces, and describe their specificity below. In this approach, the story is a narrative space constructed on the basis of the contributing narrative spaces, via the processes of conceptual integration. More accurately, I will refer to the narrative construct resulting from the processes of reading as the emergent story. While narrative spaces are set up in the text and, at least to a degree, receive their topology from various textual devices (lexical, syntactic, and stylistic choices), the emergent story is a blend of various elements from the spaces available, which relies on counterpart relations, selective projection and emergence of new structure. Most interestingly, it also contains aspects of narrative structure which are not available in any of the input spaces alone and arises as a result of subsequent blends. Due to the complexity of narrative input spaces, though, the . For a thorough discussion of the discourse aspects of narrative comprehension see Emmott (1997).
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.5 (55)
Chapter 2. The text and the story
integration processes which yield the emergent story rely not only on selective projection, elaboration, or completion, but also on narrative-specific mechanisms. Moreover, the construction of the emergent story may also affect the topology of the inputs and relies on the construction of narrative viewpoint in ways that go beyond profiling various types of narrators. At the same time, I argue that fictional narratives use meaning-construction mechanisms closely aligned with those used in processing other texts or communicative acts. In this respect, the approach advocated here is different from much of the work on narrative-structuring mechanisms done within the approach broadly conceived as Russian Formalism (cf. Propp 1968; Lotman 1979; and others), where the description of the specificity of literary discourse is the goal. Fictional narratives are here treated as textual tokens of conceptual, linguistic, and memory mechanisms also used elsewhere. In what follows, then, I will present some examples of the integration processes which govern the construction of the emergent story from its input spaces. More specifically, I will focus on the concepts of narrative anchors, the specificity of emergent cross-mappings, and the phenomena of viewpoint compression and embedding.
Narrative anchors, emergent cross-mappings, and cross-input projection Narrative spaces have been presented above as a variety of mental spaces.5 Indeed, they can be defined very similarly to mental spaces in colloquial discourse, as cognitive domains activated by the use of linguistic forms, while the purpose they serve is on-line story construction. Like other mental spaces, they can become inputs to narrative blends. While all mental spaces can be maintained and continue to be elaborated for the duration of the discourse interaction, narrative spaces can remain activated for much longer periods of time, and often with substantial interruptions (a reader may put away the text for periods of days or even weeks). Any speculation on the memory mechanisms which make the maintenance of fictional stories possible is beyond the scope of this paper, but the question is certainly worth investigating. In most of the existing literature, mental spaces are described as prompted by linguistic expressions (or their visual or gestural equivalents), called space builders. . The ensuing discussion relies in crucial ways on the concept of a mental space, and on the mechanisms of conceptual integration, or blending. A thorough introduction to these theoretical concepts is beyond the scope of this paper. The reader will find the necessary explanations in Fauconnier (1994[1985], 1997), Fauconnier and Sweetser (1996), Fauconnier and Turner (1996, 1998a, 1998b, 2002), Turner (1996), Turner and Fauconnier (1995, 1999), Coulson (2001).
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.6 (56)
Barbara Dancygier
For example, expressions such as yesterday, tomorrow, in Spain, if..., he thinks that... set up temporal, spatial, or hypothetical spaces which can then be elaborated. The concept of a mental space builder in the narrative is less clear. True, the traditional introduction to a fairy tale – once upon a time, in a country far, far away – seems to also transport the reader through time and space, and into places which are not understood to be the reader’s base space. So the basic similarity is there. However, contemporary fiction is typically far less explicit and the set-up is implied in the structure of the space rather than performed through a specific expression. Many contemporary narratives start in medias res, letting the reader construct the spatial or temporal viewpoint as the text develops. Also, characters are often immediately referred to through proper names, thus announcing their unique participation in the story. For example, McEwan’s Atonement starts as follows: (1) The play – for which Briony had designed the posters, programmes, and tickets, constructed the sales booth out of a folding screen tipped to its side, and lined the collection box in red crepe paper – was written by her in a two day tempest of composition, causing her to miss a breakfast and a lunch.
The reader is thrown into the narrative space where the main character resides, and with a remarkable level of descriptive detail. At this point nothing is known about Briony’s age and surroundings, also the content of the play is not mentioned. The mental space being elaborated suggests only that Briony is not a professional writer and that the performance will not take place in a theatre. What is important, though, is Briony’s obvious passion for this event and the fact that what will be happening might be in many ways a ‘one-man’ show. The space is thus set-up, and then developed, but it is being equipped with various elements of structure at the same time. The details of its topology will be added as the text progresses, but it is difficult to find a specific expression or expressions which could be described as space-builders. Still, the grammatical choices of person and tense do suggest a viewpoint, and profile an omniscient narrator. For comparison, the initial sentence of Atwood’s The Blind Assassin seems to play a different role: (2) Ten days after the war ended, my sister Laura drove a car off a bridge.
The sentence (with the support of the ensuing comments) introduces the reader to the narrative space with a number of very salient characteristics: – – – –
the story is being told by Laura’s sister (Iris); it is told in the past tense, from a temporal viewpoint closer to Iris’s ‘now’; it is significant that Laura’s death occurred so soon after the end of the war; and, perhaps most importantly the agentive verb drive suggests that Laura may have done it deliberately – in other words, she may have committed suicide.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.7 (57)
Chapter 2. The text and the story
It is hard to point to any specific expression as a space builder, but the sentence as a whole not only sets up a narrative space, but also allows the reader to construe the event described as ‘possible suicide’. If this is the conclusion, then the narrative also promises to help the reader find out whether Laura’s death was indeed a suicide or not. What has happened, then, is that in fact three different narrative spaces have been set up: the narrator’s space (with its teller – Laura’s sister – and its ‘present’ viewpoint), the ‘Laura’s death’ space (the focal event in the past), and the ‘suicide’ space (where the expectation is created that the narrative will either confirm the suicide interpretation or offer another one). Already in the ensuing fragment, Iris assures a policemen that the death was in fact an accident (another narrative space), while telling the reader that Laura “had her reasons”, and referring to the notebooks Laura left for Iris to read in a way that suggests that they contain the explanation of what happened (the ‘notebooks’ thus constitute an opening into another narrative space). The Blind Assassin is a very complex and multi-layered narrative, but the reasons for Laura’s suicide, the need for the official accident verdict, as well as the part of the story told in the notebooks, are all crucial to the reader’s comprehension of the complete story the novel tells. All the major input spaces which will continue being elaborated throughout the text are set up in the first thirty pages of the over-six-hundred-page text. While the story in Atonement proceeds more or less sequentially from the performance of Briony’s play (although it also surprises the reader in the end), the narrative of The Blind Assassin is like a floor puzzle, where different pieces are being added all the time, but the story emerging throughout the text revolves around the spaces set up at the beginning. I will refer to the narrative-structuring expressions (such as ‘drove’, ‘accident’, or ‘notebooks’) as narrative anchors. As I have argued elsewhere (Dancygier 2007), it is not predictable what kinds of expressions will take that role in the narrative. Some of them are indeed mental space builders in the ordinary sense of the word, by relying on temporal or spatial viewpoint, on representation spaces (as photographs or paintings), or the assumed relation between form and content (as the notebooks). But they can also rely on character traits or descriptive details requiring explanation, on perceived similarity or analogy, etc. One excellent example of a narrative anchor which is not easily identifiable as an ordinary space builder is ‘the green light’ in The Great Gatsby. The final scene of the first chapter talks about Nick looking at Gatsby standing alone on his lawn and wondering whether he should call to him. (3) But I didn’t call to him, for he gave a sudden intimation that he was content to be alone – he stretched out his arms toward the dark water in a curious way, and, far as I was from him, I could have sworn he was trembling. Involuntarily I glanced seaward – and distinguished nothing except a single green light, minute and faraway, that might have been the end of the dock.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.8 (58)
Barbara Dancygier
At this point in the novel the light is meaningless. Still, the description of Gatsby’s preoccupation with some object in the distance and the positioning of the scene at the end of the chapter give the ‘green light’ the focus of an anchor. As the text develops, the reader finds out that Gatsby is looking at the light at the end of Daisy’s dock. The mental space it represents turns out to be very rich – it is linked to Daisy and her present life with Tom, but it also embodies Gatsby’s dream of winning her back. Later Gatsby tells Daisy that he was looking at the light knowing it was her dock – it stood for all his desires and hopes, but also for her. At the end of the novel, when the reader knows that Gatsby didn’t realize his dream and died, Nick returns to the green light. Here is how the novel ends: (4) Gatsby believed in the green light, the orgiastic future that year by year recedes before us. It eluded us then, but that’s no matter – tomorrow we will run faster, stretch our arms farther. ...And one fine morning – So we beat on, boats against the current, borne back ceaselessly into the past.
The return to the green light here gives it a generic meaning – it now stands for anybody’s hope that the perfect future is just around the corner. The story of Gatsby’s life is thus re-framed as a generic story, and his failure is no longer the final failure, because the hope does not die, even though it cannot be fulfilled. But the innocent green light has meanwhile achieved its full potential of an anchor – it has come to represent the true theme of the novel and it structures the reader’s understanding of the events which have been narrated. To sum up, narrative anchors are ‘place holders’ for mental spaces which cannot be fully set up or elaborated at a given point in the narrative. An anchored space remains activated and is expected to be elaborated later in the narrative; furthermore, the expected elaboration is presented as contributing to the topology or framing of the currently active space. In a sense, the identification of anchors is necessary for the reader to focus on those aspects of the story which in the end contribute to a coherent understanding of what happened. While trying to recreate the temporal sequence may be one of the common strategies, in multi-space narratives the primary task the reader has to deal with is identifying the major narrative spaces involved and constructing the cross mappings among them, so that a coherent story emerges in the end. While the reliance on narrative anchors is clearly required in reading a complex text like The Blind Assassin, the seemingly less fragmented narratives like Gatsby rely on them too. Specific narrative space builders may thus be hard to identify, but there are still expressions which have a privileged role in the narrative, in that they prompt the reader to understand the ensuing text as contributing to the elaboration of the narrative spaces thus announced. In the simplest case, a flashback may appear in the text without any specific temporal space builder suggesting a specific past moment,
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.9 (59)
Chapter 2. The text and the story
but once it starts, the reader will gradually receive the information needed to place it within the story. It seems, then, that narrative spaces may not be clearly set up at the outset, and the text continues contributing to their topology. In the process of elaboration the text may also be leading the reader up a variety of narrative garden paths; for example, in The Blind Assassin, the readers are repeatedly prompted into believing one explanation or another, until the structure of all the contributing narrative spaces reaches a satisfying degree of coherence. While misleading clues lead the reader to construe the emergent story in ways which will then be abandoned in view of new evidence, the primary narrative spaces announced early on keep on structuring the processing of the new information the text provides. The analysis proposed here focuses on some textual devices which prompt for subsequent stages of story integration. Naturally, the ways in which different text fragments are linked in the process of reading are numerous and complex, and cannot be exhaustively discussed here. The concept of narrative anchors I introduce (see also Dancygier 2007) is primarily designed to capture the major divisions within the story and the gradual emergence of cross-mappings connecting ostensibly separate parts. For comparison, Emmott’s (1997) study of narrative contextual frames and the resultant explanations of the nature of anaphora in fictional narratives provides a useful model of some of the more fine-grained aspects of narrative comprehension. What narrative anchors are meant to capture, though, is the ways in which the local linguistic choices are co-opted to play a role in the global construction of the story. Narrative anchors participate in more than mental space set up and maintenance. They are also used to prompt for cross-mappings which later lead to the emergent blend. Without such links, a coherent emergent story could not be constructed. In most stories, the major cross-mappings are explicitly built by the text as the story develops. For example, in The Great Gatsby, the temporal connection between Gatsby’s ‘old’ self as James Gatz and his ‘new’ self as Jay Gatsby is explicitly introduced, though it is still up to the reader to see the relevance of the flashback to the main story line. It is possible, though, that the input narrative spaces are not presented to the reader as connected. In The Blind Assassin, for example, the main narrative space is anchored to the time when Iris, Laura’s sister mentioned above, remembers her youth. Embedded in this narrative is a novel by Laura, published after her death. In the novel there are just two characters, the secret lovers referred to as he and she. The reader has to recognize who he and she are in order to fully understand Iris’s memoir and the only way to do it is to dynamically construct cross-mappings between the lovers in Laura’s novel and the characters in Iris’s memoir. There are numerous instances of such cross-mappings emerging from the text, even though the identities of the lovers are not explicitly specified anywhere. For example,
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.10 (60)
Barbara Dancygier
‘Laura’s novel’ text presents the man as mentioning his being an orphan, and his Presbyterian background: (5) Anyway I’ve got no taste. It’s because I’m an orphan. The Presbyterians ruined me, in the orphanage. It’s why I’m so gloomy and dismal.
At the same time, there is a man in Iris’s story, about whom little is known, whose name is Alex Thomas. His background is mentioned several times in the ‘memoir’ narrative, first when Laura talks about it: (6) ‘He’s an orphan,’ said Laura. ‘He was adopted, from an orphanage. A Presbyterian minister and his wife adopted him.’ She seemed to have winkled this information out of Alex Thomas in a very short time.
and then, twelve pages later, when Alex talks about it briefly. A careful reader will construct the cross-mapping based on that, linking the lover in one story to the man in the other. There are many such emergent cross-mappings, which gradually let the reader construct the identity of the lovers. From the point of view of the emergent story, such cross-mappings display interesting characteristics. First, they are prompted by a recurrent descriptive feature of the text, and the textual prompts only achieve any significance at the second mention. The ‘orphan’ description does not matter for the ‘Laura’s novel’ space, although it can be hoped that the reader will notice it, given that nothing at all is known about him or her and the reader should be looking for clues to their identity. It is only when the description appears again that the cross-mapping can emerge. Characteristically, all the cases I have looked at so far follow the same pattern: a descriptive detail (setting, appearance, linguistic pattern, etc.) has no special role until it re-appears elsewhere, and then the second (or any subsequent) mention prompts the emergent cross-mapping. The emergent cross-mappings may be triggered by a variety of minor descriptive details – in a way, it is hard to expect any reader to notice all of the prompts, but some, at least, should be noticed. The cross-mappings do more than just provide new connections between or among the counterparts. In the case described above, identifying ‘Alex Thomas’ and ‘the lover’ as counterparts not only links the spaces in a new way, but also triggers a cross-input projection. As a result, Alex Thomas’s identity in one input space is projected onto the he-character in the other input space, while this, in turn, adds to the description of Alex Thomas by identifying him as ‘the secret lover’. This piece of information is crucial to the emergent story as a whole not only in terms of referential links, but also in terms of the plot: the things that Alex does in one space now explain the missing plot strands in the other. The cross-mapping thus contributes to the topology of both narrative spaces it links and prompts the blending processes leading to the construction of the emergent story. What is particularly interesting about this process is that the narrative inputs which are textually in-
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.11 (61)
Chapter 2. The text and the story
dependent are maintained as separate input spaces throughout the reading, also after the emergent structure has been fully grasped, because each of the contributing stories has its own narrative topology. Grasping the connection between the ‘lover’ in one input and Alex in the other enriches both spaces, and both will contribute to the final blend, but the focus and logic of each of the stories is also maintained. Similar cross-mappings are used to build the story in The Great Gatsby. The crucial event in the novel is a car accident in which Myrtle Wilson is killed. The accident is not narrated as such, and the understanding of ‘what happened’ depends crucially on a descriptive detail – the color of the car which killed Myrtle. There are two cars taking the party of friends to New York and then back – a blue one and a yellow one. She is killed by the yellow car, but the identity of the driver, who is thus responsible for the woman’s death, is not obvious and has to be pieced together from different narrative fragments. The emergent story where Gatsby (willingly) pays the price for the crime Daisy has committed has to be constructed by the reader through emergent cross mappings and cross input projection. The crucial connection is the one between the color of the car and the identity of the driver – Gatsby owns it, so he is the ‘natural’ suspect, and the reader has to recall some details of a conversation preceding the yellow car’s departure to know that Daisy drove it. One can further speculate that the abundance of very rich descriptions of color in The Great Gatsby primes the reader’s attention to the yellow/blue distinction, which may have not been mentioned in a different kind of story. Indeed, without the attention to color that Gatsby relies on the story would have been much harder to construct. The kinds of emergent cross-mappings (and the resulting compressions) that I have looked at so far are triggered by such vital relations as identity or analogy. As Schnepf (2006) points out, emergent cross-mappings, especially as regards identity, can also be triggered by more complex role-value mappings. One of the cases she discusses in The Blind Assassin adds another dimension to the character of Alex Thomas. In the ‘memoir’ space, which is the most elaborate input to the final blend, Alex is a very mysterious character. Among others, he is suspected of arson and he is wanted by the police for his alleged role in provoking rioting and looting. His actual role in these events (if any) is never explained in the memoir, nor does it emerge from his ‘lover’ identity in ‘Laura’s novel’. However, in another narrative space, he authors a pulp sci-fi story, with no clear connection to other narrative spaces. As Schnepf argues, the sci-fi story profiles text-specific roles, such as ‘a sacrificial virgin’, ‘the king’, or ‘the blind assassin’ (hired to kill the king and provoke riots and unrest). These descriptions are easily read as ‘roledescriptors’, because the characters in the sci-fi story are not given any names and are not narrated in any psychological depth. Schnepf shows in detail how the descriptive details of the ‘stock characters’ in the sci-fi story prompt cross-mappings
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.12 (62)
Barbara Dancygier
which give new meanings to the actions of characters in other narrative spaces. For example, once a cross mapping is established between Alex Thomas and ‘the blind assassin’, Alex can be understood to be actually guilty of the crimes he was suspected of, but it also becomes clear that he was paid to commit them by another character, whose involvement was not even suggested elsewhere in the text. Consequently, crucial events ‘read-into’ the emergent story are constructed based on the emergent role-value cross-mappings. The role-value analysis proposed by Schnepf goes much further than establishing anaphora (as discussed in Emmott 1997). The reason why the cross-space identities of characters are not given explicitly and have to be established in the process of reading is that some of the facts crucial for the understanding of the plot could not be told by the primary narrator – Iris. Within the story she tells she does not really know whether Alex is guilty of arson or not; in fact, being his lover, she is willing to believe his innocence. The sci-fi story Alex tells gives the reader (and Iris as well) an opportunity to construct the missing parts of the plot, and the cross-space identity mappings established by anchors serve as prompts, not as ultimate goals. These aspects of story construction are difficult to capture outside of blending (interestingly, the critical literature on The Blind Assassin either downplays the role of the sci-fi story or misses it altogether). Earlier attempts at acknowledging the reader’s contribution to the narrative as such (consider Iser’s 1974 idea of a ‘virtual’ dimension of the text) are theoretically not specific enough to account for the crucial role the readerly contribution to what is called the ‘story’. The mechanisms described above are (at least to a significant degree) responsible for the construction of the final emergent story. The emergent story arises through a gradually increasing network of narrative anchors and emergent cross-mappings. The projections that result from the ever-growing number of connections add narrative structure to all of the spaces linked, until a coherent and complete story is constructed.
Levels of mental-space embeddings in the narrative As the examples above suggest, fictional narratives are blends of their participating narrative spaces. It is important, however, to note that many levels of narrative structure can be looked at this way.6 Among others, the phenomena of represented speech and thought, which have been broadly discussed in stylistics, narratology and linguistics (e.g., Leech & Short 1981; Banfield 1982; Genette 1980; Fludernik . See Oakley’s (1998) analysis of the graphic novel Maus, which describes the space construction processes in a text where multiple narrative levels are involved.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.13 (63)
Chapter 2. The text and the story
1993; Cutrer 1994; Sanders & Redeker 1996; Semino & Short 2004; Vandelanotte 2004a, 2004b, 2005), can be viewed in terms of multiple mental spaces and mental space embeddings (as described in Sweetser 1996; Dancygier & Sweetser 2005). Let us consider the second paragraph of The Blind Assassin: (7) I was informed of the accident by a policeman: the car was mine, and they’d traced the license. His tone was respectful: no doubt he recognized Richard’s name. He said the tires may have caught on streetcar track or the brakes may have failed, but he also felt bound to inform me that two witnesses – a retired lawyer and a bank teller, dependable people – had claimed to have seen the whole thing. They’d said Laura had turned the car sharply and deliberately, and had plunged off the bridge with no more fuss than stepping off a curb. They’d noticed her hands on the wheel because of the white gloves she’d been wearing.
The fragment features a number of mental spaces – Iris’s narrative, the policeman’s explanations, Iris’s speculation on why he sounded perhaps overly respectful, the reports of the eye witnesses. The main narrative space here is Iris’s narrative, with her thoughts and the policeman’s hypothetical speculations embedded in it, and then with the witnesses’ words subsequently embedded in the policeman’s report. The third paragraph then picks up on Iris’s narrative: “It wasn’t the brakes, I thought. She had her reasons”, thus adding another level of Iris’s reflections on the accident. The embeddings in the paragraph can be described as most of the major categories of represented speech and thought, introducing different voices and viewpoints and thus giving depth to the narrated event. They also display some of the basic features of mental space embedding, by shifting the personal pronouns and tense forms to mark the inherited viewpoint. For example, the policeman’s thoughts and words are represented in the third person, and his report of the witnesses’ accounts is introduced through a past perfect form (They’d said), to mark the double-past viewpoint (Iris is remembering a past conversation with the policeman, while the policeman is reporting a conversation that took place before he met with Iris). All the mental spaces activated here (Iris’s narrative, the circumstances of the accident/suicide, conversation with the policeman, policeman’s thoughts, conversation with witnesses, the witnesses’ representation of the accident, etc.) have their own structure and topology, their own viewpoint, etc. However, these contribute to the topology, at the macro-narrative level, of the space which will continue to be elaborated throughout the text, the ‘suicide’ space. It appears, then, that the story construction progresses through a number of levels of mental space embedding and blending. Between the micro-level of the circumstances of Laura’s death and the immediate reaction of the witnesses and the macro-level of constructing an explanation of her reasons for killing herself
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.14 (64)
Barbara Dancygier
there are, of course, scenes and sub-plots which also have a structure of mental spaces – for example, the story of Laura’s childhood or the scene of her conversation with Iris which immediately precedes her death. This seems to suggest that a mental space analysis would be possible and useful in an account of all of the stylistic specificities of narrative prose. Also, while a lot has been said about the micro-level of represented speech and thought, as well as the accompanying pure narration, the higher levels are often too elusive or too complex to warrant a disciplined discussion. However, the macro-level of the narrative – the construction of the emergent story – is apparently governed by principles which are only partially similar to the sentential level. Most visibly, the sentential level relies to a large degree on the grammatical features related to time, viewpoint, focus, or anaphoric reference across spaces. The markers of space embedding (mainly verb forms), as well as the space builders, are more easily identifiable. For example, in the paragraph quoted above, the use of tense, modals and aspect distinguishes clearly between speculation and report, while also marking the viewpoint spaces (e.g. Iris’s report, the policeman’s report, the witnesses’ report) and the focus spaces (what all the various people thought or said). Also, the use of definite and indefinite articles (the car, a retired lawyer) suggests unambiguously which referents are maintained throughout the fragment and which are being temporarily introduced as the report develops. The verbs such as inform, say, claim, or notice are space builders. An analysis relying on these tools can probably do justice to the meaning of the fragment in terms of mental-space structure and it specific mode of representing speech and thought. The narrative macro-level, however, cannot be easily subjected to such a finegrained analysis. The formal exponents of narrative space embedding are not systematic and cannot be viewed in terms of grammatical forms, there are no space builders which reliably operate on that level, and there are many more levels of viewpoint and focus. Perhaps more crucially still, the relationships of embedding are not necessarily built through textual clues alone. As I suggested in the discussion above, the relationships among spaces at the higher levels of the narrative have to be constructed by the reader in the process of arriving at the emergent story, while part of the construction is prompted by emergent cross-mappings and narrative anchors. To maintain a terminological difference between the micro- and macro-levels of narrative structure, I will reserve the term narrative space for the mental spaces which organize the topmost level of a fictional narrative and contribute directly to the emergent story. In view of this hierarchical understanding of the narrative, narrative anchors should also be seen as textual devices appearing at the sentential level which participate in the narrative construction of the emergent level. Macro-level embeddings have often been the subject of narratological interest, starting with the model proposed by Genette (1980) where narratives can be
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.15 (65)
Chapter 2. The text and the story
talked about as framed by higher level narratives (that is, hypodiegetic, hypohypodiegetic, etc.). Recent work by Herman (2006) discusses the role of framed narratives and the relations among individual levels in terms of distributed cognition, thus adding a cognitive dimension to the inquiry. What we should note, however, is that most of the research in this area assumes some default or primary narrative level (the diegetic level), where the story as such is told, so that the subsequent levels of embedding are seen with respect to the central one (either as containing it or as contained in it). Also, the discussion of narrative macrolevels (though they are not termed as such) is seen as independent of the speech and thought representation phenomena (or the micro-level). There are numerous questions which such an analysis cannot address, such as the problem of the interaction between the levels, the emergence of the higher level based on narration at the lower level, or horizontal connections across different parts. The approach proposed above, where narrative construction is less hierarchically organized, seems to address some of the issues narratological analysis cannot easily explain. The interaction among levels (whether horizontally or vertically related) can be naturally represented via cross-space mappings, embeddings, and the attendant inheritance phenomena. The role of micro-level narration in the construction of narrative spaces can be naturally accounted for – as in the case of the initial paragraphs of The Blind Assassin, where the major narrative spaces are set up on the micro level first. Finally, there is no need to assume the central (diegetic) role of any of the levels, since the story as such does not need to be located in a distinguishable part of the narrative, but is constructed based on the cross-mappings and blends projecting structure from all of the levels. What is more, the issues concerning further aspects of narration, such as the narrator, the narratee, the role of framing narratives, etc., will naturally result from the analysis of the participating narrative spaces and the links connecting them.
Viewpoint compression and blending So far, we have looked at fiction narratives from the point of view of their organization and the emergence of the sequential story. However, a narrative is also expected to profile a narrator (or narrators) – some ‘consciousness’ (or ‘consciousnesses’) which is (are) responsible for telling the story in the specific way. The narrator may be constructed as existing outside of the story itself (omniscient narrator) or the role can be assigned to a participant in the story’s events. There are of course many fine differences among texts in how the concept of the narrator is instantiated (see Booth 1961; Chatman 1978; Bal 1985; Simpson 1993) and there is a large body of research on how these choices affect the narrative itself. In what
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.16 (66)
Barbara Dancygier
follows, however, I would like to focus on the role that the construct we call ‘the narrator’ plays in the emergence of the final story. In earlier work (Dancygier 2004b, 2005) I argued that the position of the narrator (or, as I would prefer, the story’s teller) is co-aligned with the viewpoint of the narrative space it is profiled in. Furthermore, the micro-level shifts of viewpoint do not result in confusion or lack of coherence, because of what I termed viewpoint compression. One of the examples, from J. Raban’s Old Glory, is repeated here: (8) The TV news went local. An Englishman had left Minneapolis that day in a small motor boat [...]. In the picture on the screen his face had a cheesy pallor. [...]. He looked to me like a clowning greenhorn [...]
The teller of this travel narrative is taking a boat trip down the Mississippi. The day after starting out, he watches the local news and sees himself being interviewed just before departure. The whole story is told from his point of view as the ‘traveler’, but his account of what he saw on TV is narrated as if there were no connection between the traveler he is and the one on the news. In the mental space set-up of the fragment the traveler-narrator adds a ‘TV viewer’ space to the main ‘story of the trip’ space, while decompressing his own persona between the two spaces. The ‘TV viewer’ persona can now look at the ‘traveler’ persona with some objectivity. For the duration of the discussion of the TV show, the narrative viewpoint is identified with that of the ‘viewer’, while the ‘traveler’ viewpoint temporarily disappears from the story. The decompression of the narrator’s identity thus results in two different viewpoints being set up, but then one of them is selected to take over the narrative viewpoint for the duration of the blend’s operation in the text – this is what I refer to as viewpoint compression. The reader is thus given two different viewpoints, but the narrative space and the selected decompressed space have their viewpoints compressed; this allows a temporary shift of viewpoint without the loss of continuity in the narrative. Examples like (8) may create a misleading impression that a viewpoint shift (in this case, from participant to viewer) is restricted to aspects of perception. However, even in an example this simple the viewpoint involves much more – aesthetic evaluation (cheesy pallor), assessment of personality traits (clowning, greenhorn), personal information on nationality (an Englishman), but also an expectation of what qualities would be expected from someone undertaking a dangerous voyage. All these viewpoint components are inherently part of the ‘vantage point’ taken. What is more, the mostly negative picture emerging is also indicative of all the parallel viewpoint components in the ‘participant’ space – when being filmed that day the traveler probably did not see himself as ‘a clowning greenhorn’, so that the viewpoint implied (and then enriched via compression) is that of someone undertaking a dangerous trip with confidence and expertise. The understanding of viewpoint emerging here, and more directly applicable to narrative discourse, is that
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.17 (67)
Chapter 2. The text and the story
viewpoint is a characteristic of the space, and thus depends on the space topology. The discussion of the relations between viewpoint and focus spaces (cf. Fauconnier 1997; Cutrer 1994), has concentrated on specific phenomena, such as tense and aspect, where not much topology has to be explicitly mentioned. However, applying mental space analysis to conditionals (cf. Dancygier & Sweetser 2005) shows clearly that the topology needed to explain conditional reasonings involves a number of complex viewpoint elements (linguistic and social norms, epistemic states, inferential chains, etc.). Recent work by Verhagen (2005) and Pascual (2006) also shows how mental space topology is used in intersubjective construals and in the evocation of independent discourse patterns, thus further adding to the variety of viewpoint topology the spaces use in interaction. When narrative spaces are taken into consideration, it becomes clear that a space which is the designated viewpoint space at any point in the narrative will have to be richly structured, profiling information about space, time, epistemic states, inferential patterns, subjectivities, etc. In fact, the expanded understanding of viewpoint which narrative analysis requires can fruitfully be used to explain specific aspects of narration choices. For example, an omniscient narrator in a third person narrative does not need to be thought about as a ‘consciousness’. On the contrary, one could argue that we (as readers, at least) tend to construe such narrators as ‘consciousnesses’, because the epistemic and inferential richness of the viewpoint space thus set up naturally calls for such a construal. But the richness is the feature of a narrative space, not of a specific human-like mind. As I argue below, the concept of a ‘teller’ is in fact less crucial to the narrative than the concept of narrative viewpoint, and can be subsumed by it. The paragraph from The Blind Assassin quoted in (7) is naturally interpretable in terms of viewpoint compression, on micro- and macro-level. For example, there is clearly a difference between the possible ‘accident’ viewpoint, represented by the policeman’s speculations, and the ‘suicide’ viewpoint, emerging from the witnesses accounts embedded in his description of what happened. These viewpoints will be maintained throughout the novel, although they will be represented by different participants in the story, but at the micro-level of the paragraph the policeman’s account given to Iris compresses them into the ‘police report’ viewpoint, which does not seem to favor either interpretation. This is further compressed with Iris’s viewpoint, since she is presented as the focal point of the whole story, in whose hands all the viewpoints converge and are given their proper role. In the text that follows, the double role of ‘teller’ and ‘participant’ that she represents at the start is developed further: Iris believes it was suicide, and this will be her viewpoint as the teller of Laura’s story, but she will have to maintain the official ‘accident’ verdict as a participant in the events. The compressions thus build up all the way to the highest levels of the narrative. There is a possibility, then, to see particular tellers, at different levels of narrative structure, as having no narrative role other than introducing or maintaining
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.18 (68)
Barbara Dancygier
the viewpoints which structure the story. This seems to be the case in the quoted fragment from The Blind Assassin, and also in the further presentation of the story. Immediately following Iris’s brief introduction, which suggests suicide, there is a newspaper clipping, which gives an official version of the events. The clipping makes it clear that the family insisted on the accident verdict. The reader is thus offered two viewpoints – of Iris, who knows what happened, and of the official version maintained by other members of the family. Even though they are not equally prominent in the remainder of the novel, they do both contribute to the readers’ understanding of ‘what happened’, and any indication of one version of events being true immediately reflects on the perception of the other version of events and the motives behind it. The compressed viewpoint established early on is crucial to the construction of the story as a whole, especially to the understanding of the roles of individual participants. What these examples suggest is that the position of a teller is typically profiled to introduce and maintain a viewpoint, but the same viewpoint may be represented by more than one teller, and a teller may be construed in terms of more than one viewpoint. In what follows, I will explore the idea that the construal of a narrator is not necessary for the emergence of a specific narrative viewpoint. I will consider two macro-level examples where narrative viewpoint is to a large degree independent of the profiling of a narrator, or in fact stimulates a choice of a teller to match the viewpoint. The first case I will consider is Jan Potocki’s The Manuscript Found in Saragossa – a story of interweaving stories (over a hundred of them altogether), written by a Polish aristocrat, in French, over the period between 1797 and 1815. Apart from its complex structure and astounding cultural collage, which have attracted critics and a variety of admirers, the text has a great number of narrators: from the main character/narrator linking the stories, named Alphonse van Worden, through a number of passing storytellers along the way. It is not surprising, then, that the position of a teller is different in a text of this kind, but it does not follow that he or she does not matter at all. The story told in the Saragossa Manuscript describes, at least at some level, Alphonse’s journey to Madrid through the Sierra Morena mountains. Early in the journey, he decides to spend the night at an inn, abandoned because of its reputation of being haunted. As night falls, he is led to a secret chamber, where he meets two beautiful sisters, Emina and Zubeida. They turn out to be Moorish women, of noble birth, and they appear to want to seduce him. However, when the night ends, Alphonse finds himself awakening at the foot of the gallows, in the company of two decomposing bodies. The cycle of enjoying the company of the sisters and ending up under the gallows repeats itself, and during one of the amorous encounters Alphonse promises to never mention the sisters to anyone.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.19 (69)
Chapter 2. The text and the story
Main narrative space: the manuscript Teller: Alphonse van Worden Alphonse’s story Teller: Alphonse Father’s story Teller: Alphonse Landulpho’s story Teller: Father
The inquisition story Teller:Alphonse Pacheco’s story Teller: Pacheco Trivulzio’s story Teller: Theologian
Figure 1.
The story is thus strongly suggesting that Alphonse is in fact playing with devils, or demons of the two hanged men, tempting him into sin. When the idea is established in the text, Alphonse starts to tell his own story (but is constantly interrupted by new events in his journey). The resulting set of stories is diagramed in Figure 1. Each of the stories in Figure 1 has a different teller – from the main narrator, who tells the past story of his youth, as well as reports being surprisingly captured by the inquisition, through Pacheco, a crippled and emotionally shattered man he meets, to his father or a theologian sharing their household. Although the tellers are clearly connected to the time and space of the stories they inhabit, which are distant from the space of Alphonse’s journey, what they say does put the main narrative space in a new light. Alphonse’s father wanted him to grow up to be an honorable man – never show fear and never break his word. The stories of Landulpho and Trevulzio are both stories of valiant men who are tormented by demons in the shape of beautiful women. Both were told to a very young Alphonse, to teach him that if he were to find himself in a similar predicament, he should never show fear. After the first story, young Alphonse is frightened, after the second one, he learns his lesson and behaves bravely. Although Alphonse does not make that connection for the reader, and there are many other stories told in the same narrative strand, it is clear that the recurring theme of women-demons links the young man’s memories with his recent experience with Emina and Zubeida (via emergent cross-mappings). However, the stories also establish their own shared viewpoint – that encountering demons may cause fear, but an honorable man cannot yield to it and has to keep his word.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.20 (70)
Barbara Dancygier
Main narrative space: the manuscript Teller: Alphonse van Worden Demons can take the shape of women. Be honorable. You cannot show fear or break your word.
Be faithful to the teachings of your church. Confess your sin.
Figure 2.
Thus, regardless of the actual tellers and protagonists, the stories jointly set up a viewpoint space that Alphonse is familiar with – the space of being brave above all other things. In the second strand, the religious authority of the inquisition and of the hermit who takes care of insane Pacheco is developing a different viewpoint. Pacheco lost his eye and his sanity in an encounter with two beautiful women, who appeared to be demons (again!). But he made a confession and is on his way to recovery – presumably, if Alphonse confessed his sin to the church, he would also be saved. As a result, the two strands of stories seem to be constructing the viewpoint spaces between which Alphonse will have to choose (see Figure 2). Both strands of stories are suggesting that Emina and Zubeida are devils, and Alphonse has reasons to believe that this might be true. But they are also specifying the options available to him, both of which agree with his own convictions and upbringing. He has to choose the more important value, and chooses honor over faith. At the end of the story his choice is handsomely rewarded. It seems, then, that in a text like the Saragossa Manuscript the presence of the many tellers does not necessarily evoke a proportional array of viewpoints. Also, the viewpoints are not constructed with respect to the individual spaces, but emerge as shared generic spaces, linking various stories and giving them relevance to the narrative as a whole. None of the stories or tellers alone can flesh out Alphonse’s moral dilemma, but they can, at best, represent one of the viewpoints. Still, it is the cross-space reinforcement of the repeated moral message that helps construct Alphonse’s story as a whole. This is, then, a case where the viewpoint is constructed, but the tellers are only marginally important to the process. In fact, other narrative mechanisms – such as the emergent cross-mappings, seem to play
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.21 (71)
Chapter 2. The text and the story
suicide
v-point: Iris
Iris remembers v-point: Iris
accident
v-point: official
Blind Assassin 1 (Laura’s novel) notebooks v-point: Laura
v-point: Iris/Laura
Blind Assassin 2 (sci-fi story) v-point: Alex
Emergent story Figure 3.
the key role in viewpoint construction. Also, one could argue that the emergence of viewpoint in the strands of the story discussed above also relies on viewpoint compression – several narrative spaces thus receive a shared level of structure, while maintaining their independence as coherent narratives. An even more interesting case of the relationship between narrative viewpoint and the profiling of a teller is presented by The Blind Assassin. The novel, as described above, consists of several, ostensibly independent, narratives, which prompt the reader to construct a unifying, over-arching story. All in all, the major narrative spaces of the novel could be presented as in Figure 3. In the diagram, the spaces are described in terms of their narrative viewpoint, although it is not necessarily the case that in each space the narrative viewpoint is unambiguously associated with a person-like teller. For example, the ‘accident’ space is identified mainly through a number of newspaper clippings spread throughout the text, which serve as a kind of running commentary on what Iris’s story reveals. It is, then, another example of a viewpoint not attributed to a specific teller, given that the assumed reporter who wrote the words is not in any way related to the story itself. The most interesting space in the network is the novel ‘The Blind Assassin’ (numbered 1), introduced early in the text as written by Laura and published by Iris after Laura’s death. The novel is written as the present tense, third person narration, with the narrator referring to the two (there are only two) characters as he and she. The reader is soon led to believe (through emergent cross-mappings and anchors, see Dancygier 2007) that the lovers in the story have counterparts in the main narrative – in Iris’s memoir. As I was showing above, the reader soon finds out that the man in the story is Alex Thomas, but the identity of the woman is more mysterious. Still, it is also becoming clear that the story of the secret lovers
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.22 (72)
Barbara Dancygier
Iris remembers Teller: Iris Characters: Iris Chase Laura Chase Alex Thomas
Novel by Laura Chase The Blind Assassin (1) Teller:
??Laura
Characters: she he
Laura is Alex’s lover Figure 4.
in Laura’s novel represents the viewpoint anchored to the woman. In other words, the author/narrator and she are one and the same, in spite of the third person style. Knowing that the published author is Laura, the reader concludes that Laura was Alex’s lover, and Iris’s narrative offers numerous hints supporting that view. The structure can be represented as in Figure 4. The diagram represents the set-up in which a character in the ‘Iris’ narrative space is ostensibly profiled as the teller of the ‘novel’ space. As regards the characters, the identity of the man is soon discovered via emergent cross-mappings, while the identity of the woman is selected to match the viewpoint represented in the space. Initially, based on the available information about authorship and on the coherence across the two spaces, the woman (and also the teller) is identified as Laura. An additional contribution to the emergent story is thus the information that Laura was Alex’s lover. However, the text of the ‘novel’ and of Iris’s memoir then keep on suggesting that the woman in the story of the lovers is in fact Iris. As the story develops, it becomes clear that the viewpoint it represents is more accurately attributed to Iris, not to Laura. In other words, the identity of the teller has to match the viewpoint (temporal, spatial, epistemic, etc.), not the other way around. At some point in the reading (which might vary from reader to reader), the relationship between the central narrative spaces has to be adjusted. In its final form, it is represented in Figure 5. Iris is now identified as the woman in the ‘novel’ space – she is the only person who could know what ‘she’ knows. This makes the viewpoint of the ‘novel’ space Iris’s viewpoint, and designates Iris as the teller – the actual author of the novel published under Laura’s name. In terms of the emergent story, it also prompts the conclusion that Iris, not Laura, was Alex’s lover.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.23 (73)
Chapter 2. The text and the story
Iris remembers Teller: Iris Characters: Iris Chase Laura Chase Alex Thomas
Novel by Laura Chase The Blind Assassin (1) Teller:
Iris
Characters: she he
Iris is Alex’s lover Figure 5.
Another consequence is that Iris’s viewpoint is now represented in both major narrative spaces of the text. Since the spaces constitute the core of the emergent story, the whole narrative is thus subordinated to Iris’s viewpoint. The viewpoints of the individual stories, in turn, are thus compressed, so that Iris’s memoir now represents both what her narrative says, and what the ‘novel’ narrative adds to it, within the scope of one coherent narrative. What is remarkable, however, is that a text of this kind may profile more than one possible teller, and the choice is made based on the viewpoint characteristics of the narrative space. To put it in somewhat simplified terms, narrative viewpoint constructs the teller, not the other way round. It is the space topology that decides how viewpoint will be construed, not the profiling of a specific narrator or ‘consciousness’. It should also be noticed that the two central spaces of The Blind Assassin are narrated in different ways. One of them profiles a narrator who is also a character, the other profiles a narrator outside of the space; one is told in the past tense, the other in the present tense; one relies heavily on referential expressions such as proper names, definite descriptions, etc., the other uses only third person pronouns. And yet they do tell the same story, thanks to the power of narrative anchors and emergent cross-mappings. The observation seems to suggest that the construction of narrative viewpoint relies on a number of different kinds of viewpoint – temporal and spatial, first of all, but also epistemic (what does the teller know) and evaluative (what does the teller think).7 In many narratives, all the aspects of viewpoint converge around the construction of a teller, but in those stories . Epistemic and evaluative aspects of mental space construction have been discussed in Dancygier (1998), Dancygier and Sweetser (2005), Harding (2003, 2004).
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.24 (74)
Barbara Dancygier
which are not consistent and unambiguous in their construal of a teller, various aspects of the viewpoint marking a given narrative space may be foregrounded instead. As the texts briefly analyzed above suggest, the teller is a narrative space construct, crucially dependent on other aspects of the space topology, and related to, but not identical with the viewpoint the space represents.
Conclusion In the sections above, I have outlined some of the ways in which mental space theory can be fruitfully applied to narrative discourse. First, the analysis shows clearly that, contrary to common belief, mental spaces and blending framework is useful in the consideration of meaning construction phenomena reaching beyond the scope of single expressions or utterances. In fact, it offers an elegant and useful approach to much more complex discourse types, such as fictional narratives. In particular, the mental space approach gives us an opportunity to use one coherent framework to look at the construction of the text as such (including the context in which it was produced), and, simultaneously, to represent the processes of reading comprehension, including the cases of varying degrees or pace of comprehension. We can now discuss the text of a narrative as providing the prompts (lexical, syntactic, stylistic, or pragmatic) for the construction of narrative meaning – the emergent story. Furthermore, a fractured text (such as The Blind Assassin) may be coherently described in terms of the function of different prompts provided, as only some prompts are directly relevant to the emergent story right away, while others may be meant to mislead or confuse the reader. This interaction between the text and the reader is part of what the enjoyment of reading often relies on – if fictional narratives did not contain any element of mystery or suspense, they would be a lot less appealing. At the same time, analysts can now compare different texts in terms of the kinds of prompts they provide and the kinds of reading skills (or grounding assumptions) they require. Consequently, the issue of narrative coherence can be explored at the level of the text as a whole. At the beginning, I mentioned the need to flesh out the relationship between the text as such and the story that the reader gets as a result of interacting with the text. I have argued further on that what has been referred to as the story is more or less equivalent to the emergent story. However, the term story has also been used to talk about film adaptations – in a sense, the same story can underlie the text of a novel and a film adaptation. The nature of the text/film relationship is a highly controversial issue, given that film adaptations often significantly change important aspect of the story (characters, plot, setting), sometimes beyond recognition, and that some texts have been filmed more than once. It has been argued that a
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.25 (75)
Chapter 2. The text and the story
movie and a book can tell the same story,8 but those who read the book typically do not believe that the movie represents it accurately, or even well (one of my favorite quotes warns: “Never judge a book by its movie”). In fact, trying to be specific about a narrative structure which is identical for a book and its film adaptation does not seem to be a fruitful line of investigation. Anecdotal cases of bad or good adaptations aside, the mental spaces approach to the narrative could suggest an explanation of what the inevitable changes do in terms of the narrative space set-up, how the textual prompts differ from visual prompts, or, perhaps most importantly, why some aspects of the text cannot be ‘translated’ into a movie. It is not necessarily the case that the emergent story of a novel will be the same as the emergent story of its adaptation, and, quite interestingly, movie viewers familiar with the book usually do not compensate for missing story structure by recalling the text. The movie remains an independent ‘text’, with its own emergent characteristics. In fact, the visual mode makes certain kinds of narration impossible – for example, it is hard to imagine how a movie could keep the viewer hesitating whether Iris or Laura is Alex’s lover; the face would reveal the answer right away, and it would be awkward to have an actress hiding her face through most of the story. Still, the mental space approach seems to promise an interesting way of specifying the differences and the similarities. At the same time, investigating narrative discourse should prompt the mental spaces/blending framework into a more complete representation of the ‘on-line’ aspect of the ‘on-line meaning construction’. Narratives provide a particularly complex example in this respect, but colloquial discourse relies to a large degree on some of the similar phenomena: maintaining the activation of a space already set-up, elaborating many spaces at the same time, blending the spaces into a higher level structure which gives coherence to the discourse event, constant readjustment in the input spaces and adding new counterpart connections, etc. It should appear possible, in the long run, to view narrative discourse both as different from spoken interaction, and as relying on some of the same mental space mechanisms.
References Bal, M. 1985. Narratology: Introduction to the Theory of Narrative. (Translated by Christine van Boheemen). Toronto: University of Toronto Press. Banfield, A. 1982. Unspeakable Sentences: Narration and Representation in the Language of Fiction. Boston: Routledge and Keegan Paul.
. For an extensive discussion of film narratives, see Chatman (1978, 1990). Some comments on blending and mental space structure in film narratives can also be found in Turner (2006), Dancygier (2006), and Rohrer (2005).
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.26 (76)
Barbara Dancygier
Booth, W. 1961. The Rhetoric of Fiction. Chicago: Chicago University Press. Carston, R. 1993. ‘Conjunction, explanation and relevance.’ Lingua 90:27–48. Chatman, S. 1978. Story and Discourse: Narrative Structure in Fiction and Film. Ithaca: Cornell University Press. Chatman, S. 1990. Coming to Terms: the Rhetoric of Narrative Fiction and Film. Ithaca: Cornell University Press. Coulson, S. 2001. Semantic Leaps: Frame-shifting and Conceptual Blending in Meaning Construction. New York and Cambridge: Cambridge University Press. Cutrer, M. 1994. ‘Time and Tense in Narratives and Everyday Language’, Ph.D. Dissertation, University of California, San Diego. Dancygier, B. 1998. Conditionals and Prediction. Cambridge: Cambridge University Press. Dancygier, B. 2004a. ‘Identity and Perspective: The Jekyll-and-Hyde Effect in Narrative Discourse’, in M. Achard and S. Kemmer (eds) Language, Culture, and Mind. Stanford: CSLI Publications. Dancygier, B. 2004b. ‘Visual Viewpoint, Narrative Viewpoint, and Mental Spaces in Narrative Discourse’, in A. Soares da Silva, A. Torres, and M. Gonçalves (eds) Linguagem, Cultura e Cogniçao: Estudos de Linguistica Cognitiva, vol. 1/2 Dancygier, B. 2005. ‘Blending and narrative viewpoint: Jonathan Raban’s travels through mental spaces.’ Language and Literature 14.2: 99–127. Dancygier, B. 2006. ‘Preface: What Can Blending Do for You?’ Language and Literature 15.1:5– 15. (Special Issue on Blending; Guest Editor: B. Dancygier) Dancygier, B. (2007) ‘Narrative Anchors and the Processes of Story Construction: The case of Margaret Atwood’s The Blind Assassin’. In Style 41.2: 133–152. Dancygier, B. and Sweetser, E. 2005. Mental Spaces in Grammar: Conditional Constructions. Cambridge: Cambridge University press. Emmott, C. 1997. Narrative Comprehension: A Discourse Perspective. Oxford: Clarendon Press. Fauconnier, G. 1994 [1985]. Mental Spaces. 2nd ed. Cambridge: Cambridge University Press. Fauconnier, G. 1997. Mappings in Thought and Language. Cambridge: Cambridge University Press. Fauconnier, G. and Sweetser, E. (eds) 1996. Spaces, Worlds, and Grammars. Chicago: University of Chicago Press. Fauconnier, G. and Turner, M. 1996. ‘Blending as a Central Process of Grammar’, in A. Goldberg (ed) Conceptual Structure, Discourse, and Language, pp. 113–130. Stanford, California: CSLI Publications. Fauconnier, G. and Turner, M. 1998a. ‘Conceptual Integration Networks’, Cognitive Science 22–2: 133–187. Fauconnier, G. and Turner, M. 1998b. ‘Principles of Conceptual Integration’, in J. P. Koenig et al. (eds.) Discourse and Cognition, pp. 269–283. Stanford, California: CSLI Publications. Fauconnier, G. and Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Fludernik, M. 1993. The Fictions of Language and the Languages of Fiction. London: Routledge. Fludernik, M. 1996. Towards a ‘Natural’ Narratology. London and New York: Routledge. Genette, G. 1980. Narrative Discourse: An Essay in Method. (Translated by J.E. Lewin) Ithaca: Cornell University Press. Haiman, J. 1980. ‘The Iconicity of Grammar: Isomorphism and Motivation.’ Language 56:515– 40.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.27 (77)
Chapter 2. The text and the story
Harding, J. Riddle 2003. ‘Evaluative Stance, Counterfactuals, and The Wife of His Youth.’ Paper presented at the 8th International Cognitive Linguistics Conference. University of La Rioja, Spain. 22 July 2003. Harding, J. Riddle 2004. Simple Regrets: Counterfactuals and the Dialogic Mind. Ph.D. Dissertation, University of Maryland. Herman, D. 2002. Story Logic: Problems and Possibilities of Narrative. Lincoln and London: University of Nebraska Press. Herman, D. (ed.) 2003a. Narrative Theory and the Cognitive Sciences. Stanford, California: CSLI Publications. Herman, D. 2003b. ‘Stories as a tool for thinking.’ In Herman 2003a (ed.). 163–193. Herman, D. 2006. ‘Genette meets Vygotsky: narrative embedding and distributed intelligence’. Language and Literature 15(4): 357–380. Iser, W. 1974. The Implied Reader: Patterns of Communication in Prose Fiction from Bunyan to Beckett. Baltimore, MD: Johns Hopkins University Press. Labov, W. 1972. Language in the Inner City. Philadelphia: University of Pennsylvania Press. Labov, W. and Waletzky, J. 1967. ‘Narrative Analysis: Oral Versions of Personal Experience’, in J. Helms (ed.) Essays on the Verbal and Visual Arts. Seattle: University of Washington Press. Leech, G. and Short, M. 1981. Style in Fiction. London: Longman. Lotman, J. 1979. ‘The Origin of Plot in the Light of Typology.’ Poetics Today 1(1–2). 161–184. Oakley, T. 1998. ’Conceptual Blending, Narrative Discourse, and Rhetoric’, Cognitive Linguistics 9 (4): 321–360. Pascual, E. 2006. ‘Fictive Interaction within the Sentence.’ Cognitive Linguistics 17–2: 245–267. Propp, V. 1968. Morphology of the Folktale. [1928]. Trans. Laurence Scott.Austin, TX: University of Texas Press. Rimmon-Kenan, S. 1983. Narrative Fiction: Contemporary Poetics. London: Methuen. Rohrer, T. 2005. ‘Mimesis, Artistic Inspiration and the Blends we Live By.’ Journal of Pragmatics. 37: 1686–1716. Sanders, J. and Redeker, G. 1996. ‘Perspective and the Representation of Speech and Thought in Narrative Discourse’, in G. Fauconnier and E. Sweetser (eds), pp. 290–317. Schnepf, J. 2006. ‘Role-Value Mappings and Narrative Structure in Atwood’s The Blind Assassin.’ Paper presented the Literature and Cognitive Science Conference. Storrs, CT, April 6–9, 2006. Semino, E. and Culpeper, J. (eds) 2002. Cognitive Stylistics: Language and Cognition in Text Analysis. Amsterdam/Philadelphia: John Benjamins. Semino, E. 2002. ‘A Cognitive Stylistic Approach to Mind Style in Narrative Fiction’, in E. Semino and Culpeper, J. (eds.), pp. 95–122. Semino, E. 2006. ‘Blending and characters’ mental functioning in Virginia Woolf ’s Lappin and Lapinova, Language and Literature 15:1: 55–73. Semino, E. and Short, M. 2004. Corpus Stylistics: Speech, Writing and Thought Presentation in a Corpus of English Writing. London: Routledge. Simpson, P. 1993. Language, Ideology and Point of View. London and New York: Routledge. Stockwell, P. 2002. Cognitive Poetics. New York: Routledge. Sweetser, E. 1990. From Etymology to Pragmatics. Cambridge: Cambridge University Press. Sweetser, E. 1996. ‘Mental Spaces and the Grammar of Conditional Constructions’, in Fauconnier, G. and Sweetser, E. (eds), pp. 318–333. Toolan, M. 2001. Narrative: A Critical Linguistic Introduction (2nd edition. London: Routledge. Turner, M. 1996. The Literary Mind: The Origins of Thought and Language. New York: Oxford University Press.
TSL[v.20020404] Prn:7/01/2008; 13:33
F: PB17002.tex / p.28 (78)
Barbara Dancygier
Turner, M. 2003. ‘Double–scope stories.’ In D. Herman (ed). 117–142. Turner, M. 2006. ‘Compression and Representation’. Language and Literature 15:1 17–29. Turner, M. and Fauconnier, G. 1995. ‘Conceptual Integration and Formal Expression’, Metaphor and Symbolic Activity 10(3): 183–203. Turner, M. and Fauconnier, G. 1999. ‘A Mechanism of Creativity’, Poetics Today 20 3): 397–418. Vandelanotte, L. 2004a. Deixis and grounding in speech and thought representation. Journal of Pragmatics 36 (3): 489–520. Vandelanotte, L. 2004b. From representational to scopal ‘distancing indirect speech or thought’: A cline of subjectification. Text 24 (4): 547–585. Vandelanotte, L. 2005. Types of Speech and Thought Representation in English: Syntagmatic Structure, Deixis and Expressivity, Semantics. Ph.D. dissertation. Catholic University of Leuven. Verhagen, A. 2005. Constructions of Intersubjectivity: Discourse, Syntax, and Cognition. Oxford: Oxford University Press.
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.1 (79)
chapter
Fictive interaction blends in everyday life and courtroom settings1 Esther Pascual Vrije Universiteit Amsterdam and University of California, Berkeley
This chapter deals with ‘fictive interaction blends’ (Pascual 2002), namely simplex blends structured by the frame of the ordinary face-to-face conversation. Fictive interaction is presented as the unifying pattern underlying blends previously analyzed separately. A parallel is drawn between these and blending examples from legal settings, representing the different trial phases. These involve the conceptualization and presentation of: (i) attorneys’ serial monologues as simultaneous turn-taking; (ii) legal evidence as speaking; and (iii) the verdict as an audible message. The examples discussed appear at the levels of the discourse structure and content, the sentence and the grammatical constituent. I conclude that the conversation frame as well as the subframe of the fictive trialogue constitute fundamental structures of thought, language, and discourse. Keywords: Fictive interaction, conversation frame, conceptual blending, legal argumentation
Introduction This chapter analyzes instances of mental space mapping and conceptual blending in everyday life and institutional discourse. The focus is on what could be called ‘fictive interaction blends’ (Pascual 2002, see also Brandt this volume). These are . The research presented in this paper was supported by post-doctoral Veni fellowship 275-70012 from the Netherlands Organization for Scientific Research (NWO). I am indebted to Line Brandt, Jet van Dam van Isselt, Kashmiri Stec, Eve Sweetser, the Berkeley Viewpoint Group, and the editors for useful comments and suggestions for improvement. All shortcomings are of course my own. Contact details: Dept. of Language and Communication, Faculty of Arts, Vrije Universiteit, De Boelelaan, 1105, 1081 HV-Amsterdam, The Netherlands. Telephone: +31 (0)20 598 6411. Fax: +31 (0)20 598 6500. E-mail: <
[email protected]>; <
[email protected]>.
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.2 (80)
Esther Pascual
simplex blends resulting from the conceptual integration of a mental space with the frame of the face-to-face conversation. Fictive interaction blends are structured by the cultural model of speech as informational, such that in the default case what one says is regarded as entailing what one believes and also what is objectively true (Sweetser 1987: 47–48). The peculiarity of these blends is that their conceptual configuration shows an interactional structure which more often than not does not directly mirror the observable communicative situation in which they are set up or some sort of Reality or Fiction Space, for that matter. Fictive interaction is presented as the unifying pattern underlying blends previously analyzed separately, namely Fauconnier and Turner’s (1996, 1998, 2002) “Debate With Kant” (see also Brandt this volume); Turner’s (2002) “The Dream of the Rood;” and Coulson and Oakley’s (2006) “Voting as Speaking.” A parallel is drawn between these and different blending examples from American criminal court cases. The legal examples to be analyzed in most detail involve: (i) the conceptualization of a sequence of communicative performances of attorneys at trial as simultaneous turn-taking for the sake of the jury; (ii) the introduction of material evidence – or the lack thereof – as the deceased victim testifying in open court; and (iii) the presentation of the final verdict as the jury’s audible message to the defendant and the community at large. It will be sustained that the basic underlying configuration of the courtroom examples reflects the conceptualization of the trial event itself as a sequence of overt as well as covert interactional turns with an underlying trialogic structure. The examples discussed appear at the levels of the discourse, the sentence and the grammatical constituent. The legal examples selected for detailed analysis come from official court transcripts, televised material, and ethnographic notes from three recent high-profile murder trials in the United States. All italics and underlining in the data transcripts are mine. For privacy reasons, all names have been changed.
Courtroom interaction Language-in-interaction plays a crucial role in the courtroom. In the adversarial Anglo-American system, the presentation of evidence occurs through the questioning of witnesses by the prosecution and the defense; and the institution’s suggestion for its interpretation occurs in the discourses of the attorneys to the jury. The importance of verbal interaction in court is translated in a set of preestablished and highly regulated communicative structures (cf. Atkinson & Drew 1979; Adelswärd et al. 1987). At trial, conversational roles are strictly allocated and turns are extremely constrained. The prosecution and the defense cannot address each other during the trial, nor can the judge or jury make their feelings about the case explicit, pose questions directly to attorneys or witnesses, or interrupt the
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.3 (81)
Chapter 3. Fictive interaction blends
discourse of the legal professionals. The court’s fixed interactional pattern motivates the emergence of what I have characterized as ‘fictive interaction’ (Pascual 2002, 2006a). This constitutes an invisible – although equally present and critical – channel of communication between fictive participants, who may or may not correspond to those in the actual situation of communication. Such interaction is not imaginary or fictitious, since it is not conceptualized as occurring in a fantasy world or even in a hypothetical or counterfactual scenario. Instead, it is fictive as opposed to factive, in the sense of Talmy ([1996] 2000), as it is entirely conceptual in nature, even though it may be motivated by the actual interaction in the situation of communication. Indeed, most communicative exchanges in court do not occur for their own sake, but as ‘display talk’ for a third party (Goffman 1981: 137). A clear example of this is the witness testimony phase. This appears to be a dyadic interaction, but is in effect a ‘multi-party’ one (Cotterill 2003: Ch. 4). Since the members of the jury are the ultimate trial adjudicators, they constitute the primary intended recipients at whom the whole interaction between attorney and witness is aimed, even though (in American law) they are not verbally involved in the questioning process. Also, in the default case, the words of an attorney at trial – even when addressed at a witness – are produced in order to challenge or counterargue the (anticipated or previously expressed) views of the opposite team (Pascual 2006a). Thus, a good attorney will formulate questions and head for answers that map the types of questions and answers that would satisfy the jury on the one hand and hurt the opposite team on the other hand. Therefore, the examination of witnesses involves an unvoiced fictive communicative channel between the attorney and witness factively exchanging turns, and the silent jury and opposite team ‘overhearing’ the exchange. In this paper I discuss fragments of legal argumentation in which a fictive interaction structure is set up in discourse in order to introduce, make mental contact with or say something about apparent non-interactional realities. These are proposed to reproduce the fundamental conceptualization of the main trial stages, that is, the presentation of evidence, its evaluation, and the subsequent decision-making towards a verdict. The cases discussed occur at different levels, namely the discourse content, the sentence, the clause, the phrase and the lexical item. I try to show that the skeletal conversational structure of the examples discussed remains the same regardless of the level at which they appear. Before each courtroom blend is discussed, I will first deal with what I assume are their everyday versions, as studied in the literature on conceptual blending.
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.4 (82)
Esther Pascual
Many communicative events as one This section focuses on the common presentation of serial monologues as different turns in a fictive dialogue. This involves the conceptualization of communicative events objectively occurring at different times as compressed into one sole simultaneous conversation or debate in the blend. The examples to the discussed are the celebrated “Debate with Kant” blend and the closing arguments of attorneys at trial.
Debate with Kant One of the earliest and most well-known examples in the conceptual blending literature is the “Debate with Kant” blend (Fauconnier & Turner 1994, 1996, 1998, 2002, see also Brandt this volume). Imagine a modern-day philosopher saying in a seminar: (1) I claim that reason is a self-developing capacity. Kant disagrees with me on this point. He says it’s innate, but I answer that that’s begging the question, to which he counters, in Critique of Pure Reason, that only innate ideas have powers. But I say to that, what about neuronal group selection? And he gives no answer.
Here, two mental spaces containing the work of the two philosophers appear integrated in a Blended Space. In that space, the claims that the long-deceased German philosopher made in his books are construed as arguments and counterarguments to the modern professor’s. Also, the claims that the modern professor made centuries later become counterclaims of and questions to Kant’s propositions. Since we come to know about Kant’s ideas through his writings, this blend is first allowed by the conventional blend in which reading is conceptualized as the writer speaking to the reader(s) directly (Herman 1999; Fauconnier & Turner 2002: 210–211). The blend is further structured by the debate frame, and more schematically by “the cultural frame of a conversation” (Fauconnier & Turner 1998: 145). It is also important to bear in mind that the fictive debate between the two philosophers does not occur for its own sake. The different ideas are presented in order to convince or at least instruct a particular audience, namely the students of the modern-day professor’s. It is by ‘overhearing’ the fictive argument between the two philosophers that students are to get an idea of what their opposed philosophical positions are. Thus, in the objective situation of communication, we have the lecturer (i.e. factive addresser) telling students (i.e. factive addressees) about his own work as well as that of Kant’s, which was addressed at a German-reading audience. In the blend, however, we have the lecturer (i.e. fictive addresser and addressee) discussing with Kant (i.e. fictive addressee and addresser) for the sake of the lecturer’s students (i.e. fictive overhearers). At the same time, since the lec-
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.5 (83)
Chapter 3. Fictive interaction blends
ture is primarily meant to instruct the students, one would expect the lecturer to make mental contact with the students’ common ground, thus providing answers to questions they may have. Hence, it seems safe to say that the lecturer is also involved in a fictive conversation with the silent students. This is consistent with the classical idea of monologue as dialogue (Bakhtin1981 [1975]). In short, I suggest that in the blend, the professor’s presentation of his views and those of Kant in what objectively is a monologue to his students in fact underlies a trialogue between the modern philosopher, Kant, and the students. This is a common structure of philosophical as well as political (Lakoff 2004, 2006) and academic debates (Latour 1987) in general.
Legal monologues as fictive trialogues Contrary to the case of the two philosophers from different centuries, the main ‘arguers’ in the adversarial trial, i.e. the prosecution and the defense, are not only invariably contemporaries, but are also to present their side of the case at the same trial event. This notwithstanding, the strict interactional rules of the Western court procedure make it impossible for them to debate the relevant issues in an ordinary face-to-face discussion in which they exchange turns. As pointed out before, in the Anglo-American system the prosecution and the defense cannot address each other during the trial. Their views on the case are to be presented through separate speeches to the jury, what are called ‘opening statements’ and ‘closing arguments,’ which may not be interrupted. First, the prosecution team presents their view on the case to the jury, then the defense does so, and (in the closing argument phase) the (American) prosecutor may subsequently deliver a final speech. Even though in the Normative Space – and in the Reality Space of any Western trial – opening statements and closing arguments are presented in the form of a monologue, I suggest that they are not conceptualized by participants as serial monologues. Rather, they seem to be construed as different conversational turns in an ongoing discussion. Take for instance the following extract from an on-line forum: (2) A man is standing trial for murder. The prosecutor says to the defense, “The defendant committed the crime. Now, go and prove to me that he did not.” [. . .] But then the defense replies, “I have evidence that he did not commit the crime. Now, prove that my evidence does not exist.” To which the prosecutor promptly replies, “But I have evidence that counters your evidence. Prove that my evidence does not exist.”
In this example, the two attorneys are presented as exchanging turns, which licenses the use of imperatives and the second person pronoun to refer to their opponent (rather than the jury). The arguments of attorneys to the jury appear as a simultaneous debate between the two sides, with the jury as audience, in much the same way as in Debate With Kant.
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.6 (84)
Esther Pascual
I believe that this is the case regardless of the length of each discourse and the amount of conversational contributions involved. By way of illustration, consider the choice of words in the following explanation of what a closing argument rebuttal is, which one of the plaintiff lawyers gave to the jury in the popular O.J. Simpson trial (Los Angeles 1996: Vol. 49): (3) Right now, what I’m getting up here for is to take part in what we call rebuttal; that is, I just want to touch on some of the arguments that [the defense counsel] has made over the course of a couple days here. And after I touch on these certain things in a more general sense, [the prosecution team] will be getting up and responding in kind to the comments he made, and responding in some detail to things he put out to you during the course o- out to you during the course of his argument.
Note that in the O.J. Simpson trial, the closing arguments of the prosecution and the defense, as well as the prosecution’s rebuttal lasted one day and a half each, in which different lawyers took turns to speak for the plaintiffs and the defendant, objections and side bars occurred, and the judge interrupted on numerous occasions to instruct the jury on how to interpret the argument phase. Still, the interaction frame seems to prevail as the basic underlying structure of the entire closing argument phase. In the blend, different speech events which occurred during different days – each being objectively addressed to the jury – appear compressed in one sole conversation between the attorneys for the sake of the jury. The understanding of the closing argument phase as a fictive conversation or debate between the two sides also seems to be present when there is no possibility of rebuttal, as in the Spanish system. Take the following piece of discourse extracted from a prosecutor’s closing argument to the jury in a murder trial I observed in the Barcelona county court in 1997 (my translation): (4) The prosecutor is the first one to speak, because the law says that the prosecutor is the first one to speak, because the law wants him to speak first. If the prosecutor does not speak after the defense attorney has spoken, this is not because he does not want to or because he has been convinced by the defense attorney, but because he cannot.2
In (4), the prosecutor in this extract seems to anticipate that the Spanish jury may conceptually integrate the attorneys’ serial monologues with the ordinary conversation frame and wants to make sure that they do not project too much into the
. Spanish original: “El fiscal habla el primero porque la ley dice que el fiscal hable el primero, porque la ley quiere que hable primero. Si después del abogado defensor el fiscal no habla no es porque no quiera o porque el abogado defensor le haya convencido, sino porque no puede.”
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.7 (85)
Chapter 3. Fictive interaction blends
blend. Indeed, in an ordinary debate, it is commonly accepted that the interactant who ‘has the last word’ wins the argument. Interestingly, the fictive interaction blend does not only seem to become manifest in the attorneys’ vocabulary and fictive quotations when speaking about the argument phase, but also in their syntactic choices. Consider for instance the discourse fragment below from a prosecutor’s closing argument rebuttal in a high-profile murder trial which I did fieldwork on in California in 2000 (Pascual 2006a: 391–393): (5) Now, Mr. Loeber [defense attorney] questions, Well, how could the blood get on the end of the poker, because the poker is not hitting her in the head?’ [. . .] The reason why blood gets on the end of the poker [. . .] is centrifugal force.
Here, the prosecutor presents and subsequently answers a question that appears to have been previously raised by his adversary. However, the defense attorney actually never produced such a question. In fact, he used no interrogatives in his attempt to cast doubt upon the accuracy of the prosecutor’s argument. The attorney’s exact words were: (6) And we know from dr. Stone’s [forensic expert] testimony and from our own common sense, when we look at these unfortunate, sad photographs of Rachel [victim] from the coroner, that there were no wounds there that correspond to the end of a fire poker. They’re linear wounds. That’s why we have linear, linear, linear. But to get that castoff spatter we have to have blood on the end of the poker, and that would get there most likely – we’ve had no other explanation – by the end of the poker hitting Rachel’s head.
In (6), the defense attorney is pointing out an apparent paradox arising from two supposed states of affairs: (i) the victim’s wounds do not correspond to the end of a fire poker, and (ii) there must have been blood from the victim’s head at the end of a fire poker. By presenting these two contradictory scenarios, the defense attorney challenges the prosecution’s interpretation of the facts. According to the prosecution, the fireplace poker that is missing from the couple’s home corresponds to the murder weapon, which was never found. What is most striking about (5) is naturally the re-presentation of the defense attorney’s entire argument in (6), which consists of a set of assertions addressed at the jury, as a single sentence, namely a question that he then proceeds to answer. Significantly, when I asked him to comment on this extract, the prosecutor used the question-answer pattern again and characterized his counterargument as a response: ‘Albert [defense attorney] was saying that, you know, if it was a poker, why aren’t there poker marks in her head? [. . .] so, I was just responding’ (Int.2-DA: 10). In the Reality Space of the actual trial we have the two opposed attorneys arguing their case in serial monologues addressed to the jury. First, the prosecutor
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.8 (86)
Esther Pascual
delivers his argument to the jury at time 1, then the defense attorney presents his alternative argument to the jury at time 2 (extract 6), and subsequently the prosecutor addresses the jury one more time at time 3 (extract 5). In the fictive interaction blend, the defense attorney is arguing with the prosecutor in a simultaneous debate in which the prosecutor’s first argument raises a challenging question by the defense, which the prosecutor subsequently responds to, with the jury as overhearer. Also, the long discourse of the defense in (6) becomes compressed into one utterance, namely an interrogative. This choice of sentence type is particularly fortuitous, since it allows the prosecutor to use it both as a concise paraphrase of the defense’s previous challenging argument (since rhetorical questions are understood as negative assertions) and subsequently reframe it as a challenging information-seeking question, which he can then proceed to answer. Thus, the overall configuration is not only structured by the frame of the ordinary conversation, but more specifically by the question-answer pattern. In the blend, the answer to the question represents the last word in the fictive debate, and thus the indication that the prosecutor has won. It should be noted at this point that legal monologues are also generally conceptualized as involving an inaudible ‘two-way communication’ or ‘dyadic conversation’ between the attorney speaking and the judge/jury counterpointing the attorney’s words with an inner discourse of their own (Harré 1985; Walter 1988; Stygall 1994). Indeed, the fictive argument set up between prosecutor and defense attorney occurs for the sake of the jury, as the ultimate evaluator. Since the jury cannot make their feelings about the case explicit or pose questions directly to attorneys or witnesses, attorneys need to anticipate the questions they may have and make sure to provide satisfactory answers to all of them. Thus, the prosecutor in this case was not only taking the voice of the opposite attorney and responding to him in a fictive interaction blend. He was simultaneously also posing questions the jury might have (and which they might have asked him about, were that allowed). Hence, the rhetorical question posed and subsequently answered by the prosecutor is not only mapped onto the negative assertions previously produced by the defense attorney. It also represents doubts or questions the prosecutor believes may be in the minds of skeptic jurors.3 This complex network is schematically represented in Figure 1 (where the prosecutor is P, the defense attorney is D, and the jury is J).
. This is very much how the prosecutor explained his use of interrogatives in his closing argument in my interview with him: (i)
I understand the defense will bring up those questions in their closing argument, so I bring them up myself, and I defeat them. So when the defense brings them up, they’ve
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.9 (87)
Chapter 3. Fictive interaction blends
Figure 1. The prosecutor’s fictive trialogic question
Presenting a constructed rhetorical question ascribed to one’s adversary and subsequently responding to it – having therefore reframed it as an informationseeking one – is an effective argumentative strategy, since it simultaneously serves to: (i) present the opponent’s challenging argument in a concise manner; (ii) ask a question the jury may ask themselves; and (iii) set up the cognitive and discursive grounds for a counterattack. Furthermore, I believe that the use of this questionanswer structure is not only conceptualized and used as an ordinary adjacency pair so that the former merely prepares the addressees and overhearers for the latter. It also seems to reflect a conceptualization of the trial event itself as a sequence of conversational turns, in which what has been said earlier or what is anticipated to come later are commonly explicitly or implicitly either questioned or answered by legal professionals for the sake of the judge/jury. already heard it! They, th-the jury knows what my response is the defense argues [for]. (Int.2-DA: 12) (ii) I’m answering questions that I think the jury will be asking in, in the jury room, [. . .] I’m just anticipating! I’m asking what a logical person might ask. Now, ‘what about this, mister Geisberg?’ ‘What about this?’ I wanna answer all the questions! (Int.2-DA: 13)
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.10 (88)
Esther Pascual
Such a conceptualization may also be reproduced in instances of fictive interaction within the sentence (Pascual 2006b, see also Brandt this volume). Consider for instance the examples below from the prosecutor’s discourse in the same murder trial (Pascual 2006a: 394, 396): (7) a.
Express malice means, simply, was it an intentional killing[?], okay? Did the person who killed think about it? Did they have a choice? b. Now, was there an attempt to kill? This is kind of a ‘who’s buried in Grant’s tomb[?]’ argument, but [sly smile] you have to think about this.
In (7a), the definition of a legal term is presented as a set of interrogatives. This involves a conceptual integration of the Law Space, with its rules and definitions, and the Deliberation Space, in which these rules and definitions are to be used in the evaluation of the facts. By using questions for the definitions of a legal term in the Law Space, the prosecutor succeeds in presenting the meaning of the term through the expected reasoning process that the actual jury in the Deliberation Space will have to go through when trying to decide whether or not this term applies to the facts in the case at issue. Thus, the questions presented by the prosecutor as definitions are mapped onto hypothetical questions jurors may ask themselves and each other in the jury room in the Prosecutor’s Belief Space. Note too that these questions correspond to the issues the defense has questioned in the Trial Space. Presenting this definition in terms of a question also prepares the cognitive and discursive grounds for answering it in a way that is favorable to the prosecutor speaking. Indeed, the extract in (7a) was followed by: “Well, let’s apply this. If you apply it to this case, was there planning? Of course there was planning.” In (7b) an interrogative at the lexical level is used to characterize the prosecutor’s own argument. The question ‘Who’s buried in Grant’s tomb?,’ which comes from American children games (to be blended here with the courtroom situation), is peculiar since its answer is presupposed in the question itself: Grant is the person who is buried in Grant’s tomb. The prosecutor indirectly warns the jury that even though his argument may state the obvious, he still needs to make his point, since the defense had previously cast doubt upon – or questioned – his theory of the case, obvious though it might seem to anyone who had followed the case. Thus, the absurd question from the children’s game is mapped onto the defense’s entire critical argument as a challenging question to the prosecutor, which skeptical members of the jury might also want to ask. In his argumentation to follow, which would correspond to the answering of the self-answered question in the game, the prosecutor is fictively responding to the defense’s prior and anticipated future challenges as well as to the jury’s possible doubts. Thus, even though in these examples only one fictive utterance is overtly expressed, I suggest that this one ‘utterance’ needs to be understood as a fictive conversation turn in the fictive trialogue which
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.11 (89)
Chapter 3. Fictive interaction blends
is assumed to characterize the participants’ conceptualization of the most relevant communicative events in Western courts. I hope to have shown that the rhetorical effectiveness of questions in legal monologues is due to the attorneys’ ability for compression and blending, the resulting network being structured by the frame of the ordinary face-to-face conversation. This involves taking the perspective of the final evaluators as well as that of the opposite team. More specifically, questions in legal monologues generally show an underlying triadic structure, as they serve a double persuasive function, namely: (i) turning silent addressees into co-constructors of discourse and (ii) challenging the version of the facts proposed by the opposite team.
The inferable as speaking This section deals with the presentation of the source(s) of a particular inference as speaking to the one(s) to draw this inference. This blend is extremely common in everyday thought and language. We talk about an event as being very telling, an object as saying something about its owner, or a thunder as announcing a coming storm. The examples discussed in this section are the image of a cross speaking to a sinner and the presentation of legal evidence as ‘speaking for itself.’
The speaking Cross Consider the old Anglo-Saxon poem “The Dream of the Rood.” This literary work has been analyzed by Turner (2002) as an example of a complex conceptual integration network, involving various interesting blends. The one that is relevant to the present discussion involves the presentation of the Rood, the Holy Cross, as appearing to a sinner in a dream, and speaking to him of his experiences. The Cross’ verbal behavior is presented as real in the poem’s Dream Space. As Turner points out, this image is a vivid exploitation of the conventional blend in which an observer’s inference on the history of a physical object appears as that object actually speaking to the observer. It is not unusual for archaeologists to speak of what a mummy or an ancient vase tells them, for instance. Similarly, Oakley and Coulson (this volume) show how the interpretation of different events as being related to one another can be construed and presented in discourse as dots screaming at the ones that should connect them and draw the relevant inferences. It should be noted that this blend also seems to be allowed by the general blend in which non-human animals and objects become personified and speak, regardless what one may inference from their appearance. The characters speaking in fairy tales and children’s games are not always human. Also, we adults often talk to plants or machines, and speak of an appetizing piece of pie, for instance, as “saying eat me,” thereby projecting intentionality in the blend. In “The Dream of
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.12 (90)
Esther Pascual
the Rood” blend, the Cross appears as the addresser, the sinner as the addressee, and the reader of the poem as the overhearer of the conversation between Cross and sinner.4
The speaking evidence Since legal discourse is not significantly different from literary discourse (Pascual in press), or ordinary conversation for that matter, one should expect to find fictive interaction metaphors in legal or law-related discourse. Indeed, this seems to be the case. The first example of this kind is from a brainstorming session among three prosecutors from a prosecution’s office in California, broadcast by the American NBC television channel in July of 2002. In this case, the prosecution accused the defendant of sexually assaulting a 23-year-old woman. The defense argued that the sexual contact between defendant and victim was entirely consented by both parties. At a point in the discussion among the three prosecutors, one of them said: (8) He left all kinds of evidence that he won’t tell us out of his mouth! The broken jaw, the semen. . . I assume there’s physical findings. . . [. . .] So, basically, her body is telling us what he won’t!
In this case, the statements of the victim and the defendant as to the nature of their sexual contact were contradictory. Since both statements had the same legal value, the prosecution’s accusation in this case relied almost exclusively on physical evidence from forensic doctors, who would be called to testify in court and show the jury pictures of the victim’s bodily injuries. This involves an EFFECT FOR CAUSE metonymy (Panther & Thornburg 2000). It is through having found injuries on the victim’s body (i.e. effect) that one can conclude that she was sexually abused by him (i.e. cause). By so doing, the victim’s battered body can ‘tell’ the story. The presentation of the physical evidence on the victim’s body as the body telling its observers about its violent past allows for the setting up of an identity mapping between the confession that is conspicuous by its absence in the police interrogation of the defendant (structured by a negation network) and the inference that can be drawn from the victim’s injuries after the attack. This makes comparison of the one with the other easier, and thus helps draw the inference that the defendant is lying, that the victim is telling the truth, and consequently that their sexual contact was not consensual. Note that this comparison would be rhetorically less straight-forward if the SEEING IS KNOWING metaphor (Lakoff & Johnson 1980) had been used instead,
. This conceptual network is in line with Tobin’s (2006) work, which shows that readers are mostly construed as overhearers.
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.13 (91)
Chapter 3. Fictive interaction blends
as in the alternative formulation “her body is showing us.”5 Just as was the case for Turner’s speaking cross blend discussed previously, the configuration of the speaking body is based on a conventional blend in which an interpretation or inference is conceptualized in interactional terms as the source of that interpretation or inference speaking to the one that draws it. In English and other languages this culturally meaningful blend is reflected in the polysemous meaning of verbs of speech (e.g. “see what it’s telling you,” approximately paraphraseable as “see what it’s inviting you to infer,” Baynham 1996: 74). Hence, it should not be surprising that in (8) this same verb of communication is used (“her body is telling us...”).6 Finally, note that the body can only tell the story to those who want to listen, that is, to those who have been looking for answers in her body. Indeed, when asked the right ‘questions,’ legal evidence can be presented as speaking in a similar manner as a witness may. Consider now an example from the same murder case from which examples (5)–(7) come. The defendant in this case was a financial manager accused of brutally killing his wife in the couple’s home. No clear evidence or alibi was provided to prove the defendant not guilty and he was the only witness for the defense. This notwithstanding, the defendant insisted on his innocence and testified under oath that he was at work at the time of the crime. The example to be discussed, which is from the district attorney’s closing argument rebuttal, is (cf. Pascual 2002, forth.):7 (9) But interestingly enough, Rachel did, in a way, testify through circumstantial evidence, and that is this: the defendant readily admits on the August the 26th interview that Rachel had no enemies. Everybody loved her. There wasn’t one person who came into this courtroom over the last three weeks and said, “Boy, Rachel’s a bad person.” There isn’t one person who has a motive to kill Rachel Coff. There isn’t one person who was stalking Rachel or saying anything bad about Rachel or that Rachel had a boyfriend on the side or anything like that. . Thanks to Jannis Papalexandris for this observation. . Consider also the pieces of discourse below, from a prosecutor’s discourse in murder trials occurred in New York and California (Pascual 2002: 161–162): (i)
That’s not what the bullet is telling me.
(ii) . . . there is an absence of spatter on those pills that tells you that the pills had to be deposited after her injuries, [. . .] which is what tells you the killer had time. (iii) I actually prefer it when no big coverage is given [of a verdict] because it tells the defendant we don’t care about you any more, you don’t get to be on the first row, we don’t have to hear what you have to say. Goodbye! Go live in a cell! . For a more detailed analysis of this example, examined through the eyes of the participants’ knowledge of the embedding discourse and trial as well as their conceptualizations and depictions of what a trial is, see Pascual (forth.).
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.14 (92)
Esther Pascual
Rachel had no enemies. People describe her as articulate, witty, assertive. She was a career woman. But because of that, she speaks out, because with no enemies, there’s nobody who’s gonna break into that home, and there’s no signs of forced entry in the home and there’s nothing stolen from the home. So it points to the defendant.
In (9), different bits of circumstantial evidence, which were presented to the jury through the attorneys’ examination of over thirty witnesses for two weeks, appear “in a way” compressed into one sole testimony, namely the testimony of the victim’s in the trial for her own murder. This image succeeds in presenting various pieces of diffuse evidence in a compressed and culturally meaningful human scale scene. Various conventional blends are involved in this image. An identity mapping is first established between the circumstantial evidence presented by the prosecution team in the Present Reality Space of the ongoing trial and the victim’s lifestyle in the Past Reality Space prior to the crime. This evidence is mostly in the negative. The prosecutor’s case is primarily based on the lack of evidence that would support a different hypothesis from the one he is arguing for. Hence, the conceptualization of the accusatory evidence itself involves conceptual integration in a negation network (cf. Fauconnier & Turner 2002: 241). It is through the sum of these missing pieces of evidence that the victim’s fictive voice can be heard. In its turn, the negative evidence from the victim’s life attains its legal value by being conceptually mapped onto elements in the prosecutor’s interrogation of various witnesses in the Under-Oath Communicative Space of Past Witness Testimony: the defendant ‘admit[ted]’ that the victim had no enemies; nobody ‘said’ that she was a bad person; all witnesses ‘describe[d]’ her as articulate, etc. Critically, since the inference that “there’s nobody who’s gonna break into that home” comes from consideration of the victim’s life, the image of the victim speaking up involves the conventional blend in which a source of inference speaks to those who draw it. At the same time, since jurors are the final evaluators, it may also be accurate to postulate that the victim is “in a way” speaking directly to them, as they are invited to look for the answers to their questions on the case in the victim’s life. As it is, were the victim still alive, she would have been called to testify. In that case, the jury would mainly come to know about the circumstantial evidence in the case through her answers to the attorneys’ questions in witness testimony. Critically, it is not uncommon for attorneys and the jury to understand and talk about the attorney-witness exchange as the witness speaking to the jury. In fact, attorneys often overtly ask witnesses to address the jury directly when answering the attorney’s questions. Hence, just as was the case for the examples discussed previously, the overall conceptual network set up by the image in (9) is structured by the frame of the ordinary face-to-face conversation. This frame is in its turn being fur-
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.15 (93)
Chapter 3. Fictive interaction blends
Figure 2. The testimony of the deceased
ther structured by the law-specific frame of under-oath testimony (see Figure 2), since the victim not only “speaks out,” but is actually presented as “testifying.” This is crucial, since in the American system a direct identity relationship is established between the content of under-oath testimony – with no contradictions or proven lies – and the truth. Also, since sworn testimony is technically the only means for the jury to gain access to the facts to be evaluated, the jury is instructed to consider only that which they have heard through testimony. Note too that given the law’s assumption of equivalence between sworn testimony and the ‘truth,’ that which is heard through testimony constitutes direct evidence. Thus, the presentation of the murder victim ‘testifying’ in (9) actually involves the construal of circumstantial evidence as direct evidence of guilt. This is non-trivial since, although circumstantial evidence against the defendant seemed rather compelling in this case, no direct accusatory evidence could be provided by the prosecution. The crime was committed in the seclusion of the couple’s bedroom and there were no eyewitnesses. Moreover, the defendant’s sworn testimony in self-defense technically constituted direct evidence of his innocence. Bearing this in mind, I believe that the victim’s fictive testimony in the Current Discourse Space of the prosecutor’s argument needs to be construed within the light of the defendant’s factive testimony in the Post-Crime Reality Space of the
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.16 (94)
Esther Pascual
actual trial. Indeed, the victim’s testimony embodies the evidence that “points to the defendant,” and without that testimony, it actually does not make sense. The victim’s exemplary life seems to be implicitly presented as a counterargument of or a response to the defendant’s testimony in self-defense. If the victim had been alive, able and willing to testify, her testimony would have been weighted against the equally direct evidence of the defendant’s testimony, which was the only evidence of his innocence. Hence, in (9) the fictive argument between the prosecution and the defense for the sake of the jury appears as an implicit verbal confrontation between those from whom evidence for the one and the other version of the facts was obtained: the victim (through her life, described by witnesses) and the defendant (through his testimony under oath). This configuration is consistent with the common understanding of the (often long) sequences of attorney-witness exchanges as simultaneous turn-taking among witnesses. Take for instance the way in which an attendee at the trial at issue explained how the daughter-in-law of the victim and the defendant was purposely called to be questioned immediately after her husband’s testimony (Int.8-Nov: 11): (10) of course Tracy [daughter-in-law] followed Tom [son], but they couldn’t speak between each other’s testimonies, and Don [prosecutor] used her to contradict her husband.
Even though these two witnesses did not engage in any verbal exchange during their testimonies, the various interactional sequences between the prosecutor and the second witness are presented as the witness’s fictive response or counterargument to what the previous witness had told the prosecutor as addressee and the jury and court as overhearers. Significantly, in this particular trial the crossexamination of the defendant served the prosecutor’s case particularly well, since it resulted in a long set of incongruous allegations and transparent contradictions, which revealed the defendant as an untruthful witness. Therefore, it seems that, when running the blend, the jury should easily agree on which testimony to believe. In (9), reframing lose pieces of (negative) circumstantial evidence as one concrete testimony of the best eyewitness succeeds in presenting the evidence against the defendant in the same form and in the same conceptual domain as the evidence in favor of him. Also, presenting the victim and the defendant – rather than the attorneys – as the ones implicitly debating the opposed versions of the facts is argumentatively effective, since it uses the individuals the case is actually about with no intermediaries or representatives in between.8 In sum, the overall config. The rhetorical choice of having the defendant implicitly ‘arguing’ against the victim’s fictive testimony fits well with the characteristics of the case at hand, in which, counter to what is customary, the defendant – rather than the defense team – seemed to be the one to be most
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.17 (95)
Chapter 3. Fictive interaction blends
uration successfully brings together: (i) the crime and subsequent trial on it; (ii) the presentation of evidence through testimony; and (iii) its subsequent evaluation compressed in the same Blended Space. The compression is as human-scale as can be. The image of a deceased person being conceptually called for a posthumous testimony is often used in modern (American) litigation and was even recommended by Quintilian (see 1921 translation). On occasions the murder victim is explicitly called upon to answer questions one can answer oneself by considering the evidence on the case. An example is the prosecutor who ended his closing argument by presenting the counterfactual scenario of a “miracle,” in which the deceased victim was brought back to life and told the jury that she had already told them who her killer was (Coulson & Pascual 2006: 171–175). Finally, since the blend of the source of inference as speaking is culturally meaningful, a piece of evidence can be presented as “crying out for an explanation” or “shouting out to someone” for instance. What the evidence is fictively telling the ones who understand it can also be overtly expressed through a string in the direct speech. Consider for instance the examples of embedded fictive verbal interaction in (11) below. These examples come from a personal weblog, a defense attorney’s conversation with the judge in a criminal case, and an online forum on the aftermath of the September 11th attack: (11) a.
....if you can find definite solid evidence that screams to you “yes! Jesus is real!”, then... b. we don’t really know because there’s some physical evidence that says, hey, we got this injury. . . c. all those people who were told to shut up about their disagreement have so much “I told you so” evidence, that some serious re-alignment has to occur.
In (11a), a hypothetical piece of evidence is presented as screaming what can be inferred from it to the one(s) to find the evidence. In (11b), the noun phrase “some physical evidence” is modified by a restrictive relative clause presenting this evclearly counterargued. Consider for instance, how the prosecutor began his discourse following the defense’s (Vol. 6, 1452: 10–15): (i)
The purpose of my rebuttal closing argument is to specifically address some of the points that Mr. Loeber [defense attorney] addressed. And what I think is important that you realize before I make this argument is my comments are not directed at Mr. Loeber personally. The comments are directed at the defendant and the law as it applies to the defendant.
In fact, in a four-hour feedback interview, the chief deputy defense attorney admitted to me that defense team was convinced of the defendant’s guilt (Int.9-DC: 10–11).
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.18 (96)
Esther Pascual
idence as speaking in conversational style and including the verb “to say” as a space-builder of the fictive address. In (11c), a direct speech constituent (i.e. “I told you so”), showing the first and second person deictic pronouns, is used as a modifier of the noun “evidence.” The referent of this noun needs to be construed as the one saying “I told you so” to those who had not believed that type of evidence would ever be provided. In sum, it is not uncommon to conceptualize the source of an inference as speaking to the one(s) to draw the inference in a fictive interaction blend. Hence, it makes sense to present a murder victim as speaking up, or even testifying through circumstantial evidence on her life, as well as to characterize a type of evidence through a direct speech constituent representing what that sort of evidence fictively says. In the courtroom, legal evidence may be presented as speaking to the professionals or to the jury, always as a counterargument to what the evidence of the opposite team may suggest.
The final decision as a moralistic address This last section deals with the presentation of a final evaluation in conversational terms. The examples discussed are the Voting As Speaking blend and the construal and presentation of the verdict in court as the jury saying something to the defendant and the community at large.
Voting as speaking Coulson and Oakley (2006: 54–55) discuss a political letter in which electoral polls and democratic elections appear as the voice of the people, with the ability of sending a “message” to a political party. The relevant fragment of the letter reads: (12) All the public opinion polls [. . .] have said the same thing over and over: The American public does NOT want impeachment. Yet, Congress has decided to tell the public to take a flying $#@& and has moved ahead with the impeachment process anyway. The only way to send a true message to the right wing is to throw every Republican out of office. The message would be loud and clear to all these new Democrats – THE AMERICAN PUBLIC WANTS THE AGENDA OF THE (so-called) CHRISTIAN RIGHT REMOVED FROM THE HALLS OF OUR UNITED STATES CONGRESS!
Coulson and Oakley analyze this fragment as involving a metaphor that presents a political process as an “interpersonal argument,” involving massive compression in a fictive interaction blend. The public polls are first personified and used metonymically to stand for the American people as a whole and not just the voters questioned. These polls have then the ability to verbally express the opinion of the
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.19 (97)
Chapter 3. Fictive interaction blends
people. The Congress ignores that advice and speaks to the public in a joint voice. Then, the public is encouraged to respond to the Congress’ ‘words’ by sending “a true message to the right wing,” namely “to throw every Republican out of office.” Even though the election polls, the Congress decision regarding the impeachment process and the actual elections occur at different times, in the blend they appear as an integrated event scenario. The event in the blend is a verbal argument or debate between the American public and the politicians in power. In this argument, the sum of all the individual acts of voting for Democrats are compressed into one sole voice representing “the citizenry’s turn in conversation” (Coulson & Oakley 2006: 54). Unlike the vote in the Reality Space of politics, this conversational turn is one that can be heard. Thus, the number of votes maps onto the loudness of the message that voting is presented as sending. Following our cultural understanding of everyday conversation, the louder the message the more conviction we attribute to the sender(s). In this letter, Moore suggests that if enough citizens vote for Democrats, the message to the opposite party will be so forceful as to end the public debate. The presentation of voting as speaking discussed by Coulson and Oakley is certainly not a one-time configuration. Sapir ([1949] 1986: 104) for instance, speaks of (my italics) “such simple acts of communication as that John Doe votes the Republican ticket, thereby communicating a certain kind of message.” Note too that voting can be construed as communicating something to fellow voters and not only to the political candidate parties to be elected. An example is the cartoon drawing of a pie chart, whose labels appear to be different conversational turns in a heated political debate between the percentage of voters sharing particular opinions (Pascual 2002: 16–17): (13) Thinks country is ‘divided’ Is not! [. . .] Does my opinion count? [. . .] Don’t blame me! I didn’t vote! Don’t blame me! I voted for everyone!
It could also be added that in the Voting as Speaking blend, somebody (i.e. the group of voters), speaks to somebody else (i.e. the losing party), with a third group as audience (i.e. the entire population of voters and non-voters to be ultimately affected by the outcome of the debate). Thus, the fictive conversation is an ‘adversarial’ one that seems to suggest a trialogic structure.
The jury verdict as an audible message The courtroom equivalent of voting in a political election is coming up with a verdict. Interestingly, in the same way as voting can be construed as an act of communication, so can reaching a verdict. Consider for instance the following extract
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.20 (98)
Esther Pascual
from a prosecutor’s opening statement to the jury in a no-body trial aired by NBC television: (14) . . . at the end of this trial I want you to do two things. One, I want you to tell the defendant that he is guilty of murder. And two, I want you to tell the defendant that there’s no such thing as a perfect crime.
Much as they might want to, in the strict sense jurors in the Reality Space of the trial in which this piece of discourse was produced could never do what the prosecutor asks them to in (14). In the Normative Space of the Law, jurors are not allowed to speak during the trial or sentencing phase, let alone address the defendant directly, as the judge or family members of the victim may at sentencing. Moreover, jurors are not the ones to read the verdict; a clerk reads it instead. All jurors do is assert when asked by the clerk whether the verdict read was their “true verdict.” In fact, the members of the jury do not even write down the verdict in their own words. After the deliberation, once they have come to a unanimous agreement on all the relevant issues concerning the case, the foreperson of the jury fills in and signs up the verdict form prepared by the court. This form shows a set of binary options to chose from (“guilty/not guilty”; “did/did not”; “was/was not”). These seem to be conceptualized as yes/no questions on the case that the court asks the jury to answer. In fact, in American law the verdict form is sometimes explicitly presented as a list of questions, something that is the default case in the Belgian jury system. It is through the jury’s answers to these questions that the defendant, the court and the public in general can learn what the jury’s views on the case are. Thus, even though the jurors’ addressed at in (14) did not have to decide whether there is “such thing as a perfect crime,” that they thought this was not the case could be inferred from their accusatory verdict. The presentation of the jury’s verdict as them telling something to the defendant in this example is not a one time occurrence. Also, I believe this is not a mere rhetorical device, often as (American) attorneys use it. Rather, I suggest that it is a reflection of the way in which attorneys and jurors alike conceptualize the final verdict. In order to support this, I will discuss fragments of the full transcript of a real-life jury deliberation in a death-penalty case for double murder, which occurred in an Ohio court in 2004. The transcript was released by ABC television as complementary material relating to the documentary series ‘In the Jury Room.’ Just as was the case for voting, the jurors in this case seemed to understand their verdict as a “message.” As it is, in their deliberation the word “message” was used 13 times in relation to their verdict. Consider for instance the example below, produced by the foreman at the beginning of the deliberation on sentencing: (15) . . .this is the sentencing phase, to approach the sentencing options carefully and with forethought. To see what message we’re sending or presumably send-
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.21 (99)
Chapter 3. Fictive interaction blends
ing with each one of the four counts [...] What message we’re sending A to Mr. Donald [defendant] and what message we’re sending to[,] as the conscience of the community, we’re sending to the community with our deliberation, with our final suggestion. (juror 1, JurDel-B, p. 182)
In the Present Reality Space of the ongoing deliberation, which is mapped onto the Normative Space of the Law, the jury only get a set of “sentencing options” from the court, which cannot be changed or paraphrased. This notwithstanding, in (15) their choice between them – or rather what can be inferred from their choice – is presented as the jury’s own message. The verdict is not portrayed as a decision on what type of sentence the defendant deserves, but as a concrete act of communication addressed to the one who committed the crime that requires punishment. Also, by being the defendant’s ‘peers,’ who are selected (roughly) at random, jurors metonymically stand for “the conscience of the community.” Therefore, since in the Normative Space of the Law the jury’s decision is unbiased, it maps onto any future verdict on a similar case. This allows the jury’s message to be presented as also addressed to the community at large. By ‘responding’ to the court’s ‘questions’ through their verdict, the jury is construed as fictively speaking to the defendant and the community directly as overhearers of the fictive verbal exchange between court and jury. Interestingly, just as was the case for the voting as speaking example, the jury’s fictive message to the defendant and the community can appear verbalized in a concrete fictive utterance: (16) What is the message that we are sending A. to Mr. Donald and B. to the community? To the community we are saying this punishment that we are meeting out is harsh; it is commensurate with the crime and murders must be paid; there’s a payback for murder and it is losing your freedom for the rest of your life [. . .] There are other possible messages and those other possible messages in my opinion are based on our sentencing. [. . .] The other mercy factor, the other messages, excuse me, is that in some way, because of the mitigating circumstances, surrounding this case, that we looked at carefully, we have assessed as a jury that we empathize with your situation, Mr. Donald and we are not going to punish you up to the full letter of the law because we feel that the mitigating circumstances are equal to the aggravated circumstances. (juror 1, JurDel-B, p. 184)
In this fragment, the jury’s final decision, which in the Present Reality Space of the ongoing deliberation was reached through multiple turns among jurors for two full days with time left for individual reflection, is not presented as a description. Rather, it is presented demonstratively with conversational style as a harsh speech to the community at large and a compassionate speech to the defendant himself. Since in the Normative Space of the Law the verdict is neither addressed to any-
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.22 (100)
Esther Pascual
body in particular nor written down by the jurors themselves, these ‘utterances’ cannot be actual quotations of actual previous ones and hence constitute instances of ‘constructed dialogue’ (Tannen 1986, 1989). They represent the compressed verbal counterpart of the feeling of all jurors, using what Tannen (1989) calls ‘choral dialogue.’ In the blend, the jury’s suggested punishment appears as the joint fictive voice of all jury members speaking as one fictive individual (Langacker 1999) engaged in a fictive verbal argument. Thus, the jury’s reasons for punishment are metonymically linked to the punishment by an identity connector, which seems to represent the mental version of the conventional Cause-Effect vital relation. Also, the jury’s moralistic justification for their choice in (16) corresponds to what one could infer from their decision, rather than the text in the actual verdict form, which does not include a section on verdict motivation. The presentation of the jury’s fictive address to the defendant in (16) constitutes a clear case of fictive interaction within the sentence (Pascual 2006b). There, the verb ‘to assess,’ which is a verb of reasoning rather than communication, is followed by ‘that’ plus a string in the direct rather than the indirect speech. A grammatical blend seems to be involved, in which formal and functional properties of direct and indirect speech are integrated. Note too that whereas in the verdict form the defendant is referred to in the third person and with full names and surnames, the fictive interaction blend, which turns the verdict into a face-to-face conversation, allows the use of a vocative such as “Mr. Donald” and the second person pronoun to refer to him. In the cases just discussed a decision arisen from a long set of conversational turns among jurors is compressed to human scale into one sole act of communication in the blend involving the jury and the defendant as fictive participants. On occasions even more compression of the diffuse is involved in running the blend. As pointed out in the discussion on example (15), it is not unusual for juries to view their decision as one that stands for the decisions of jury counterparts in Past and Future Hypothetical Spaces. Consider for instance: (17) . . . the fact is he [defendant] got out and continued on a course of action that landed him in jail again and after a stent in jail he got out and he has been incarcerated again. And how many times does he have to be incarcerated before we say, you know what? You’ve had all these chances. You’ve had them. When does it stop Mike? When does it stop? (juror 7, JurDel-B, p. 150)
In this fragment, the utterer seems to indicate that the jury needs to teach the repeat offender a lesson by coming up with a severe verdict. The jury member speaking is framing the sentencing decision at hand as their final scolding of the defendant, after he has been given a second chance by two previous juries. Not only do we have the presentation of the joint fictive voice of the jury as a fictive individual, but also the voice of their counterpart juries in previous trials. Since the jury is an institutionalized legal entity and a trial a public matter, a metonymic
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.23 (101)
Chapter 3. Fictive interaction blends
Figure 3. The jury’s verdict as the last community’s scolding of the defendant
mapping is established in which the present jury stands for all the juries who have been entrusted with the task of judging the defendant’s conduct. Thus, in (17) the verdict at hand is contextualized within the defendant’s entire criminal history, understood as a heated argument. More specifically, the jury’s present message to the defendant represents one conversational turn between the defendant and the law in a fictive debate that he started with the commitment of his first criminal act. By so doing, the utterer of (17) succeeds in presenting her suggestion for sentencing together with the motivation behind it, as well as frame the crime within the defendant’s entire criminal history. The image presented suggests that if their message is severe enough, it will be the last word to end the debate (see Figure 3). In sum, whereas in the Reality Space of the trial the defendant is not the one to which the verdict is addressed, as it is read in open court for all to hear, in the cases analyzed here the defendant appears as the direct (fictive) addressee of the jury’s verdict. He can thus be presented as addressed in the second pronoun and his proper name can be used in the vocative. Note too, that since the frame of the ordinary conversation is used in running the blend, there is also conceptual space for overhearers. Indeed, the victim’s family, who might not have been present at the trial, may appear as overhearing what the jury fictively says to the defendant. Consider for instance the extract below:
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.24 (102)
Esther Pascual
(18) . . . we are giving mercy to Vanessa’s family and to Dan’s [victims] family, uh, by having sentenced this man already and giving [them] the conviction and the satisfaction that they can sleep knowing that we said, okay, yeah, you did murder Dan, you did murder Vanessa. (juror 12, JurDel-B, p. 134)
In this fragment, it is by having fictively overheard the jury’s verbal accusation to the defendant that the family of the two victims are to learn how their family members passed away. Finally, the understanding of a relation between verdicts and their possible translation in some verbal exchange also seems to be present in the overall conceptual configuration underlying direct speech constituents used to categorize a type of verdict. Consider for instance the examples below, from an on-line forum on a trial for sexual assault, a television interview with a prosecutor in a drunk-driving case, and an information flyer on trial skills seminars for defense attorneys: (19) a. . . . the jury should not hand down a verdict that says, “you are free to go.” b. You can’t come up with a verdict on I feel sorry for this guy. c. Real-life examples of techniques that have attained “Not guilty!” verdicts.
In these examples, the strings in italics do not constitute a descriptive means of characterizing a type of verdict, as in the alternative wordings “exculpatory verdict(s)” or “a verdict (based) on sympathy.” Rather, they are demonstrations of the jury’s joint fictive voice verbalizing the verdict’s content or motivation. In (19a), the main fictive interactional structure that seems to characterize the jury’s conceptualization of their verdict becomes manifest in the presentation of the verdict itself as doing the fictive talking. In the same way as the law-makers’ fictive voice may be heard when one speaks of the law as ‘saying’ this or that, the verdict is presented as taking the joint voice of the jury in telegraphic and straight-forward speech presenting the gist of the ‘message’ to the defendant. In (19b), the fictive speech of a juror standing for the whole group serves to metonymically refer to a potential string of thought or decision-making process, which may be expressed as such to fellow jurors, the court or the community. In (19c), the exclamation “Not Guilty!” is used to refer to exculpatory verdicts, which are expressed through the jury’s underlining of the words not-guilty in their jury form, subsequently read aloud by a court employee, and then agreed upon verbally by the jury. The exclamation mark seems to indicate that the string in italics is an exclamation of joy and thus needs to be ascribed to the defense team. In short, it seems common to conceptualize the outcome of a weighting process, such as voting or reaching a verdict, in conversational terms as those who come up with the decision fictively addressing the one(s) to be most affected by it. Such a conversation is understood as having been started by politicians in the case of voting or the defendant(s) in the case of coming up with a verdict. As a con-
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.25 (103)
Chapter 3. Fictive interaction blends
versation or discussion, it leaves space for overhearers (e.g. fellow citizens, victim’s family). This conceptualization does not only become manifest in speakers’ lexical choices (e.g. “message” for “vote” or “verdict”). It is also reflected in their use of direct speech – in different grammatical positions – in order to demonstrate the content of this message and/or the motivation behind it, as well as to characterize the type of verdict, for instance.
Fictive interaction as a fundamental cognitive process In the previous three sections I discussed instances of fictive interaction blends in ordinary and legal settings. These involved the conceptualization and presentation of: (i) serial monologues as simultaneous trialogues between the speakers and the audience; (ii) the source of an inference as directly speaking to those who draw the inference or learn about it; and (iii) the result of a deliberation on some issue as an act of communication. I postulate that the use of the conversation frame in general as well as the subframe of the fictive trialogue in ‘display talk’ (Goffman 1981) and apostrophe (cf. Richardson 2002) in particular is not restricted to these cases. Quite differently, I argue that it constitutes a fundamental structure of ordinary argumentation as well as thought, language, and discourse. Take for instance the following example of apostrophe from the same jury deliberation on which subsection 5.2. was based: (20) Juror 2: Did you wet yourself at that point? Juror 7: That’s what I said when you said that. Juror 11: Oh, my Lord, have mercy. Keep that conversation down. (JurDel-B, p. 211)
Instead of directly telling juror 2 and 7 that they should ‘keep their conversation down,’ in (20), a juror addresses a deity begging Him to make sure this happens. Regardless the jurors’ religious beliefs, the juror’s address is a fictive one, since it is not meant for God to hear, but for the participants in the actual situation of communication, who are thus turned into fictive overhearers. A similar structure can be used by fictively addressing the individual talked about directly, as in the following extracts from the same jury deliberation: (21) Juror 2: It’s the defense fault, but that’s why it cannot come down [on] Mark [defendant], is because the defense did not do their part. Juror 1: He chose his lawyers. He chose them. [. . .] He chose his lawyers! [. . .] Juror 12: If you don’t like the way your lawyer is doing it, then fire him and get another one. Juror 2: Okay. Juror 12: Fire him and get another one. (JurDel-B, p. 170)
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.26 (104)
Esther Pascual
In this example, as in (20), the utterer of the strings in italics steps out of the ongoing situation of communication in the here and now in order to communicate something to her fellow interlocutors. Juror 12 fictively addresses the individual that the conversation is about, namely the defendant, whom jurors had never addressed and would most probably never address. By fictively telling the defendant that if he does not like his lawyers he should fire them, juror 12 is letting his fellow jurors know that the defendant had a choice to stay with his defense team. Consider also the following example from the same deliberation: (22) Juror 7: The Doctor says. . . The Doctor says that he is a pathological liar. That’s what the Doctor says. The Doctor says he is competent, so it balances itselfJuror 1: You got to weigh it. Juror 7: It cancels itself out. [. . .] It cancels itself out. [. . .] He says on one hand he’s okay, and on the other hand he is not okay. Well pick a side Doc. Is he all right, or is he not all right. (JurDel-B, p. 91)
In (22), juror 7 fictively asks the (now absent) expert witness talked about to take a stand on a particular issue. By so doing, this juror can make it clear to fellow jurors that the testimony of that witness was not consistent and should probably be disregarded, without having to explicitly express it. Fictively addressing an individual in order to say something about them is not restricted to argumentation. Take for instance the examples below, from my interview to a novelist who attended a murder trial, a juror’s comment to the press in the Michael Jackson case, and an interview with a lawyer: (23) a.
I took one look at him [defendant] and I thought [takes breath] “Oh! I don’t like you at all!” (Int.8-Nov.: 35-36) b. I disliked it intensely when she [witness] snapped her fingers at us [...] I thought, “Don’t snap your fingers at me, lady.” (New York Times, June 14, 2005) c. I want the jury to get annoyed at the other guy. . . Hopefully the jury is figuring: “Shut up. We wanna hear what he has to say. Sit down, you nerd!” (Walter 1988: 80)
These examples illustrate what could be called thinking as speaking. Thinking something about somebody or about their behavior is presented as directly speaking one’s thoughts to that individual. Note that in the situations described the utterers could never speak to the individuals talked about. A trial attendee cannot address the defendant on the stand and the jury cannot talk to witnesses or attorneys in court. This notwithstanding, what they thought about them or about their behavior is still presented through direct speech as words fictively directed to them. It is through this fictive address that the conversational participant(s) in the actual
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.27 (105)
Chapter 3. Fictive interaction blends
situation of communication can learn what the utterer thought of the individual talked about. Since a relation is assumed between language and conceptualization, this seems to suggest that thought as such is interactionally structured.
Conclusions In this paper I suggested that fictive interaction constitutes the unifying structure of well-known blends, previously analyzed separately, namely Fauconnier and Turner’s “Debate With Kant,” Turner’s “The Dream of the Rood,” and Coulson and Oakley’s “Voting as Speaking.” Thus, the study of basic blending types, rather than anecdotal tokens, may help us understand the operations underlying their particular instances, as well as shed light on the reasons why some blends are more successful than others. Analysis suggested that the novel construals that arise in fictive interaction networks – as well as in any other conceptual configurations – emerge from situational constraints and are rooted in extant frames and cultural models – such as the cultural model of speech as informational (Sweetser 1987) – as well as overall knowledge of the context. I further proposed that the courtroom examples dealt with reflect the conceptualization of the different trial phases (i.e. the presentation of evidence, the argumentation upon it, and its evaluation) in terms of overt and covert verbal exchanges between the main trial participants showing an underlying triadic structure. Such a conceptualization becomes manifest at the levels of the discourse structure and content, the sentence, the clause, the phrase and the lexical item. The strategic use of fictive interaction as in the examples discussed seems to be motivated by the importance of language in Western courts as well as its strict interactional structure and adversarial system (Pascual 2006a, forth.). Thus, legal argumentation involves the conceptual construction of fictive realities different from the objective facts, which still determines the outcome, specially when the stakes are particularly high (Coulson & Pascual 2006; Pascual in press). I further conclude that the existence of the fictive interaction blending type suggests that if ordinary interaction can serve as a frame to a common blend, consideration of the interactional context should enlighten understanding of the overall network. More generally, I postulate that the conversation frame in general as well as the subframe of the fictive trialogue in ‘display talk’ and apostrophe in particular constitute fundamental structures of thought, language, and discourse. In sum, I argue that we not only rely on our direct bodily experience for structuring our mental world (Lakoff & Johnson 1980, 1999; Johnson 1987; Sweetser 1990; Talmy 2000), but also on our social experience as individuals constantly exposed to and engaged in situated verbal exchanges with fellow speakers.
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.28 (106)
Esther Pascual
References Adelswärd, V., Aronsson, K., Jonsson, L., and Linell, P. 1987. The unequal distribution of interactional space: Dominance and control in courtroom interaction. Text 7 (4): 313–346. Atkinson, J. M. and Drew, P. 1979. Order in Court: The Organisation of Verbal Interaction in Judicial Settings. London: Macmillan. Bakhtin, M. M. 1981 [1975]. The Dialogic Imagination, M. Holquist (ed.), C. Emerson and M. Holquist (trans.). Austin: University of Texas Press. Baynham, M. 1996. Direct speech: What’s it doing in non-narrative discourse. Journal of Pragmatics 25 (1): 61–81. Brandt, L. This volume. A semiotic approach to fictive interaction as a representational strategy in communicative meaning construction. Cotterill, J. 2003. Language and Power in Court: A Linguistic Analysis of the O. J. Simpson Trial. Houndmills: Palgrave Macmillan. Coulson, S. 2001. Semantic Leaps: Frame-Shifting and Conceptual Blending in Meaning Construction. Cambridge: Cambridge University Press. Coulson, S. and Oakley, T. 2006. Purple persuasion: Conceptual Blending and deliberative rhetoric. In Cognitive Linguistics: Investigations across Languages, Fields, and Philosophical Boundaries, J. Luchjenbroers (ed.), 47–65. Amsterdam/Philadelphia: John Benjamins. Coulson, S. and Pascual, E. 2006. For the sake of argument: Mourning the unborn and reviving the dead through conceptual blending. Annual Review of Cognitive Linguistics 4: 153–181. Fauconnier, G. and Turner, M. 1996. Blending as a central process of grammar. In Conceptual Structure, Discourse and Language, Goldberg, A. (ed.), 113–130. Stanford: CSLI. Fauconnier, G. and Turner, M. 1998. Conceptual integration networks. Cognitive Science 2 (1): 133–187. Fauconnier, G. and Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Goffman, E. 1981. Footing. In Forms of Talk, 124–159. Oxford: Basil Blackwell. Haiman, J. 1989. Alienation in grammar. Studies in Language 13 (1): 129–170. Harré, R. 1985. Persuasion and manipulation. In Discourse and Communication, T. van Dijk (ed.), 126–142. Berlin: Walter de Gruyter. Herman, V. 1999. Deictic projection and conceptual blending in epistolarity. Poetics today 20 (3): 523–541. Johnson, M. 1987. The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason. Chicago: University of Chicago Press. Lakoff, G. 2004. Don’t Think of an Elephant: Know Your Values and Frame the Debate. Chelsea Green. White River Junction: Chelsea Green. Lakoff, G. 2006. Thinking Points: Communicating Our American Values and Vision. New York: Farrar, Strauss, Giroux. Lakoff, G. and Johnson, M. 1980. Metaphors We Live By. Chicago: University of Chicago Press. Lakoff, G. and Johnson, M. 1999. Philosophy in the Flesh: The Embodied Mind and its Challenge to Western Thought. New York: Basic Books. Langacker, R.W. 1999. Virtual reality. Studies in Linguistic Sciences 29 (2): 77–103. Latour, B. 1987. Science in Action: How to Follow Scientists and Engineers Through Society. Cambridge: Harvard University Press. Oakley, T. and Coulson, S. This volume. Connecting the dots: Mental spaces and metaphoric language in discourse.
TSL[v.20020404] Prn:25/01/2008; 8:05
F: PB17003.tex / p.29 (107)
Chapter 3. Fictive interaction blends
Panther, K.-U. and Thornburg, L. L. 2000. The EFFECT FOR CAUSE Metonymy in English Grammar. In Metaphor and Metonymy at the Crossroads, A. Barcelona (ed.), 215–232. Berlin: Mouton de Gruyter. Pascual, E. 2002. Imaginary Trialogues: Conceptual Blending and Fictive Interaction in Criminal Courts. Utrecht: LOT. Pascual, E. 2006a. Questions in legal monologues: Fictive interaction as argumentative strategy in a murder trial. Text & Talk 26 (3): 383–402. Pascual, E. 2006b. Fictive interaction within the sentence: A communicative type of fictivity in grammar. Cognitive Linguistics 17 (2): 245–267. Pascual, E. In press. “I was in that room!”: Conceptual integration of content and context in a writer’s vs. a prosecutor’s description of a murder. In New Directions in Cognitive Linguistics, Evans, V. and Pourcel, S. (eds.). Amsterdam/Philadelphia: John Benjamins. Pascual, E. Forth. Text for context, trial for trialogue: A fieldwork study of a fictive interaction metaphor. In Cognition in Court: Cognitive Linguistic Approaches to Law and Litigation, Aldrige, M., J. Luchjenbroers and E. Pascual (eds.). Berlin and New York: Mouton de Gruyter. Quintilian. 1921. Institutio Oratoria. [Translated by Butler, H.E.] 4 vols. Cambridge: Harvard University Press, IV, 1, 28. Richardson, A. 2002. Apostrophe in life and in romantic art: Everyday discourse, overhearing, and poetic address. Style 36 (3): 363–385. Sapir, E. [1949] 1986. Selected Writings in Language, Culture, and Personality. D.G. Mandelbaum (ed.). Berkeley: University of California Press. Stygall, G. 1994. Trial Language: Differential Discourse Processing and Discursive Formation. Amsterdam/Philadelphia: John Benjamins. Sweetser, E. 1987. The definition of lie: An examination of the folk theories underlying a semantic prototype. In Cultural Models in Language and Thought, Holland, D. and Quinn, N. (eds.), 43–66. University of Chicago Press, Chicago. Sweetser, E. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. Talmy, L. [1996] 2000. Fictive motion in language and ‘ception.’ In Toward a Cognitive Semantics: Concept Structuring Systems. Vol. I, 99–175. Cambridge: MIT Press. Tannen, D. 1986. Introducing constructed dialogue in Greek and American conversational and literary narratives. In Direct and Indirect Speech, 311–322, Coulmas, F. (ed.). Berlin: Mouton de Gruyter. Tannen, D. 1989. “Oh talking voice that is so sweet”: Constructing dialogue in conversation. In Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse, 98–133. Cambridge: Cambridge University Press. Tobin, V. 2006. Joint attention in reading communities: Overhearer effects, routinization, and reading Sherlock Holmes. Paper presented at the Conceptual Structure, Discourse and Language Conference. University of California, San Diego, November 3–5, 2006. Turner, M. 2002. The cognitive study of art, language, and literature. Poetics Today 23 (1): 9–20. Walter, B. 1988. The Jury Summation as Speech Genre: An Ethnographic Study of What it Means to those Who Use It. Amsterdam/Philadelphia: John Benjamins.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.1 (109)
chapter
A semiotic approach to fictive interaction as a representational strategy in communicative meaning construction Line Brandt University of Aarhus
This chapter examines a phenomenon characterized in Pascual 2002 as fictive interaction (FI): signifying states and attitudes by the enactment of an invented verbal interaction. The phenomenon is examined as a representational strategy in meaning construction at two levels of linguistic manifestation: discourse and syntax. Data include examples of utterances that function metonymically at the syntactic level, as well as four examples of FI in discourse, among them Fauconnier & Turner’s “Debate with Kant” example. The four discourse examples are claimed to each exemplify one of four general types of semiotic blends. While I argue that analysis in terms of semiotic integration of mental spaces can account for examples at the discourse level, I propose to view examples at the syntactic level in terms of an invented communicative scenario evoked by an embedded dramatized enunciation which can be characterized as either generic or ascribed to a particular person. Keywords: fictive interaction, meaning construction, enunciation, conceptual integration, mental spaces, blending, semiotic base space, representation, virtuality, dramatization, construal, fictive motion, intersubjectivity, pragmatics
Introduction The subject of this chapter is a phenomenon whose existence provides an interesting case for our understanding of representation in communicative cognition, a phenomenon called fictive interaction. ‘Fictive verbal interaction’ is a term proposed by Pascual in characterizing this much overlooked phenomenon in linguistics: fictive pieces of conversation “used metonymically to stand for whole mental and emotional states, activities, and events” (Pascual 2002: 16). Fictive interaction is typically manifested in the form of an imagined utterance whose non-actuality is signaled grammatically and accompanied, in spoken language, by intonational
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.2 (110)
Line Brandt
shifts and other expressive-gestural indices. As a linguistic entity, such an imagined utterance “constitutes a self-sufficient discourse unit conceptualized within a nonfactive communicative occurrence, which functions syntactically and semantically as a grammatical constituent” (Pascual 2006a: 262). Speakers employing this representational strategy use the situational frame of verbal interaction to gain mental access to these represented entities, entities which need not themselves belong to the experiential domain of communicative interaction, and which are not necessarily related to communication at all. The phenomenon thus exemplifies “a specific case of departure from the direct reference to or description of actuality” (p. 2). The pervasiveness of the phenomenon, its conventionalization, and the ease with which we produce and comprehend instances of it, further provide indication that our experience with face-to-face interactions with others affects “our conceptualization of experience as well as the internal architecture of language structure and use”, as noted by Pascual in her book (p. 16). This chapter presents a framework for distinguishing different kinds of fictive interaction in online meaning construction. Meaning, within cognitive semiotics, is taken to refer to the signified (signifié) side of signs occurring in communication and other expressive practices, and ‘construction’ is taken to be a mental endeavor engaging multiple minds, as the exchange of signs (semiosis) is essentially an intersubjective enterprise. The study of semiosis (semiotics) hence concerns the ways in which minds construct and share mental content. Two overall uses of fictive interaction are outlined: semiotic blends and embedded metonymic enunciation. One type makes use of conceptual blending in the construction of a sign, and the other embeds a fictive enunciation in the base enunciation. These two rhetorical strategies afford a number of possibilities for making the discourse more engaging to the participants. This cognitive, and pragmatic, reward, it is argued, is due not only to the shift from abstract thought content to representations at “human scale” (cf. Fauconnier & Turner 2002) but to how we experience representations when they are dramatized in our mental simulation, that is, when particular instantiations of concepts are acted out theatrically on the stage of our inner vision. Fictive interaction is thus suggested to utilize the cognitive affordance of dramatization as an important representational strategy. As Pascual points out, these are instanced of “concrete subjective (re)enactment” rather than “abstract objective description”. (Pascual 2006a: 261) The cognitive motivation for using this representational strategy has to do with the experience of the communication itself: “The discoursive power of this usage should not be underestimated. The use of direct speech serves to (re)create a staged verbal performance in the current interaction as though it were occurring at the time of speech. Thereby, a story becomes more vivid [. . .]; it constructs a sense of immediacy [. . .]; and it attains conversational involvement [. . .].” (p. 261).
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.3 (111)
Chapter 4. A semiotic approach to fictive interaction
In extension of Pascual’s syntactic categorization of fictive interaction specifiers (Ch. 5 in Pascual 2002; see also Pascual 2006a, and 2007 [in prep.], cf. for instance cases where an embedded utterance modifies the head noun, e.g. “‘I do!” Dishes’ as a brand name for wedding china), I examine the enunciative aspect of this functional category, and examples are given of two different subtypes of the attested phenomenon, characterized here as embedded metonymic enunciation. In Pascual’s analysis, all examples are described as examples of “fictive interaction”. The distinction I make is based on differences in enunciation. One subtype is characterized by the fictive enunciation of a specific enunciator, believed to exist. The other has a generic enunciation; the enunciator is represented generically. Though Pascual’s approach to fictive verbal interaction is generally compatible with the semiotic approach, another difference, which is as much methodological as it is ontological, has to do with the extent to which conceptual integration theory and the notion of mental spaces are thought useful for the analysis of the phenomenon (cp. Pascual 2002). As can be surmised from my analysis of embedded metonymic enunciation, and also from the distinction I make between “embedded metonymic enunciation” and “semiotic blends”, I do not think all fictive verbal interaction phenomena benefit equally from being represented in terms of “conceptual blending” of mental spaces. Since I do not believe the blending framework renders the phenomenon of metonymically represented fictive interaction more intelligible, I do not apply it in my analysis of this type of fictive verbal interaction. However, as I will demonstrate, I do find it useful, both in giving a cognitively realistic account of, and diagrammatically representing, fictive interaction in the case of “semiotic blends” (semiotic integrations of conceptual content in what I call “virtual (mental) spaces”). Taking as a starting point the case of the Debate with Kant, an example given in Fauconnier & Turner 2002 to illustrate the fictivity involved in certain complex blends (featuring a contemporary teacher fictively debating Kant to illustrate his own philosophical point), further examples are given of blended integrations of fictive interaction in complex sign structures occurring in situated communication. The generalizations from the proposed semantic analysis of the Debate with Kant example are extended to the rest of the examples (borrowed from the Michigan Corpus of Academic Spoken English, and Google), to illustrate the similarities between these four pieces of discourse and to compare their rhetorical functions, suggesting four subtypes of semiotic blends utilizing the cognitive resource of representing fictive interaction. This perspective offers a method for analyzing semiotic blends, integrating the pragmatic and semantic dimensions of language in cognitive analysis of natural language phenomena, as Fauconnier set out to do when mental space theory was first launched as a cognitive paradigm in the philosophy of language (cf. Fauconnier 1994 [1985] and 1997), thus shaping the preliminary stages of conceptual integration theory (cf. Fauconnier 1997).
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.4 (112)
Line Brandt
The notion of mental spaces is re-examined from a phenomenological viewpoint and applied to diagrammatic analyses of sign integrations in argumentational discourse. This pragma-semantic, context- and relevance-oriented style of analysis adds a semiotic dimension to blending theory which is intended as a way to advance the theory, specifically to address the pragmatic aspect of linguistic meaning and to aid in distinguishing semiotic blends from other types of conceptual integration.
Semantics and pragmatics The first appearance of the Debate with Kant blend was in Fauconnier & Turner (1995),1 an example that is revisited at the beginning of Chapter 4 of The Way We Think (2002). One important thing that this example served to illustrate was that it is possible, and indeed common, for us to construct online fictions – not for their entertainment value, as one might think, but to create inferences about some real-world circumstance. As Fauconnier showed in his second book on mental spaces, there are space constructions manifested in natural language that “are not intended to serve in direct matchups with the real world but can nevertheless yield important real-world inferences.” (Fauconnier 1997: 69) One such kind of fiction is the counterfactual, common in everyday thinking and discourse, and especially prevalent in historical discourse, since the writing of history consists not only in depicting what happened, but also, significantly, what did not happen. Counterfactuals, as Fauconnier writes in his 1996 paper on the same subject,2 are understood via the construction of appropriate mental spaces and projection of structure from one domain [read: mental space] to another: from an alternative situation to an actual situation. One example is the following statement, which entails imagining a fictive identity relation: If I had been Reagan, I wouldn’t have sold arms to Iran.
The intended meaning of this utterance is that the speaker disapproves of the arms sale, which is not explicitly stated. The “alternative situation” described above can be said to be fictive, in that it is not vested with speaker belief. The speaker nonetheless manages to make a point – namely about the subject matter referred to in the
. M. Turner & G. Fauconnier 1995. ‘Conceptual Integration and Formal Expression’. Metaphor and Symbolic Activity, 10:3, 183–203. . G. Fauconnier 1996. ‘Analogical counterfactuals’. In Gilles Fauconnier, Eve Sweetser (Eds.), Spaces, Worlds, and Grammar. Chicago: University of Chicago Press, pp. 57–90.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.5 (113)
Chapter 4. A semiotic approach to fictive interaction
actual space (US selling arms to Iran). It would make little sense to exclude this entailment from a semantic analysis of the meaning of the utterance. Well, that’s pragmatics, some will say, and indeed it is – but the counterfactual story is pointless if we don’t know how to link it to the intended meaning. We need to know what the point is of constructing an alternative situation; what does the alternative situation have to say about the one we’re in? (p. 59)
It is clear from this example that only one of the spaces is the reference of the utterance, namely the “actual space” which we may call the Reference space, inasmuch as it represents the signified side of the counterfactual blend. The fiction (that the speaker be Reagan) is the counterfactual scenario by which the reference is presented. We will call this the Presentation space. The blend of the two is the counterfactual: a semiotic construction, yielding inferences about the reference topic.3 Though Fauconnier does not describe such blends (of a fictive and a non-fictive scenario) in terms of their status as signs, this view is perhaps precipitated by his intuition that meaning, undeniably a semiotic subject, should be described in terms of its manifested existence in communication and the kind of self-talk we call “reasoning”. In his view, which is still only rarely embraced outside the field of cognitive linguistics, a “shortcoming of modern work, found in this case both in linguistics and in philosophy, is the sharp emphasis on separating components (e.g., syntactic, semantic, pragmatic) and attempting to study the grammatical or meaning structure of expressions independently of their use in reasoning and communication.” (Fauconnier 1997: 5) As is generally known, there is a long tradition of avoiding context in linguistic analysis, since it is thought to only “complicate things” unnecessarily, but as Fauconnier observes: “When a sentence is examined in isolation, and its interpretations are studied, it is necessary to construct implicitly a discourse in which to interpret it.” (1997: 55) Even when the linguistic data is made up, and even when there is no explicit context in which to interpret the sentence as an utterance about something, one must imagine some minimal context for the sentence to mean anything at all. In this naturalistic view of language, sentences are regarded as utterances, or potential utterances to be manifested by a speaker in some situation of communication. Sentences bring together, in one linguistically homogenous form, heterogeneous and incomplete information as to the cognitive constructions to be . See L. Brandt & P. Aa. Brandt 2005a for a step-by-step description of blending theory in a semiotic framework. See L. Brandt & P. Aa. Brandt 2005b for an application of the semiotic blending framework to literary texts.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.6 (114)
Line Brandt
performed within a context for the purpose of constructing meaning. Meaning ensues when such operations are performed, but is not itself directly assignable to sentences. (Fauconnier 1994: xx)
Sentences are thus not surface-structure/deep-structure pairings with an inherent semantic content to be interpreted and analyzed independently of any pragmatic circumstances, as generative grammar would have it. Nor is the meaning of a sentence determined by its truth-value in a proposition/state of affairs pairing, as proposed in analytic philosophy. Rather they are prompts that minds use to facilitate the construction of representations in inter- and intrapersonal cognition. In this view, then, semantics and pragmatics are not only intrinsically related but pragmatics is naturally ‘built into’ semantics; meanings are motivated by an intention in an addresser, a mind addressing itself or a 2nd person addressee, and this pragmatic relation is formally present in closed-class forms in language, notably in the pronominal forms of 1st, 2nd and 3rd person,4 and is also present at the functional level of syntax, for instance manifested as embedded utterances functioning as phrasal modifiers, as we shall see in the following. Think for instance of the 2005 Microsoft Office “New Era” advertising campaign. This campaign features advertisements which all include a punchline utterance proclaiming the beginning a new era in software. These sentences all have the same structure: “The [utterance] era is over.” One advertisement reads: “The OOPS I HIT REPLY ALL era is over”. What is interesting about the structure of these sentences is that the noun (era) is modified by an embedded utterance (“OOPS I HIT REPLY ALL”). In most cases compound nouns consist of two nouns, one modifying the other, but in cases such as these, the head noun is modified by a whole other sentence, conceptualized as an utterance. Phenomena such as these suggest a natural blurring of the boundary between the lexical and the pragmatic levels of language, since instead of a lexical class (nouns in this case), we find a structure above the level of lexemes and phrasal structures, namely the level of the sentence as utterance. This level, in turn, belongs to the realm of pragmatics. It thus seems that linguists and theorists within the field of syntax and of semantics who have made a virtue of keeping the different linguistic components methodologically separate, would benefit from rethinking their underlying philosophies of language and incorporating into their analyses the pragmatic aspects of sentence structure and sentence meaning.
. See E. Benveniste 1971.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.7 (115)
Chapter 4. A semiotic approach to fictive interaction
‘Fictive verbal interaction’ Counterfactuals are one kind of fiction, construed with the purpose of making a rhetorical point (rather than an aesthetic one). In the following we will focus specifically on discourse involving the construction of fictive verbal interaction, that is, language use exploiting the easily accessed conceptual frame (or Idealized Cognitive Model, if you will) of face-to-face interaction, as we know it from the ontological domain5 of communication, to signify something that in itself does not involve such face-to-face interaction. The object of interest is expressive representation of human interaction of the linguistic sort – as opposed to extra- or non-linguistic forms of interaction. It is therefore specified as verbal interaction, aka dialogue. The representation of verbal interaction is further specified as fictive; the verbal interactions in question are represented as fictive as opposed to real, which means they are not presented as conceptualizations of reality on the part of the enunciating cognizer; they are not vested with ‘speaker belief ’. Consider the following example: Imagine that a contemporary philosopher says, while leading a seminar, I claim that reason is a self-developing capacity. Kant disagrees with me on this point. He says its innate, but I answer that that’s begging the question, to which he counters, in Critique of Pure Reason, that only innate ideas have power. But I say to that, What about neuronal group selection? And he gives no answer. (Fauconnier & Turner 2002: 59).
As Fauconnier and Turner point out, this passage represents a situation in which a contemporary philosopher is carrying on a debate about the origin of reason with a philosopher who is long dead. What is more, the contemporary philosopher emerges as the victor in the debate, having argued Kant into silence. As fantastic as the scenario may seem from the perspective of accuracy and factuality, Fauconnier and Turner rightly maintain that such fanciful conceptual constructs – debating . The term ‘ontological domain’ is an adaptation of E. Sweetser’s idea of ‘semantic domains’, cf. Sweetser 1990, which are few and fixed in number as opposed to ‘experiential domains’ which are as numerous as the differing individual and context-dependent, historically and culturally determined framings of “what exists”. Ontological domains are based on cognitively universal distinctions between different kinds of phenomenal realities: physical reality versus social reality, for instance, or social reality (work, traffic regulations, etc.) versus the intimacy and ethics of the domain of face-to-face communication (the speech-act domain in Sweetser’s terminology). The question of what ontological (or ‘semantic’) domains exist is thus a question of natural ontology; a phenomenology of the world, as experienced by humans. On ‘semantic domains’ in this ontological sense, see also P. Aa. Brandt:‘The Architecture of Semantic Domains’ in Brandt 2004: 33–67.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.8 (116)
Line Brandt
with dead people – is a common practice in everyday life. Pascual offers additional evidence that these apparent impossibilities, which she calls fictive interaction, are indeed commonplace ways of thinking and communicating, and shows specifically how these fictive interactions operate in the context of courtroom communication (Pascual 2002, 2005, 2006b). Though it is common practice in philosophy to carry on dialogues – indeed the dialogic form is one of the defining features of philosophical discourse – these exchanges often take place over vast time spans, ranging beyond individual lifetimes, even centuries, thanks to the possibility of written discourse. What is peculiar about the passage above is that the exchange is construed as real-time dialogue, affording the possibility of responding and countering, as if it were a discussion to be settled on the spot. As in the Microsoft Office advertising example above (“The OOPS I HIT REPLY ALL era is over”), a fictive verbal interaction is constructed, only in the Debate with Kant example the represented fictive exchange refers to a non-fictive interaction, namely the (more abstract) philosophical dialogue which is naturally less personal and less geared towards immediate outcomes. The phenomenon examined here is thus constructed fictive dialogue understood as “a conceptual channel of communication underlying the observable interaction between participants” (Pascual, personal communication), employed for expressive purposes to signify non-fictive referents. It is interesting that it makes sense to engage in these fictions, that we make sense by engaging each other’s minds in fictions, not with the purpose of enjoying the fictions as such, but to make mental contact with a referent in a way that is affectively engaging. Vivifying concepts by having them be played out in the imagination, animating them, rather than just having them be pointed to as meanings abstracted from experience makes these discourse referents more present experientially which appears to increase the rhetorical effect. This is a general point about cognition from a cognitive-phenomenological perspective, though for the present purposes we will focus on conceptualizations involving the representation of fictive interaction, as outlined above, analyzing and giving a tentative typology of its different manifestations, followed by a discussion of the theoretical implications for mental space theory in a pragmatically oriented style of research. I delineate two overall kinds of fictive interaction in meaning construction: fictive interaction metonymies and fictive interaction blends. These, in turn, can be divided into subtypes. I will describe a possible typology, based on examples from Pascual 2002, MICASE (Michigan Corpus of Academic Spoken English), Google and Fauconnier & Turner 2002. The two overall types of fictive interaction: 1) Fictive interaction metonymies 2) Fictive interaction blends
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.9 (117)
Chapter 4. A semiotic approach to fictive interaction
Metonymic fictive interaction Consider the following quote by financial economist Mark Skousen: When Congress passes a minimum wage law, they are essentially giving up on the poor. They are saying, “We don’t believe you are capable of making your efforts to be paid a decent wage”.
The quote embedded in Skousen’s enunciation is not to be understood as a literal quote. Congress is not literally “saying” anything; it is a fictive quote. This fictive enunciation stands metonymically for an attitude, namely the attitude that according to Skousen led Congress to pass a particular law. This is an example of fictive enunciation. The utterance does not purport to be a true quote. It is not framed as an accurate, or even approximate, rendition of an actual utterance; hence ‘fictive’. This fictive enunciation is embedded; the quoted utterance belongs to the fictively enunciating subject – Congress – and not to the ‘here & now’ base enunciation of Skousen himself. The embedded quote represents the fictively enunciating subject’s real attitude or belief (as framed by Skousen as he is the author of the editorial viewpoint6 ); the expressed attitude is the metonymic referent of the fictive enunciation. Skousen does not believe Congress ever said ”We don’t believe you [:the poor] are capable of making your efforts to be paid a decent wage”; but he does believe Congress has the kind of attitude toward the poor that this utterance expresses. The embedded utterance is a sign standing metonymically for an attitude or belief. One could speculate that the metonymic link between ‘saying’ and ‘thinking’ is grounded in a general experiential correlation between attitudes and beliefs and the expression of these states of mind.7 The fictive enunciation is ascribed to a particular subject in this example. The enunciating subject is specific; the enunciation is personal. When the fictive enun. I propose the term editorial viewpoint here, to clarify the difference between framings that are embedded and framings that belong to the matrix enunciation. The editorial viewpoint is the framing viewpoint of the enunciator. Every text, and every cinematographic film for that matter, has an editorial viewpoint, to be distinguished from any embedded viewpoints, which, along with any other represented content, is framed by the editing authority – the author, the cinematographer, the speaker. . This correlation between states of mind and the expression of these states of mind can be extended metaphorically to a relation between affordances of inanimate objects – mindindependent states of affairs – and verbal expressions. This metaphoric connection is fairly common in some languages, French for instance, and is even grammaticalized in certain languages. Examples of this phenomenon can be found in Pascual 2006a, and 2006 (in prep.).
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.10 (118)
Line Brandt
ciation is presented as belonging to a particular subject – a person, a group of people, or, in this case, an institution – it can be described as embedded fictive personal enunciation, or simply embedded personal enunciation. Another kind of metonymic fictive interaction is characterized by embedded generic enunciation. By contrast to embedded personal enunciation which has people with actual existence as enunciators, embedded generic enunciation has generic speaker roles as their subjects of enunciation. Consider these examples: (1) The OOPS I HIT REPLY ALL era is over (Microsoft Office “New Era” advertising campaign, 2005) (2) I do! Dishes (Pascual 2002) (3) People here are so “What’s in it for me?” (heard in NYC)
In example (1), “OOPS I HIT REPLY ALL” refers to a state of affairs where the software does not sufficiently prohibit human errors in the office workplace. “Oops. . .” is the kind of outburst any office worker might utter (or think to himself) under the given circumstances. The utterance is generic. It serves to frame the concept of office mishaps in a particularly affect-laden way by seeing them from the perspective of someone who has just realized he has made a – possibly disastrous – error. The embedded fictive enunciation thus metonymically evokes the concept of a particular unfortunate state of affairs. “I do! Dishes” (cf. example (2)) are a brand of wedding china sold in the US. “I do!” is a central performative speech act in a standard wedding ceremony and thus evokes the concept of a wedding. Evoking the pathos of this significant part of the wedding – the romantic climax of the ceremony – is rhetorically powerful. This rhetorical effect is achieved by prompting the enactment of this critical speech act in one’s imagination. By framing the concept of a wedding in this manner – establishing mental access through a vividly imagined part of the ceremony – the mercantile enunciator creates the desired affective response to the product (dishes). In examples (1) and (2), the embedded utterances function as modifiers of a noun phrase in a compound construction. In example (3), the embedded utterance – “What’s in it for me?” – functions as an adjective, modified by the emphatic specifier “so”. This demonstrates, though does not exhaust, the syntactic variation associated with the phenomenon of embedded enunciation. As in examples (1) and (2), the embedded utterance in (3) has a generic enunciator. The utterance evokes a scenario where a subject – a person, group or institution – wants something from a 2nd person, who instead of responding
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.11 (119)
Chapter 4. A semiotic approach to fictive interaction
altruistically, responds from the perspective of self-interest. This imagined interaction, with its generic participants, vivifies the concept of a “New York attitude”, as framed by the editorial viewpoint (that of the speaker making the judgment). This is accomplished by the theatrical enactment of the enunciation “What’s in it for me?”, which by the already entrenched ‘saying for thinking’ metonymy described above comes to stand for a self-serving mindset. “What’s in it for me?”, with its associations to a specific kind of interaction, and with its dramatized performance in the base enunciation, is, or becomes, an easily recognizable emblem of the referent state of mind. At the pragmatic level, the emphatic so also functions as a space builder introducing the generic quote.8 In any instance of production of these and similar examples an enunciator will also typically dramatize the performance of the fictive quote by other means than strictly linguistic ones, signaling the embedding of another enunciation. Posture and facial expression along with a shift in intonation are common means of theatrically indicating an embedding. As Fauconnier notes, mental spaces are set up “not just by explicit space-builders, but by other more indirect grammatical means, and also by nonlinguistic pragmatic, cultural, and contextual factors” (Fauconnier 1994: xxxiv). In all the above examples, an embedded generic utterance stands metonymically for a referent belief state or state of affairs: it becomes a synecdochal emblem of the referent. Whereas the metonymic connection in embedded personal enunciation can be characterized as an experiential link between ‘saying’ and ‘thinking’, the metonymic link in embedded generic enunciation is synecdochal (synecdoche: part for whole): a dramatized utterance is conceptualized as a part standing emblematically for some whole (a situation/state of affairs). One could say, more generally, that both in the case of personal and generic enunciation, an embedded fictive enunciation functions as a framing device – to evoke a particularly engaging representational instantiation of a concept by dramatizing the referential access, i.e. the signifying sign presenting the intended referent. Metonymic semantization is clearly involved in these kinds of fictive in. “Like” and “all” have a similar space-building function. See Pascual 2005 on the function of “like” as a space builder staging enacted dialogues in discourse. An example of the use of “like”:“i got in this really bad argument with my friend the other day. i told her i was going out with super hunky popular guy and she was like no and i was like ya and she was like no and i was like ya and she was like no and i was like ya and she was like no and i was like ya and she was like no and i was ya and she was like no and i was like ya and she was like no and i was like ya and she was like no and i was like no and she was like no and i was like haha i got you! and she was like no and i was like ya and she was like no and i was like ya and she was like no and i was like ya and she was like no and i was like ya and i was like whatever.” (Google, Posted September 01, 2005 08:31 PM)
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.12 (120)
Line Brandt
teraction phenomena, only, as shown above, there are slight variations as to what kind of metonymy is employed. Perhaps the distinction between personal and generic enunciation needs to be elaborated a bit in relation to the notion of fictivity. One could argue that embedded enunciation can be represented generically even in cases when the quoted utterance is not fictive. Consider the following example (Pascual 2002: 219): Interactant A: Oh, poor thing! here you are. . . Interactant B: Oh, don’t you dare [[poor thing] me]. This [dialogue] was produced by two colleagues and friends at work well past three in the morning. Interactant A is busy and is sorry to keep B awake, since they usually ride home together. She insisted that she is ok leaving alone, to which the friend reiterated that he really does not mind waiting for her. Finally, A expresses her feelings by exclaiming “poor thing!” referring to the faithful friend who is willing to stay up to keep her company. Again, this is a factual rather than fictive language production. B’s response, however, takes this same expression – originally intended as factual – and turns it into a creative fictive interaction verb. Incidentally, a grammaticalization, even one created on-line as the one in [this example] involves a level of generalization, which takes us one step away from particular language uses. (p. 220)
Not only is the utterance (“poor thing!”) not fictive in this example, the (embedded) interactants in B’s utterance are not the same particulars (A and B) represented in A’s utterance. As Pascual notes, at the end of this exposition, “the utterer of “poor thing” used as a verb is no longer A or B. It is a fictive interactant that stands for any individual who may utter the expression in any situation similar to the one occurring at the time of actual language production. The originally factual exclamative “poor thing” is used fictively in the next conversation turn as a verbal phrase to mean “to pity” or “feel sorry for” somebody.” (p. 220, my underlinings) Pascual characterizes this “any-ness” as fictivity (“It is a fictive interactant. . .”, “The . . . exclamative . . .. is used ficitively. . .”). I would suggest viewing it as genericity rather than fictivity, since the interactants are not represented as if they were real, which we would expect to be the case if they were represented fictively. Fictive entities are specific. They are specific, but are vested with non-belief (on the part of the conceptualizer/enunciator). Under this view, the (embedded) interactants in B’s utterance are represented generically rather than fictively. Pascual’s account points toward a more general feature of language, however, as there is an element of fictivity to genericity, if fictivity is taken to mean ‘existing in imagination only’ (as opposed to the more common notion of presenting a sign as referring to something fictive rather than something real). When one lets a token individual stand for a whole category (as in “A man needs to eat”) one does so without manifesting a belief in the existence of
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.13 (121)
Chapter 4. A semiotic approach to fictive interaction
any such particular individual, and so, in a sense, the individual signifying the category is fictive. This idea of fictivity is somewhat misleading, though, as it rests on an analytic-philosophical concern with What Really Exists and not on a semiotic concern with the ontological status of signs. The latter ontology resides in cognition and concerns mental spaces and mental constructs in a cultural rather than strictly physical reality. It seems to me a more fruitful approach to fictivity in language to view it from a cognitive standpoint rather than from a standpoint abstracted from and exterior to the communication. From the philosophical standpoint of embodied physicalism, the mainstream philosophy in cognitive linguistics today, any concept will be judged fictive that does not have a counterpart in physical reality. As a consequence, any reality that is not physical can be deemed fictive. Following this line of thinking, “fictive” becomes a synonym for “conceptual” – it now designates any mental content whatsoever. At the end of his paper on virtual reality, Langacker finds himself pondering the conundrum of the seeming ubiquity of fictivity, a conundrum which I believe can be ascribed to a philosophical confusion in the culturally oriented (especially the linguistic) cognitive sciences between mind-independent ontology, and its aspirations towards objective descriptions of physical reality, and the ontology of mental spaces as represented by conceptualizers. I quote: “At this point I have no definite idea of how far it is useful to push the notion of fictivity. [...] Should we go all the way and say that everything is fictive? Since our entire conceptual world is in some sense a mental construction, should we not just admit that the only kind of reality we have access to is VIRTUAL REALITY? I will leave that to philosophers.” (1999: 101) I think it is wise to leave the question of Reality an sich to philosophers and move the attentional focus, as Langacker proceeds to do in his concluding remarks, to the different ways in which fictivity is represented in language. Insofar as the object of study is linguistic cognition it seems to me a mistake to base determinations of fictivity/reality on extra-linguistic, mind-independent Olympic truths. From a cognitive perspective the only kind of fictivity that is relevant to semantic descriptions is represented fictivity: fictivity from a conceptualizer’s – or a group of communicating conceptualizers’ – point of view. Returning to the question of genericity, the – supposedly fictive – token individual is a representative of the (generic) referent, and is not itself a referent. It is important to note the difference between referring to something as a fiction (i.e. presenting something as a fiction) – whether this fiction is an end in itself (as in, for instance, literary ‘fiction’) or a predicative means to signify another referent (as in the Congress example above) – and imagining entities (and motion and change) in the construal of the representation of a given referent. Perhaps “construal” is not the right word for this phenomenon, since “construal” can also refer to a selective description (aka framing) of a state of affairs
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.14 (122)
Line Brandt
by syntactic means and invites judgments as to truth value or accuracy. It should be replaced by a term that relates specifically to the mental simulation of a representation. Perhaps it would be useful to distinguish between “simulational” and referential construal. Examples of referential construals would be selective descriptions – framings – of an action, event or state of affairs leading to certain evaluations; these construals can be judged to be accurate, or inaccurate, descriptions of a referent. They are truth-conditional. By contrast, mental simulations are not presented in a veridical mode. The properties of a mental simulation of a representation (“a mentalation”?) are not referential; the singurality of the entity standing for a category in the case of genericity, and the motion and change imagined in the case of “fictive motion”9 (cf. Talmy 1996, 2000: Ch. 2) and “subjective change” (Matsumoto 1996), aka “fictive”/“virtual”/“abstract”/“conceptual” change, are not ascribed to the referents of the representations. Generically represented singular entities stand for a plurality of entities. In Langacker’s terms, a virtual (or ‘arbitrary’) instance designates a category, and a category has an openended set of instantiations in actuality (1999: 96). In the case of subjective motion and change the referents are static states of affairs. Consider this example from the novel Everything is Illuminated by Jonathan Safran Foer: “[. . .] he had [. . .] a scar from his eye to his mouth, or his mouth to his eye. One or the other.” (p. 11) This depiction humorously points to the fact that the two construals are equivalent. The truth-value does not change, but for whatever intents and purposes one may be deemed more appropriate than the other. Langacker has a similar example in Langacker 2001 (p. 9): (4) a. An ugly scar extends from his elbow to his wrist. b. An ugly scar extends from his wrist to his elbow. c. An ugly scar {extends/goes/runs/reaches/streches} from his wrist to his elbow.
These examples illustrate the difference between the temporality of the reference and the temporality of the act of representing. Langacker makes a useful distinction between conceived time and processing time. Conceived time is referential: time as an object of conception. Processing time is time as a medium of conception; it has to do with the temporality of the mental simulation, which he calls “mental scanning”, and which is also sometimes described as the very act of “building up” a conception (cf. Langacker 1999: 84), much in line with what I am claiming here. Processing time is the time through which the conceptualization process . An example of fictive motion – from Dan Chaon’s collection of short stories Among the Missing: “We lived in a little house behind it, and behind our house was the junkyard, and beyond that were wheat fields, which ran all the way to a line of bluffs and barren hills, full of yucca and rattlesnakes.” (p. 50)
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.15 (123)
Chapter 4. A semiotic approach to fictive interaction
unfolds. In the three examples above, the state of affairs referred to is a static physical configuration, whereas the representation is construed sequentially; the mental simulation is dynamic. As Langacker explains: These expressions are truth-conditionally equivalent, describing precisely the same objective situation. Yet they clearly differ conceptually, and since the differences are determined by their form, they must be accepted as aspects of linguistic meaning. The contrast between (1a) and (1b) resides in the direction of mental scanning, i.e., the conceptualizer’s path of mental access in building up to a full conception of the overall configuration. [...] These various expressions construe the same situation in contrasting ways.10 (2001: 9–10)
A mental simulation is a dynamic process of representation and it makes little sense, from this perspective, to apply notions of reality/fictivity. The dynamic process of representation does not correspond to anything outside itself. Forcedynamic, figurative, temporally dynamic processing in representational imagination belongs to the realm of imagination but it is not fictive, if by “fictive” we mean “not represented as real”. It is not represented at all; it is the process of the representation coming into existence. To rephrase the above observation, then, we could say that: There is a difference between representing something as a fiction and imagining entities, and motion, and change, in the mental simulation of the representation of a given referent. I advise the reader to see Langacker’s exposition on the subject of genericity and virtuality in his article on Virtual Reality (Langacker 1999).11 For the present purposes, it will suffice to note, in conclusion, that embedded utterances can be either fictive or generic, while embedded interactants (embedded in the matrix enunciation) can be either specific or generic. In the Poor Thing example the embedded utterance is real, and the embedded interactants are generic.
. The omitted part of the quote ([...]) reads:“The alternatives in (1c) employ different verbs of motion for the metaphorical description of what is actually a static scene. What is crucial is that the directionality and metaphorical dynamicity are conceptually imposed on the situation rather than being inherent in it objectively.” I question Langacker’s interpretation of the description (employing different verbs of motion) as “metaphorical”. The mental simulation (aka the mental scanning) is not “imposed on the situation” – at least not in the referential sense. Metaphorical descriptions are referential – they directly concern the referent – and since the simulation – the mental scanning, with its temporal properties and its directionality – is not referential the description cannot be said to be metaphorical in any meaningful sense. . See also Langacker 1996 and 1997.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.16 (124)
Line Brandt
It follows from this that the above example is not an example of fictive interaction. It is interesting nonetheless. What is interesting about it is that it illustrates a cognitive propensity for representing concepts in terms of verbal interaction. Going back to the question of classifying different kinds of fictive interaction, some constructive revisions may be necessary. It may be to our advantage to begin seeing fictive interaction as a special (and pervasive) sub-type of the more general phenomenon of embedded enunciation or embedded interaction (Émile Benveniste, who first launched the linguistic notion of ‘enunciation’,12 always thought of it in terms of communicative interaction, so one nomenclature is as good as the other). To summarize the earlier findings in the light of this insight, the phenomenon of embedded metonymic enunciation can be subdivided into personal and generic enunciation. In metonymic personal enunciation, the utterance is fictive (as in the Congress example above), and the interactants specific. In metonymic generic enunciation, the utterance is generic – whether it is born generic (example (1)– (3)) or is “generized” from a real occurrence (cf. the Poor Thing example) – and the interactants are generic. There may very well be further subtypes which remain to be described in this framework. I will leave this to be sorted out by future research. In the following sections I give examples of semiotic blends involving the mental act of representing fictive interaction (henceforth: FI): that is, blends that utilize the semantic script of face-to-face dialogue to present a referent. Fictive interaction blends: – – – –
contrastive co-temporality blends virtual being blends contrast blends analogy blends
Contrastive co-temporality blends Recall the Debate with Kant blend described in The Way We Think: Imagine that a contemporary philosopher says, while leading a seminar, I claim that reason is a self-developing capacity. Kant disagrees with me on this point. He says its innate, but I answer that that’s begging the question, to which he counters, in Critique of Pure Reason, that only innate ideas have power. But I say to that, What about neuronal group selection? And he gives no answer. (Fauconnier & Turner 2002: 59) . Cp. “l’énonciation”, Problèmes de Linguistique Générale 1966, see Benveniste 1971.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.17 (125)
Chapter 4. A semiotic approach to fictive interaction
Fauconnier & Turner’s blending diagram (p. 62) illustrating the semantics of this piece of discourse consists of two input spaces: “Input 1” and “Input 2”, mappings from the inputs to a “generic” space connecting them, projections from the inputs to a “blend”, and a “frame” to complete the blend. The so-called “generic space” is a shared structure space which consists of a list of roles instantiated in the inputs: the presence of a “thinker”, “claims & musings”, a “mode of expression”, a “language”, an “issue”, a “purpose” and “time”. In the inputs these roles are filled by specific instantiations. In one input, we have Kant, his claims & musings, his mode of expression: writing, the German language, ‘reason’ as the issue, search for truth as the purpose, and the year of publication of his claims & musings: 1784. In Input 2, we have the speaker (the deictic “me”), his claims & musings, his mode of expression: speaking, the English language, ‘cognitive processes’ as the issue, the search for truth as the purpose, and the historical point in time where this contemporary philosopher’s musings take place: 1995. Aside from these elements which have counterparts in Input 1, the speaker has a concept of Kant as a philosopher who is no more but whose ideas live on in his writing. He is aware of Kant but, for natural reasons, Kant cannot be reciprocally aware of the speaker. The blend of these two inputs is framed by a “Debate frame” specifying what “rhetorical actions” and what language use can normally be expected in a debate (the use of “argumentational connectives, affirmation and negation”, e.g. “however”, “on the contrary”, “yes” and “no”). In the blend, the contemporary philosopher is debating with Kant as if he were right there, alive and up for a lively discussion. They are thus mutually aware of each other’s existence and each other’s arguments, and engage in a dialogue where claims and counterclaims are made. In the blend, they are discussing cognition, a subject of interest to Kant, though in the 18th century it was not thought of as “cognition”, and they are doing it in English. While it is true that the analyzed discourse is in English, my objection to including ‘language’ in the diagram is that the choice of language is not part of the meaning, if meaning is understood in the semiotic sense, that is: in terms of the communicated intention. This raises some general questions about the analysis and its theoretical foundation: Conceptual Integration Theory (CIT, aka blending theory), a theory which offers otherwise valuable contributions to understanding cultural products like art and language. Meaning in the semiotic sense, refers to what is intended by a sign. A sign could be a gesture or an utterance, or some other unit of discourse (words, phrases, paragraphs, texts etc.), a work of art; expressive acts motivated by the intentionality of the expressive subject (the “signer”) and his or her awareness of other subjects as intentional participants in communicative interaction. In Saussurian terms, the
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.18 (126)
Line Brandt
meaning is the signifié, the signified. In Peircian terms, it is the “object” that a sign refers to. Since the enunciator of the discourse analyzed does not intend to convey that the discourse is in English, the fact that it is in English is not part of the meaning. By the same logic, the Shared Structure space that goes under the name of “generic space” is superfluous in a diagram representing the meaning of the discourse. While it may be true that the inputs in the blending network have shared structure, these analytic generalizations are not part of the speaker’s intended meaning. Insofar as there is shared structure between the inputs, this structure is not represented separately, in a mental space, and it is not part of the meaning of the utterance. This argument can be extended to other cases where mental spaces are set up and blended in a blending network. This, in turn, raises the problem of space building: What determines what the input spaces are? Or put differently: When is there more than one mental space? At a practical level this concern amounts to the question of when differentiation (i.e. pluralization) of spaces is warranted in the diagramming of the meaning of a piece of discourse (or whatever expressive sign is the object of study). This question needs to be settled before analysts can rationally argue for one analysis over another. It is thus a methodologically central issue in conceptual integration theory, and a challenge to the CIT community, which needs to furnish the theory with clear and explicit criteria on which its practitioners can base their arguments for and against particular analyses, if they are to be counted as scientific arguments. For the purposes of the present task – to describe different types of fictive interaction blends – my suggestion would be to distinguish mental spaces that are ontologically differentiated, and spaces that are semiotically differentiated in a given conceptualization. Spaces that are semiotically differentiated occur in blending networks whose main characteristic is that one input space is about another input space. In such a semiotic network, the blend is always a blend of two spaces, since one space signifies the other. Ontological differentiation is a matter of the conceptualizer’s epistemic stance towards a scenario or fact. When an event or state of affairs is represented, it is inscribed by the conceptualizer as being the case in actuality, or as being hypothetical, or counterfactual, or it is inscribed as a desirable or undesirable scenario that has only imaginary existence, for instance, or it may be represented as being contained within a fiction. In mental space terms this amounts to the sensible intuition that there is a difference between real and fictive spaces (cp. Fauconnier’s notion of domain types, Fauconnier 1994, Ch. 1, and Fauconnier 1997: 138) and between factual and counterfactual spaces, etc. There is much more to be said on the matter; however, these distinctions will suffice for the present purposes.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.19 (127)
Chapter 4. A semiotic approach to fictive interaction
In Fauconnier & Turner’s analysis of the Debate with Kant blend the two input spaces are represented as separate spaces because Kant existed in the 18th century and his philosophical investigation (“critique”) of what he called “pure reason” was published in 1784, whereas the contemporary philosopher conducts his seminar and makes his statements in 1995. The differentiation between the spaces is thus grounded in time: the temporal separateness of ‘then’ and ‘now’. However, there are some deep philosophical problems with proposing time as a criterion for differentiation, as I see it. If mental space differentiation is established by temporal differentiation, what cognitive grounds do we have for determining the duration of a mental space and hence for ascertaining the onset of a new space? In other words, when exactly is it time for setting up a new space, and yet another, and so forth? How is it determined when a new moment has begun, and how do different conceptualizers coordinate these distinctions, so that they may communicate their ideas to one another? As noted elsewhere by Fauconnier communication is successful insofar as mental spaces are shared: To the extent that two of us build up similar space configurations from the same linguistic and pragmatic data, we may ’communicate’; communication is a possible corollary of the construction process. (1994: 2)
It seems to me it would be an impossible task for us to ever conceive of the same number of spaces, since there is no objective, intersubjective criteria for deciding when a new moment has arrived. It is not unlikely that we depend on a shared conception of time at the level of neural binding, but since we do not have conscious access to these processes, it is hard to see how such notions can feasibly figure in the diagrammatic modeling of semantics. If we are to have a method for validating mental space diagrams, and comparing different hypotheses as concerns specific analyses, we need a common ground for determining accuracy and explanatory power. Otherwise efforts towards falsification or argumentational support are rendered impracticable in any substantiated form. If the object of analysis is the meaning present in our consciousnesses as it manifests itself in interpersonal communicative interaction, then the methodology must, to a substantial degree, consist in “interspective” evidence (“interspective” is meant to emphasize the collaborative nature of the introspection), by which I mean that argumentational support, and challenges, are corroborated by comparative spection of the imaginal content of the minds contemplating these communicated meanings. (I am suggesting the term “imaginal” to denote a phenomenological property of representations, particularly: experienced semantic content as it plays out in the imagination (cf. Brandt & Brandt 2005b: 119).
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.20 (128)
Line Brandt
From a historical perspective, Kant and the contemporary philosopher exist in a continuum. They exist at different (durative) points in time in the continuum of the history of philosophy, spanning over centuries of conceptual evolution. Since these different times occur in the continuum of the history of philosophy, we would have no reason to represent the two philosophers separately in mental space diagrams, if what we were modeling was our conception of their placement in history. They would be represented in a single mental space. However, the object of the diagram is the semantics involved in the statements made by the contemporary philosopher in the cited example. The example is semiotic: it is a piece of discourse, which is why I propose a semio-pragmatic semantic analysis. Recalling the distinction between ontologically and semiotically motivated pluralization of mental spaces (ontological space building having to do with belief investment – e.g. representing a scenario as something that actually happened versus something that could happen or did not happen, or happened in a movie, and semiotic space building being motivated by someone’s expressive intention), it follows that semiotically motivated blends can only have two inputs: a signifier and a signified. The blend is the sign, consisting of a signifying mental space and a signified mental space, two constituent “inputs” which can also be described as a Presentation and a Reference (cf. Brandt & Brandt 2005a: 3.2 ‘Reference and presentation’: 227 sq.). In the present example, the speaker is presenting his own view on a topic (the origin of cognitive faculties) by contrasting it with that of Kant, as expressed in Kant’s treatise on the matter. From the pragmatic ground of the Semiotic Base space (where the communication takes place) he sets up a Reference space (his own point of view on the topic) and a Presentation space (Kant’s point of view) which “blend” in a complex sign (“complex” since the signified and the signifier are entire mental spaces), structured by the fictively imposed script of verbal interaction. I use the term “script” here rather than “frame”, since a script specifies what kinds of agents and actions can be expected in a specific type of situation whereas a frame is also more narrowly understood as a conceptualization of a given object of attention, whether it be a situation or merely an object or a relation or attribute. The resulting dialogue between Kant and the contemporary philosopher is fictive in the sense that the conceptualizers do not conceive of it as taking place outside of the representation: the Virtual space that is the blend. The blend is characterized by its contrastive predication and its peculiar temporal properties, allowing for the contemporary philosopher and Kant to co-exist in a purely chronological and ahistorical time – a time out of time, which is why I propose to call it a contrastive co-temporality blend. Time in the blend is not the historical time of the time-space continuum of the history of philosophy; for instance, the speaker does not intend for the students to infer that Kant’s counterclaim (“he
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.21 (129)
Chapter 4. A semiotic approach to fictive interaction
counters, in Critique of Pure Reason, that only innate ideas have power”) takes place in San Diego in 1995, or at any other specific point in the time-space continuum. The verbal exchange between the two philosophers happens in virtual time, which is sequential but not deictically grounded. Virtual time is constrained by the argumentational relevance in each instance of use. The Regatta blend described in The Way We Think (Fauconnier & Turner 2002: 63–65) is another example of a contrastive co-temporality blend. In the Regatta example, a past ship and a contemporary catamaran are contrastively blended in a virtual space structured by a race script: the two vehicles are racing each other in the blend. As in the Debate with Kant blend, the past and the contemporary agents co-exist in a Virtual space with ahistorical, sequential time, in this instance allowing for the conceptualizer to measure and compare the speed of the respective vehicles, in a dramatized, rhetorically engaging way. In the Debate with Kant example, the argumentational relevance is the dialogue script. Fauconnier & Turner’s description of the specific circumstances of the situation of enunciation is somewhat sparse but if we assume the topic of the seminar is cognition in the age of neuro-science, this is the mental space structure of the network. As described above, and as can be seen in the diagram below, the inputs to the blend are represented separately but for pragmatic reasons: to contrast two points of view, in favor of the one. The inputs are specified as a Presentation and a Reference: one space is a (predicative) sign for the other. The space building is pragmatically grounded in a Semiotic base space, consisting of the communicative “semiosis” (exchange of expressive signs) situated in a classroom setting (“a contemporary philosopher says, while leading a seminar: “I claim. . .””). In Semiotic space, the 1st person enunciator is addressing a group of students. He is making a case for “neuronal group selection” in a discussion on cognition in the age of neuro-science. The situational context framing the semiosis, and preceding discourse framing current interactions, is what I have called Situational relevance (cf. Brandt & Brandt 2005a for an introduction to the notions of “situational”, “argumentational” and “illocutional” relevance). The speaker sets up a Reference space with his own point of view (reason is a self-developing capacity) and a Presentation space with a contrasting point of view (reason is innate) which puts his own into perspective. In the diagram, I have divided Situational relevance into two (inter-related) aspects of situatedness. The Situational relevance motivating the so-called Elaboration loop is the discourse topic: cognition in the age of neuro-science. Situational relevance puts topical constraints on what is relevant for the participants (in the exchange of viewpoints) to say and do. It would seem odd or irrelevant, for instance, if the participants began discussing the weather in Königsberg, and less attention is paid to the social aspect of interaction than would be the case in other contexts.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.22 (130)
Line Brandt
Situational relevance (philosophical argument) Presentation space
Phenoworld Semiotic space
Contemporary philosopher ’s PoV (reason is a selfdeveloping capacity...) Topic: The origin of reason (cognition)
Kant’s PoV Critique of Pure Reason (reason is innate...) Topic: The origin of reason
I claim...
Situation
Reference space
Situational relevance (cognition in the age of neuro-science)
Argumentational relevance Script: argumentational discourse in face-to-face interaction
Illocutional relevance (persuasion)
Pragmatic implication: Kant is argued into
Contemp. phil. PoV represented via contrastively foregrounded PoV (that of Kant)
Virtual space
Virtual space: Contrastive coMeaning temporality blend: Virtual interaction: Dramatized representation of dialogue as face-toface argument
silence – ”I win!”
Figure 1.
Another aspect of Situational relevance, related to the former, is the contextual motivation behind the choice of styles of discourse engaged in. The situation of enunciation determines what styles are appropriate. Using a metaphor from literature, we could say that the styles, in turn, are determined by the relevant discourse genres. The relevant discourse genre in our example is philosophy, and this shared assumption in the Semiotic space warrants the introduction of argumentational discourse. Contrasting different points of view is a commonplace strategy in structuring arguments. It serves the speaker’s cause in this context to blend two spaces instead of setting up just one: the view advocated as the reference, predicated by a foreign point of view (PoV). The Illocutional relevance (see Figure 1) is the speaker’s intention to persuade, which, in turn, puts constraints on what responses from the students may be deemed appropriate or not, insofar as the intent is recognized. The Debate with Kant blend, however exotic it may look at first glance, is an example of a fairly common propensity for structuring philosophical arguments dialogically. Philosophical discourse is inherently dialogic in form, whether
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.23 (131)
Chapter 4. A semiotic approach to fictive interaction
it occurs in writing or in verbal or imagined dialogue. In person, a philosopher will debate with himself, posing hypothetical questions and anticipated interjections from a 2nd person viewpoint, and of course, if others are present, he can debate them directly (as Socrates does in Plato’s dialogues). In writing, the practice of philosophy is not much different; the enunciator takes other viewpoints into consideration, debating with philosophers past and with (hypothetical) future interpolators. I would even go as far as to say: no dialogue, no philosophy. Because of the inherent dialogic structure in the practice of philosophy, the blend (aka the Virtual space) of the contemporary philosopher’s PoV represented via Kant’s (contrastive) PoV attracts the script of face-to-face dialogue, allowing the participants in the philosophical discussion to respond to each other as if they were debating each other in person. The structure (whether it be a script or a forcedynamic/topological schema) attracted by and projected to the blend is readily available in the phenomenal world of shared experience (shared by the cognitive community of humans, with its cultural or ethnic variations). Conceptual access to the Pheno-world is crucial. Cognizers need access to background knowledge to complete a blend, whether by implicitly drawing on a familiar script, or by setting up a Relevance space, working out the structure in attentional awareness. In order for a cognizer to set up a Virtual space, his mind would need to have a concept of as-if-ness, and to select the appropriate script and apply it to this as-if space (Virtual space), he would have to recognize its Argumentational relevance. The Argumentational relevance here is the script of argumentational discourse in face-to-face interaction. As described above, this script is analogous to the script of philosophical inquiry, and Kant’s voice in Critique of Pure Reason is transformed into a voice to be reasoned with, enacted by Kant’s contemporary opponent in Semiotic space. Argumentational relevance is a general name for this particular feature in semiotic blends, irrespective of discourse style and discourse genre: it is the structure that completes the blend, as an expressive sign of a semiotic intent. The script of face-to-face discussion is structured by a schema – a skeletal, non-figurative, dynamic structure with generic roles – which is integrated with the content in Virtual space, yielding a contrastive co-temporality blend. In the diagram, Virtual space figures twice; it is the same space but the second rendering of Virtual space (aka Meaning space) represents the integrated blend. The schema consists of a 1st person role and a 2nd person role and an object of attention role. The 1st person directs the 2nd person towards an object of (shared) attention (O). With each turntaking the 1st and 2nd person roles are switched, the 1st person becomes a 2nd person to the other interactant when the other interactant takes the floor and becomes a 1st person speaker. Both parties are keeping track of the force-dynamics of the arguments presented, and – ideally – come to the same conclusions as to which arguments prove strong enough to be left standing and which ones are weakened by counter-arguments. There is an impersonal
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.24 (132)
Line Brandt
evaluator role which all involved interactants strive to fill but which can normally only be filled by approximation, since we are mere mortals and not omniscient deities. It is thus an ideal position of knowing, a position from where it is evident what is the case and what is not the case, based on the arguments presented in the course of a discussion. This impersonal position from where judgments are passed can be compared to the role of the jury in criminal courts. Ideally the jury presides over all relevant facts and are able to tell with fatal certainty what versions of the truth are correct and hence who wins the trial. In actuality, criminal courts of course have to make do with juries who make qualified guesses as to what is more likely to be the case and base their judgments on this more modest foundation of knowledge. However approximative and fallible, the ambition to occupy the impersonal evaluator role is what makes a discussion a discussion. There are other forms of verbal interaction which do not require such a role, such as exchanging greetings for instance, but insofar as propositions are exchanged with the purpose of finding out what is the case, participants monitor the exchange from an impartial position exterior to themselves. The force in this force-dynamic battle of propositions is of course epistemic. The argumentational relevance of blending the two inputs in a Virtual space is to bring about a scripted dramatization of the reference using an (argumentative) dialogue script and its structuring schema of enunciation to conceptualize the contrastive points of view. In the integrated Virtual space (Meaning space), the philosophical dialogue across centuries is dramatized as a fictive verbal interaction. The interactants virtually co-exist in the blend; they are congregated in a virtual space where they are having a virtual discussion in the virtual time frame of face-to-face dialogue, and their mutual access to the semiosis allows them to jointly arrive at a conclusion: that Kant has not adequately described the origin of reason. In Semiotic space, the 1st person role in the schema of enunciation is filled by the teacher of the seminar. The students in the classroom fill the 2nd person role. The arrows signify directed attention. The crossed lines signify enunciational shifts. There is no record of student responses, so I have marked the (non-existent) enunciational shifts (turntakings) as dotted lines. The object of attention (O) is an embedded semiosis. Another way of diagramming the blending network in Figure 1 would therefore be to represent the blend (the integrated Virtual space) as a Semiotic space from where a Reference space is set up: the topical “Reason is. . .” space of the two actors in the fictive interaction. The turntakings in the embedded semiosis are marked by “I claim. . .”, “Kant disagrees. . . He says. . .”, “I answer. . .”, “he counters. . .”, “I say to that. . .”, “And he gives no answer.” Kant spends his last turn pondering in silence.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.25 (133)
Chapter 4. A semiotic approach to fictive interaction Judgment
Fillers in Base space
Roles
Face-to-face interaction ”I” want ”you” topay attention to (O)
Impersonal evaluator role (cp. jury)
Teacher
1. p.
Students
2. p.
Discourse
(Discourse:)
Fillers in Base space
Object of attention (O)
”I claim... ”, ”Kant disagrees...”,”He says... ”, ”I answer”, He counters... ”, ”I say to that...”, ”And he gives no answer. ”
Roles ”Reason is...”
Contemporary philosopher
1. p.
Kant
2. p.
Reason
Topic
”What ”Only about (Kant 1.p.) ”That’s innate ideas neuronal ”It’s begging the have group innate....” question...” power...” selection...”
Time
Figure 2.
The contrast between the differing points of view becomes all the more apparent when imagined as two parties explicitly negating each other’s statements. The bottom part of Figure 2 is an explication of the object of attention (O) in base space: the virtual dialogue between a 1st and a 2nd person, roles filled by the contemporary philosopher and Kant, respectively, on the topic of reason (cognition). In the blend, the 1st person role in the schema structuring the discussion script is filled first by the contemporary philosopher, then by Kant; the turn-taking is marked by interrupted lines indicating that the line representing the 2nd person becomes the line representing the 1st person, and vice versa. The line at the very bottom of the diagram indicates the passing of time. There are six turns in this embedded enunciation: The contemporary philosopher (CP) says “reason is a self-developing capacity”, then Kant says “it’s innate”, but CP says “that’s begging the question”, and Kant counters that “only innate ideas have power”, but CP says “What about neuronal group selection?”, and Kant’s final silence indicates he is at a loss, giving more weight to his opponent’s point of view. Except for the question “What about neuronal group selection?”, the embedded enunciation is narrated rather than acted out; the vantage point of the telling is in the Semiotic base space. The utterances are indirect quotations. The enunciation of these utterances is explicitly narrated from base space (“I claim. . .”, “he
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.26 (134)
Line Brandt
says. . .”), and therefore Kant is spoken of in the 3rd person rather than spoken to in the 2nd. The pragmatic implication emerging from the blend is projected back to the Semiotic base space: The contemporary philosopher wins the discussion, since Kant has no come-back to his winning argument. However anachronistic the argument may be, it counts in Virtual space, where historical time is suspended, and it counts in Semiotic space where, in force-dynamic terms, Kant’s virtual lack of resistance translates into a strengthened epistemic stance toward the proposed philosophy (cf. the Reference space) in the 2nd person subject (i.e. the students) – at least ideally; this is evidently the perlocutionary effect hoped for by the speaker. The students are invited to share the judgment arising from the teacher’s framing of the facts; as the audience they are expected to occupy the “jury position” along with the teacher himself (and Kant’s ghost – having succumbed to the persuasive powers of his virtual opponent!), and decide that the origin of reason is probably best accounted for by explanations other than innateness. The next example of a fictive interaction blend is an example of the second type of semiotic blend: virtual being blends. As do contrastive co-temporality blends, virtual being blends rely on an ability to represent a reference in a Virtual space, with emerging structure and (very real) pragmatic inferences, only the predicative relation between the Presentation and the Reference is slightly different. In contrastive co-existence blends the predicative relation between the inputs is contrastive. The purpose of the blend is to predicate something about the agentive entities in Reference space. In the Regatta blend, the property contrasted is speed. In the Debate with Kant blend the purpose is to ascertain how much credence to lend to the referent agent’s philosophical beliefs. The agents remain separate in the blend, whereas in virtual being blends they are fused. In virtual being blends the predication is one of virtual identity. The reference is presented in the blend as if the agents were identical. However, while the conceptualizers believe they are not numerically identical in reality, inferences emerge as to their qualitative identity in some specific respect, specified by what is argumentationally relevant. It is the rhetorical distinctiveness of the predication that motivates the division of blends into different types on this account.
Virtual being blends In virtual being (or virtual identity) blends, the Reference is presented as being the Presentation – in the signifying blend. The referential and presentational elements do not merely co-exist in Virtual space, and they are not just compared; they are identical. The Presentation virtually IS the Reference.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.27 (135)
Chapter 4. A semiotic approach to fictive interaction
Fictive interaction in virtual being blends is characterized by the construction of a Virtual space which blends a fictive thought content with a non-fictive one. A fictive verbal interaction, represented in a Presentation space, stands for a nonfictive referent, in Reference space. Consider the following example, an excerpt from a MICASE transcript. We can call this the Enacting Kant blend: Twenty students are gathered in a classroom for an end of term review session in philosophy. The person leading the recitation group, a female graduate student (S1), is trying to help the students understand Kant’s transcendental argument and categories of understanding. S1: [. . .] and the idea is, Kant says look. . . it’s an obvious fact that we have a unified consciousness by which he means, it’s obvious that Maureen’s thought states and beliefs and desires and mental states all kind of hang together, in a unified way inside of her in the same way that Matt’s kind of hang together, inside of him, and Matt’s perception of the board and my perception of the board are sort of in two separate unified consciousnesses, [. . .] [. . .] and he says look you know, sure there are several different possible explanations for this. [. . .] you’ve got Locke, trying to do it with memory, right? And you’ve got Hume trying to do it, and, even, explicitly saying out loud that he fails to do it with his causal connection theory. . . . you’ve got, um, Descartes trying to do it with soul theory. . . . . . and Kant says look i’m gonna offer you a new explanation, one, which involves the categories of understanding. . . and it’s the best explanation, that we have for a unified consciousness. (my italics)
In her didactic exposition, S1 creates a blend of (her framing of) Kant’s arguments (in Reference space) – and a fictive enunciation (in Presentation space); the referent point of view – “the idea” – is theatrically staged (cf. italicized utterances) in a presentational performance. In Reference space, Kant speaks through his writings. In Presentation space, Kant speaks through S1. In the blend, his written speech is signified by S1’s verbal enactment, in a dialogic face-to-face realm of interaction. Kant’s utterances are not intended to be quotes. Rather they are an expressive device by which the students are invited to imagine what he might say if he were making his case in person. There are shifts in vantage point in the course of the monologue, effecting shifts in the deictic manifestations of person. Not all utterances in the above quote have their deictic vantage point in a blend. Intermittently S1 also speaks as herself (but from Kant’s viewpoint, e.g. “it’s obvious that. . .”). I have underlined the embedded enunciation in italics. In mental space terms, the vantage point in S1’s enunciation switches back and forth between Virtual space and Semiotic space. From Semiotic space, Kant is referred to in the 3rd person (e.g. “and he says”). In S1’s explanation of the underlying premise for examining the unity of conscious-
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.28 (136)
Line Brandt
Situational relevance (presentation of philosophical stance)
Phenoworld Semiotic space Kant says ”Look...”
Presentation space
Reference space
Kant enacted by S1:
Kant speaks through his writings. Audience = readers Kant’s PoV – in dialogue with other – past or hypothetical – points of view
Dramatized enunciation S1 speaks to class as if she were Kant (Kant’s implied mindset)
Situation
Situational relevance (inquiry into the unity of consciousness)
Argumentational relevance:
Kant = (virtually IS) S1 Kant’s audience = S1’s audience
Virtual space
S1’s fictive speech represents Kant’s (eternal) views.
Script: dialogue. Epistemic affordances of dialogic exchange
Virtual being blend: Meaning Illocutional relevance (didactic instruction)
Pragmatic implication: Kant’s view is to be evaluated by students.
Dramatized representation of Kant in virtual face-to-face dialogue with contemporary thinkers. Kant advocates his views and is prepared to argue for them.
Figure 3.
ness, namely the existence of unified consciousnesses, Kant’s generic consciousness is instantiated by particulars (cf. the references to specific individuals: Maureen and Matt), and the speaking “I” is S1 herself (cf. “my perception of the board”). In the blend, Kant is the enunciating “I”, speaking to a general “you”, a 2nd person audience easily instantiated by the actual 2nd person audience in Semiotic space in running the blend: conceptualizers can take advantage of this affordance and imagine Kant as talking directly to Maureen and Matt and all the other students present in the classroom. In S1’s theatrical fiction, Kant is speaking in a laid-back, informal voice, and vies for his audience’s attention by initiating his expository remarks with the interpersonal pragmatic marker ‘look’, giving the impression of being immersed in an ongoing verbal exchange: “look, you know, sure there are. . .”, “look, I’m gonna offer you. . .”, “look. . . it’s an obvious fact that. . .”. These are embedded enunciations in the base enunciation in Semiotic space (see call-out bubble in Semiotic space in Figure 3). The Semiotic space is “scripted” by the situation in which it takes place: the “situational base” is the classroom situation where S1 is talking to her students, teaching them about Kant’s philosophy on unity of mind. The Sit-
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.29 (137)
Chapter 4. A semiotic approach to fictive interaction
uational relevance of the act of signification taking place is the presentation of a philosophical stance. S1 has set up a Reference space where Kant is discussing with predecessors and hypothetical (contemporary and future) points of view – in his philosophical writings. With the enunciational shift in “. . .and the idea is, Kant says look. . .” she sets up a Presentation space with Kant’s theatrically staged fictive enunciation. The dramatized enunciation is in the form of “direct quotation”: the enuciation is presented as belonging to Kant. He is represented as the speaker responsible for the propositional content therein. Naturally, it is S1’s editorial viewpoint that structures the discourse (in Semiotic space) and S1 who is ultimately responsible for the accuracy of the representation of the propositional content in Kant’s utterances and she alone is responsible for the perlocutionary success, of both the embedding and the embedded discourse. The students are expected to give credence to the teacher’s rendition of Kant’s philosophy, as expressed (in part) by Kant’s fictive enunciation. They are not, however, expected to give credence to the extra-propositional communication taking place in the presentation. They are not expected, for instance, to walk away with the impression that Kant is a great pedagog. Given the rhetoric of the delivery of his utterances, and the interaction implied, Kant comes across as a fairly affable and sprightly fellow (compared to the more straight-laced author of Critique of Pure Reason), but this is not part of the meaning of the presentation (as sign for the Reference space). In the blend, where the presentation is represented as BEING the reference (and vice versa), S1’s fictive speech (in Presentation space) represents Kant’s eternal views (in Reference space). Kant is virtually verbalizing his views in person. S1, his advocate, has given him a voice and the opportunity to address his audience directly. In Reference space, his audience is contemporary and future philosophers – potential readers of his work (in literary terms: the (model) reader Kant had in mind). In Presentation space, the audience is a room full of students. A virtual identity relation (mapping) connects the two in the blend. As in often observed in performances of enunciational shifts (and in mental simulations of such shifts), intonation, pitch, and other verbal or non-verbal gestures, function as space builders, indicating the embedding of a new enunciation. We can assume that this is also the case in setting up the Presentation space of the Enacting Kant blend. The Situational relevance of the blend is an inquiry into the unity of consciousness and Kant’s explanatory “transcendental argument”. The discourse topic constrains the elaboration of the input spaces as the discourse progresses and helps the conceptualizers decide what mappings to perform and what parts of the presentation and reference, respectively, are relevant to the blend and what represented content is incidental. It is incidental, for instance, that Kant makes his views known in writing in the reference space, and that his writings are (or were
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.30 (138)
Line Brandt
originally) in German rather than English. The focal point is his beliefs, not the medium of expression. The Argumentational relevance of blending these inputs is creating a representation that inspires a certain intellectual attitude. The script of dialogue, with its insistence on shared attention and jointly following, or rather developing, a train of thought, provokes an intellectually fruitful dynamic, leading to a heightened sense of awareness and clarity of thought. Dialogue inspires frequent interruptions and clarifications, and the epistemic resistance offered by the constructive criticism of a 2nd person challenges the tenets and hypotheses proposed, which can alter or strengthen the veridical stance of the participants. This is the framing of the script projected to Virtual space. The epistemic affordances of dialogic exchange benefit the pragmatic objective of clarifying Kant’s philosophy; the views discussed become more clearly delineated and their veridical status more easily resolved. The cooperative element and the trust inspired by sharing representations in a dialogue also affect the illocution. The Illocutionary relevance of setting up these spaces is didactic instruction (to illustrate Kant’s argumentation). S1 constructs a virtual being blend, a dramatized representation of Kant in virtual face-to-face dialogue with her students, structured by a specific framing of the dialogue script, and as a result of this cognitive maneuver Kant appears as a figure eager to explain his views and prepared to argue for them. The Pragmatic implication of the blend is that the 2nd person subject in the communication in Semiotic space – the students – understand the history and nature of his point of view so that they can evaluate it (as one to be shared or not), or so they may at least reproduce it, at the exam.
Contrast blends The next example of conceptual integration utilizing fictive interaction for expressive purposes is an example of a contrast blend. The Virtual space in a contrast blend is a sign consisting of a Reference space – the topic scenario – and a Presentation space which is set up in contrast to the reference. A contextually specified Argumentational relevance sorts out the conditional argument structure of the blend which brings about a contrastive predication of the reference. Consider the following example which is an excerpt of a posting in an internet debate on how the then pending US election (presidential candidates: G. W. Bush vs. J. F. Kerry) might influence the situation in post-invasion Iraq. The letter is posted by “wal” (henceforth: Subject 1: S1) September 6, 2004, at whistlestopper.com, in response to one by “CyNix” (henceforth Subject 2: S2). The enunciator, S1, is arguing for electoral support for Kerry. He starts his posting by quoting S2’s previous posting, followed by a personal address:
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.31 (139)
Chapter 4. A semiotic approach to fictive interaction
CyNix – The issue is not how many countries have given approval to the occupation. The issue is how many non-U.S. troops do we have supporting the mission. Answer: Far too few.
He continues: So how will Kerry get more nations to contribute more soldiers? Well for one, he won’t position the request such that any nation contributing troops has to be seen as endorsing the invasion as Bush did. Bush never said, “Look, I respect your opposition to the war. We thought we had a handle on things but we don’t. We need your help.” All Bush’s rhetoric was very dismissive of the ”old Europe”. You are with us or you are with the terrorists is not a sincere request for help, it is a bullying threat. (my italics)
The utterance “You are with us or you are with the terrorists” is presented as a direct quote of a statement made by President Bush. This statement is contrasted by the representation of a fictive interaction – one that Bush explicitly did not engage in: ”Look, I respect your opposition to the war. We thought we had a handle on things but we don’t. We need your help.” In this fictive interaction, Bush is addressing the international community, in response to the evident and prevalent opposition to the war (“Look, I respect your opposition. . .”). This interaction is imagined to take place at some point after it became apparent that help was needed, and instead of the real interactions that did take place, of which the “You are with us or. . .” address is only one example (“all Bush’s rhetoric”), because one excludes the other. The implication is that Bush should have said (something to the effect of) “Look, I respect your opposition. . .etc.”, but since he didn’t, he brought about a situation where there are “far too few” troops in Iraq in the aftermath of the invasion. The fictive interaction blend serves to put Bush’s handling of a particular crisis into critical perspective by contrasting it with a non-factual scenario that is presented as a past real possibility, the consequences of which would have been favorable to the current situation (at the time of enunciation), and as such it serves as an argument for favoring Kerry at the election. The semiosis in base space is to be seen as one contribution in an ongoing debate over the internet. In this computer-based form of interaction S1 and S2 are debating the coming election (2004) and its implications for the prospects of attaining military support from foreign nations in post-invasion Iraq. The speaker sets up a Reference space with Bush and Kerry and their differing political agendas, as seen in relation to the critical situation in Iraq and future approaches to getting international help. A Presentation space is set up with Bush saying: “Look, I respect your opposition to the war. We thought we had a handle on things but we don’t. We need your help.” The ontological status of the space is defined by a negation (“Bush never said”), disappointing a presumption or ex-
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.32 (140)
Line Brandt
Situational relevance (debate)
Phenoworld
fictive
past
Presentation spac e Semiotic space S1
Bush never said ”Look, ...” S2
Situation
Fict ive enunciation: Bush saying: ”L ook, I respect your opposition... ”
Situational relevance (upcoming election, US approach to getting int’l help in Iraq)
Ar gumentational relevance Schema: Counterfactual reasoning
R eference space Ke rry: no request to endorse invasion Bush saying: “You are with us or you are with the terrorists” (‘bullying threat’) Result: too few troops
Framing by negation:
Virtual space
Bush’s handling of int’l relations seen in the perspective of what he did NOT say and do: Bush’s actual stance by contrast to his fictive concession
Contrast blend: Il locutionary releva nce (swaying S2 by convincing argument)
Pr agmatic implication: Electing Bush would not be in the country ’ s best interest
Criticism: Bush’s unnecessarily alienating rhetoric damaged US international relations and consequently the US will fail to get the needed international support in post-invasion Iraq under his leadership
Figure 4.
pectation that the attitude expressed in the utterance could have existed. Since it manifestly did not, the proposed scenario is fictive. The embedded (fictive) enunciation metonymically expresses a general approach – by the very common, experientially motivated, saying for thinking/believing/being metonymy described earlier. The fictive interaction in the Presentation space consists of a metonymic personal enunciation (cf. the section on ‘Metonymic fictive interaction’ above) and the communicational circumstances evoked by imagining it – notably, and importantly, including a more favorable perlocutionary effect than had the “bullying threat”. By contrast to the fictive interaction in Presentation space, the Reference space has Bush addressing his potential allies, saying “You are with us or you are with the terrorists”, an utterance which is overtly framed as a “bullying threat”. The attitude expressed in this utterance is thought to have led to a failure to get a sufficient number of international troops to help US troops in Iraq. In Virtual space, the reference is framed by negation: Bush’s actual stance by contrast to his fictive concession. The President’s handling of foreign relations is seen in the perspective of what he did not say and do.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.33 (141)
Chapter 4. A semiotic approach to fictive interaction
The Argumentational relevance is counterfactual reasoning.13 Virtual space is structured by conditional relations. The ‘bully’ approach generated little international support. If the US had taken a more respectful approach to the international community initially, more countries would be likely to offer their support in the present (the deictic present of the enunciation in Semiotic space), regardless of their stance on the invasion. If Bush had chosen the approach expressed by “Look, I respect your opposition. . .”, the US would have been in a better position to appeal to other countries to send troops to post-invasion Iraq. The meaning of the blend is a criticism: Bush’s unnecessarily alienating rhetoric damaged US international relations and consequently the US will fail to get the needed international support in post-invasion Iraq under his leadership. The pragmatic implication is that since Bush will not be able to form the necessary alliances, Kerry is the better candidate. By staging a fictive enunciation and having the implied interaction contrastively predicate the referential state of affairs, S1’s argumentation gains expressive impetus, prompting the emergence of those inferences which support his assessment of what needs to be done.
Analogy blends A last example of representation of fictive interaction for expressive purposes is an example of an analogy blend. dizzy said on August 4, 2005 01:20 AM: [. . .] Being against stealing doesn’t just mean you don’t steal yourself, it also means actively opposing others who steal. Now, given that abortion tends to lie much higher than stealing on the moral conundrum scale for persons who feel this way, it seems like the imperative is only stronger. [. . .] sparklegirl said on August 4, 2005 02:11 AM: I agree, Dizzy. That’s why I’ve never liked those bumper stickers that say “If you don’t like an abortion, don’t have one” – from the point of view of a person who genuinely believes abortion is wrong, that’s like saying “If you
. The Argumentational relevance in this example is a schema of counterfactual reasoning. It is not represented as a counterfactual mental space; counterfactual mental spaces are imagined scenarios that are presented as states of affairs that could have existed but do not, or that could never exist (but yield useful inferences nevertheless when used to signify a reference). Typically these spaces are presented with the epistemic distance of the past tense or in the subjunctive mode.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.34 (142)
Line Brandt
don’t like slavery, don’t own slaves.” There are better ways to defend abortion rights. (source: pandagon.net, my italics)
In this blend, the presentational signifier is an embedded generic enunciation, introduced by an explicit space builder denoting the embedding of a new enunciation: “that’s like saying “If you don’t like slavery, don’t own slaves.”. The embedded fictive interaction helps predicate the topic at hand: “bumper stickers that say “If you don’t like an abortion, don’t have one” ” and the attitude expressed by this communication. The two interactants in base space (“dizzy” and “sparklegirl”: henceforth S1 and S2) are cooperatively active in establishing the blend. S2 sets up the blend whose emergent inferences confirm S1’s statement that “Being against stealing doesn’t just mean you don’t steal yourself, it also means actively opposing others who steal”. The Argumentational relevance is moral judgment. The meaning of the blend is that norms for evaluating behavior as acceptable/unacceptable depend not just on private likes and dislikes but on ethical concerns for the individuals who potentially suffer as a consequence of the behavior in question. The analogies to theft and slavery bring focus to the alleged bereaved party in the case of abortion: the would-be-children. The unspoken premise of the argument is the framing of these unborn fetuses as individuals. The moral judgment shared by S1 and S2 is thus given an ethical basis, which would have to be addressed to critique the stance that “being against” (cf. S1) abortions entails “actively opposing” those who have them.14
Concluding remarks Language is inherently dialogic, from the lexical level of closed class forms (personal pronouns, negational adverbs etc.) to the level of mental spaces. The presence of a 1st and a 2nd person and a shared object of attention is characteristic of any natural language use, whether intra- or interpersonal. Dialogue is a basic experiential domain in the phenomenal world and is frequently recruited as a source for metonymic expressivity and as a presentational input or as a “relevance-making” script framing the Virtual space in semiotic mental space integrations. Taking Fauconnier & Turner’s theory of conceptual integration and Pascual’s work on the representation of fictive interaction as a starting point, I have proposed a theoretical framework for analyzing phenomena that exploit the cognitive possibility of representing imaginary verbal interactions for expressive purposes, a . See also Coulson & Pascual 2006 for semantic analyses of con-abortion rhetoric: ‘For the sake of argument: Mourning the unborn and reviving the dead through conceptual blending’.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.35 (143)
Chapter 4. A semiotic approach to fictive interaction
framework which is hypothesized to be useful not just for the special case of taking rhetorical advantage of the easy mental access to the semantic script of verbal interaction, but for the general study of conceptual integration of mental spaces in communication: a pragmatically oriented cognitive semantics. Examples of imaginary interaction metonymies were distinguished from virtual verbal interaction in semiotic blends. Two kinds of embedded metonymic enunciation were delineated: personal and generic enunciation. In metonymic personal enunciation, the utterances are fictive and the interactants specific. In metonymic generic enunciation, the interactants are generic, as are the imaginary utterances. As we saw in the analysis of fictive interaction blends, these metonymies sometimes figure at the intraspace level in more complex semiotic constructs. Whereas a lexeme, our stock example of a sign, integrates a phonological structure with a semantic content, more complex signs can be described as conceptual integrations of entire mental spaces. One mental space stands for another in a sign relation, motivated by illocutional, situational and argumentational relevance. When both the signifying presentation and the signified reference are semantic, the cognitive result is motivated predication; one input space is about the other in some specifiable respect. The division of semiotic blends into contrastive co-temporality blends, virtual being blends (identity blends), contrast blends, and analogy blends, exemplified in this chapter, is motivated by the predicative and structural characteristics of the blends. In the case of contrastive co-temporality blends, for example, “cotemporality” is a structural feature, while “contrastive” describes the nature of the predication. In virtual being blends the blended elements are represented as virtually identical. They are imagined as identical in the blended space, and the predicative effect is qualitative; the referential content is signified as qualitatively identical to the presentational content in some contextually relevant respect. Aspects of this theory are much in line with Sperber & Wilson’s Relevance Theory (cf. Brandt & Brandt 2005a), only the focus is semantic analysis rather than behavioral effects. The concept of relevance is approached from the perspective of cognitive semiotics, calling for detailed semantic analysis of the manifestations of relevance in discourse, a practice which is still in its early phases but which has so far led to certain useful distinctions (cf. above) and that appears to aid our understanding of “selective projection” and “emergence” in (semiotically) blended representations. The focus in this chapter has been on sorting out different manifestations of imaginary verbal interactions in discourse, predominantly fictive verbal interaction, represented metonymically in the case of embedded metonymic enunciation, and ‘virtually’, in semiotic blends. A distinction was made between fictive and generic
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.36 (144)
Line Brandt
interaction. Both are examples of “a conceptual channel of communication underlying the observable interaction between participants” (Pascual), but fictivity is only ascribed to representations of particulars. I suppose from the viewpoint of analytic philosophy the remarkable thing about the study of fictive verbal interaction is that it shows natural language to be oriented not just toward denotations but towards the representations themselves and the manner in which something is communicated, even when the discourse is propositional. From a cognitive perspective, however, it is a portal to gaining insight about what is going on in our minds when we communicate, and to developing a natural (non-analytical) philosophy of language based on the phenomenological study of empirical language use. How do we make sense of representations that are not vested with belief and that are not presented as fictive referents (cp. the ‘fiction spaces’ of novels, movies etc., which are referential)? I hope to have contributed to answering this question. Fictivity in representation is an interesting phenomenon, but exclusive focus may occlude a bigger picture. For one, as we have discovered, verbal interactions are sometimes imagined generically as well, and there may be further distinctions to be made. I think future explorations of the phenomenon may shift the focus from the fictive part to the verbal interaction part – towards descriptions of the ways in which the dialogue script influences language both at the level of grammar and at the level of creative ‘sense-making’. However, to establish some common ground from where to proceed, I have chosen to make fictive verbal interaction the focus of the present paper, since this phenomenon has already been established as an object of study (Pascual15 ). Part of my aim has been to show how the study of fictive verbal interaction overlaps with the study of enunciation. One noteworthy thing about all the examples of fictive/generic/virtual verbal interaction we have gone through is the staging of enunciation in the representations. It appears that the point of imagining verbal interaction, as a means to represent, is the experienced animation of the imagination that happens when concepts are dramatized. Our analysis of the phenomenon suggests that an essential component in fictive verbal interaction is the performative staging of an enunciation – what we could call “dramatized enunciation”. As far as I can tell, the notion of dramatization used here is largely interchangeable with P. Aa. Brandt’s notion of theatricality (see Brandt 2004: From Gesture to Theatricality: 219–243); . Her 2002 book has as its basic research topic “the intersection between language, interaction and cognition”: “The focus is on a phenomenon – never before studied as such – which I call fictive interaction (FI). This constitutes the use of the schematic interactional structure of ordinary communication as a common organizing frame to understand, think, and talk about verbal as well as non-verbal entities, processes, and relationships.” (Pascual 2002: 1)
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.37 (145)
Chapter 4. A semiotic approach to fictive interaction
the enunciation is theatrically performed – on the stage of our inner vision. This observation points to a broader phenomenon: the pervasiveness of dramatization of conceptual content in the phenomenal experience of representations in meaning construction. For instance, as has been shown, here and elsewhere, it is a fairly common communicative practice to dramatize a Reference space by presenting it in a Virtual space, that is, as if it were something else. The “as-if-ness” of virtuality is theatrical and only indirectly referential. It has to do with semantic framing, and as such belongs to the realm of conceptuality. I expect that continued exploration of the pragma-semantic aspects of rhetorical fiction-making, of enunciational embedding, and semiotic virtuality, will reveal more about conceptual cognition, semiotic and otherwise. Let me conclude by outlining some methodological and philosophical implications of the cognitive semiotic framework applied in this paper. Cognitive semiotics is an interdisciplinary field developed in the 1990ies at the Center for Semiotics at the University of Aarhus. Partly inspired by Talmy’s linguistics and Merleau-Ponty’s phenomenology, one of the basic observations that this branch of culturally oriented cognitive science takes as fundamental to its approach to cultural phenomena – like language and art – is that cognizers do not have conscious access to neural processes but do have conscious access to representational processes. Cognizers have (actual or potential) conscious access to meaning, insofar as meaning is taken in the semiotic sense, and as Talmy has pointed out the performance of semantic analysis depends on the analyst’s ability to introspect attentively. Just as aesthetic appreciation of music is an acquired skill that can be developed by the application of talent and effort, semantic sensibility is a talent and a skill that training can advance. Introspection is unavoidable because all meaning is located in conscious experience (cf. Talmy 2000: 6). Most of the time we observe the content of our minds without meta-awareness, but metarepresentational access remains an option to the curious and an analytical prerequisite to scholars of linguistic meaning. We could call this the phenomenological approach to semantics. If we relate this conception of semantics to conceptual integration theory, one logical consequence is that mental spaces be viewed as imaginal scenarios rather than sets. In Fauconnier’s 1994 and 1997 books on mental spaces, mental spaces were described as structured, incremental sets with elements and relations holding between them (1994: 16). A mapping between mental spaces was described as “a correspondence between two sets that assigns to each element in the first a counterpart in the second.” (1997: 1, fn. 1). This conception of mental spaces has not been contradicted in later writings, as far as I know, and it lives on in contemporary diagramming of blends, most notably perhaps in construction grammar where one input to the blend is merely a list of elements, an unintegrated set that
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.38 (146)
Line Brandt
is principally phenomenologically void: it is not and cannot be experienced by any cognizer. I suspect that Fauconnier would hesitate to describe mental spaces as sets now that he is no longer primarily addressing analytic philosophers of language. He already hinted at a skepticism toward the appropriateness of this description in 1994, when he wrote in a footnote that “[m]ental spaces may turn out to be endowed with a different and/or richer kind of structure than sets can have (e.g., image schemata or nonpropositional prototypes).” (1994: 171: fn. 11) In 1994, in the foreword to Fauconnier’s first book on the subject (in English), Lakoff and Sweetser describe mental space configurations as “mental models of discourse” (1994: xxxix). This seems reasonable enough, at least in the context of semantics, but the formulation calls for some clarification. First of all it calls for a clarification of what is meant by “mental”. Secondly, the reference to “discourse” deserves some attention. “Discourse” can refer to communicative verbal interaction, and it can also refer to the “text” that comes into existence via this interaction: the communicated semantic content. Here it refers strictly to what the communication is about and not communication itself. My view, as can be surmised from the analyses in this chapter, is that a model of semantic meaning needs to encompass both. The other important corollary of the cognitive semiotic framework that I wanted to draw attention to is that representations have meaning by virtue of their actual/potential occurrence in discourse and other expressive practices; they do not have meaning in and of themselves. The foundation of linguistic meaning is the intersubjectivity of semiosis (the exchange of signs in communication and artistic expression). We can call this the semiotic approach to semantics. The phenomenological approach rules out principally unconscious cognitive content as data in non-experimental empirical scientific studies, specifically blending analyses of inaccessible semantic content in the “cognitive unconscious,” (cp. Fauconnier & Turner 2002), and the semiotic approach rules out any essentialist approaches to the study of meaning, specifically truth-conditional semantics, as well as any approaches to conceptual blending and integration that view representations as having inherent “meaning“ outside their function in discourse and other expressive practices. The problem with the term “meaning” is that it is used in at least three different senses, which may lead to a false sense of consensus. On some occasions of use it refers to something’s significance or importance. On other occasions it is somewhat synonymous with “coherence”, as when applied to the process of gestalting perceptions or the formation of concepts (this appears to be the case in Chapter 7 [Williams]). The third use, which I hope to have made it obvious is the intended use in this chapter, is semiotic and refers to the intention motivating an act of sig-
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.39 (147)
Chapter 4. A semiotic approach to fictive interaction
nification: an utterance, a work of art, a traffic sign, whatever. Meaning in this last sense is what the study of cognitive semiotics is all about. In a time when physicalist philosophy tends to dominate the cognitive sciences I find it important to point out the fact that signifying gestures, including verbal interactions, are as basic as physical matter in human phenomenal reality. We learn to communicate before we learn to do anything else. In the light of these facts I think it is time for cognitive semantics to revise its concept of embodied cognition and come up with a more comprehensive notion of embodiment. Such a notion would take social attunement into account. This phenomenally grounded body would not just be a self-contained body pouring liquids in and out of itself (cf. Lakoff ’s famed container schema) and autistically counting its steps on its lonely walk down the road to higher abstraction; it would be a gesturing body, coordinating gestures and imaginings with some body else. It would be seeking out intimacy with others through talking, and touching. It would be moving its behind to the rhythm of music. Cognitive semantics needs a more inclusive theory of how minds perceive and conceive of human bodies, in solitary and in social locomotion. Human thought is embodied, to be sure, but what’s more, it is embootyed. The reductionist empiricist version of embodiment theory is too solipsistic to account for phenomena belonging to the realm of verbal interaction, for one, such as the possibility of representing communication in communication, metonymically, metaphorically and in creative blends such as those we have seen here. Let this final observation serve as a call for the recognition of the social aspect of the bodily experience as a foundation for conceptual advancement – a call for embootyment theory.
References Émile Benveniste. 1971. Problems in General Linguistics, translated by Mary E. Meek, Miami Linguistics Series, 8, Coral Gables: University of Miami Press. Brandt, L. & Brandt, P.Aa. 2005a. ‘Making sense of a blend. A cognitive-semiotic approach to metaphor’, Annual Review of Cognitive Linguistics 3, 2005: 216–249. Brandt, L. & Brandt, P.Aa. 2005b. ‘Cognitive Poetics and Imagery’. European Journal of English Studies, Vol. 9, No. 2 August 2005: 117–130. Brandt, P. Aa. 2004. Spaces, Domains, and Meaning. Essays in Cognitive Semiotics, Peter Lang, Bern. Chaon, D. 2001. Among the Missing, Ballantine Books. Coulson, S. & Pascual, E. 2006. ‘For the sake of argument: Mourning the unborn and reviving the dead through conceptual blending’, Annual Review of Cognitive Linguistics, Vol. 4(1): 153–181. Fauconnier, G. & Turner, M. 2002. The Way We Think, New York: Basic Books. Fauconnier, G. 1997. Mappings in Thought and Language, Cambridge University Press.
TSL[v.20020404] Prn:8/01/2008; 15:16
F: PB17004.tex / p.40 (148)
Line Brandt
Fauconnier, G. 1996. ‘Analogical counterfactuals’, in Spaces, Worlds, and Grammar, G. Fauconnier, E. Sweetser (eds.), Chicago: University of Chicago Press, Fauconnier, G. [1985] 1994. Mental Spaces: Aspects of Meaning Construction in Natural Language, Cambridge University Press. Langacker, R. W. 2001. ‘Dynamicity in Grammar’, Axiomathes 12, Kluwer Academic Publishers: 7–33. Langacker, R. W. 1999. ‘Virtual Reality’, Studies in the Linguistic Sciences 29/2: 78–103. Langacker, R. W. 1997. ‘Generics and habituals’, in On Conditionals Again, A. Athanasiadou & R. Dirven (eds.), Current Issues in Linguistic Theory 143, John Benjamins: 191–222. Langacker, R. W. 1996. ‘A constraint on progressive generics’, in Conceptual Structure, Discourse and Language, A. Goldberg (ed.), Stanford: CSLI Publications: 289–302. Matsumoto, Y. 1996. ‘Subjective-change expressions in Japanese and their cognitive linguistic bases’, in Spaces, Worlds, and Grammar, G. Fauconnier & E. Sweetser (eds.), Chicago: University of Chicago Press. Michigan Corpus of Academic Spoken English. Available at http://www.uti.umich.edu/m/ micase/ Pascual, E. & de Vries, L. In prep. 2007. ‘Hear it to believe it: On fictive verbal interaction as grammatical structure in oral English and languages without writing’, to be submitted to Function of Language. Pascual, E. 2006a. ‘Fictive interaction within the sentence: A communicative type of fictivity in grammar’, Cognitive Linguistics 17(2), 245–267. Pascual, E. (2006b). ‘Questions in legal monologues: Fictive interaction as argumentative strategy in a murder trial’, Text and Talk 26(3): 383–402. Pascual, E. 2005. ‘It’s like, why enacted dialogues?: On the multifunctionality of direct speech in the jury room’. [Submitted to Journal of Pragmatics]. Pascual, E. 2002. Imaginary Trialogues: Conceptual Blending and Fictive Interaction in Criminal Courts, Utrecht: LOT. Safran Foer, Jonathan. 2002. Everything is Illuminated, Penguin Books. Sweetser, E. 1990. From etymology to pragmatics. Metaphorical and cultural aspects of semantic structure, Cambridge studies in linguistics, Cambridge University Press. Talmy, L. 2000. Toward a Cognitive Semantics, Volume 1, Cambridge, MA: MIT Press. Talmy, L. 1996. ‘Fictive Motion in Language and “Ception”’, in Language and Space, P. Bloom et al. (eds.), Cambridge, MA and London: MIT Press/Bradford: 211–276.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.1 (149)
chapter
Designing clinical experiences with words1 Three layers of analysis in clinical case studies Todd Oakley and David Kaufer We demonstrate the value gained when combining coarse-grained methods of corpus-based genre studies with fine-grained methods of mental spaces analysis. This combination compensates for inherent weaknesses in each individual approach. Specifically, the mental spaces approaches to discourse can fail to motivate the selection of samples, relying too often on the investigator’s limited capacity to attend to the larger field of cultural production. In contrast, corpus-based approaches can miss the rich interpretative analyses available once selections are made, as the investigator’s tools have an extremely limited capacity to attend to the fine-grained variations of human communication. We advance a three-layered approach to discourse analysis as applied to a corpus of 35 clinical case studies appearing in the journal Hospital Practice, a publication for general practitioners and emergency room physicians.
Introduction Mental spaces and genre studies: A dilemma Different linguistic schools are notorious for their acerbic disagreements. Charles Fillmore characterizes one such disagreement: when the corpus linguist asks the generative linguist, “Why should I think what you tell me is true?” the generative linguist responds in kind with “Why should I think what you tell me is interesting?” (1992: 35). This exchange encapsulates a basic disagreement between empirically minded corpus linguists interested in accurate descriptions of language use and rationally minded generative linguists interested in advancing knowledge of universal grammar, a disagreement originating from the different units of measurement taken as fundamental to each school. Generative linguists take as their unit . We thank Per Aage Brandt, Alan Cienki, and Anders Hougaard for their criticisms and suggestions.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.2 (150)
Todd Oakley and David Kaufer
the whole sentence whose frequency of recurrence statistically converges on zero, and the point of doing linguistics is to figure out what the ideal speaker can do with her language. Corpus linguists, on the other hand, rely on units of measurement (words, phrases, clauses, other sub-sentential constructions, as well as token and type frequencies) that have more statistical robustness, and these units are taken to reflect what people do with language habitually. Generative linguists seek to discover “what is possible and impossible” at the expense of what is probable, likely, and typical. Corpus linguists seek to describe with greater reliability than introspection alone the things we do regularly and frequently with language at the expense of explaining what is possible and impossible (Krishnamurthy 2001). The two camps appear to be in a stalemate, with text corpora playing at best a peripheral role in generative linguistic investigations (cf. Meyer 2002: 2–5). Mental Spaces and Conceptual Integration theory (hereafter mental spaces theory) – a programmatic approach to meaning construction within Cognitive Linguistics – and Corpus-Based Genre Studies (hereafter genre studies) – currently an area of specialization within the even broader field of rhetorical and composition studies – evidence two horns of a very similar dilemma, a dilemma over methods guiding the selection of data as well as over the granularity the analysis. Mental Spaces theorists are known for analyzing seemingly singular linguistic and semiotic phenomena, as evidenced by the now famous examples, such as the riddle of the Buddhist Monk, the Debate with Kant, the Ghost of Northern Lights, and the Grim Reaper as cultural icon discussed in Fauconnier & Turner (2002) and elsewhere in the mental spaces literature. The genre theorist might criticize a mental spaces analysis thus: “Yes, those are interesting examples, but what do they tell us about the typical modes of linguistic representation in a particular kind of speaking or writing?” The mental spaces theorist might respond, “They tell us a great deal about the underlying cognitive processes unfolding in discourse and interaction, for the idiosyncrasies, as you like to call them, are really the results of the same underlying processes used in typical forms of speaking and writing.” In other words, the idiosyncratic examples are representative of the kinds of things human beings typically do with language and other symbol systems. At this point, the genre theorist might retort, “How do you know they are representative without establishing empirical criteria for differentiating the typical from the rare?” The mental spaces theorist may respond by saying, “Our own framework reveals this to be the case, as the more mundane examples fall out from a treatment of these so-called ‘odd-ball’ cases, which, in fact, are not so odd: we find them everywhere.” So the debate continues along these lines. Assuming this characterization has some truth to it, then, so characterized, the two research groups seem to be at an impasse, for what makes a linguistic phenomenon and its description and explanation true and interesting within a
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.3 (151)
Chapter 5. Three layers of analysis
mental spaces framework is precisely what makes it irrelevant and uninteresting within the genre studies framework. As researchers living in both words, we do not think this dilemma is without resolution, as appears to be the case between corpus and generative linguists. We prefer to think of the dilemma as stemming from complementary (rather than contradictory) notions of what it means to study language at the level of discourse. Mental spaces theory has a long history of providing detailed accounts of specific instances of meaning making in language and other modes of representation. Most of the compelling developments and applications of mental spaces issue from detailed studies of singular phenomena. Indeed, the contributions in the volume aim to show how this approach advances understanding of a diverse range of discursive and interactive processes, and systematic application of a mental spaces approach to the study of discourse is to be applauded. Genre theorists by contrast tend to focus on capturing regularities of form and meaning within specific domains of written discourse. Much work in rhetoric and composition studies aims to generalize across a range of textual phenomena with the aim of understanding the functional characteristics of written genres, understood not as frozen forms but as “typified responses to recurring situations” (Miller 1984: 152). Under this view, the primary goal of genre theorists is to explain how global expectations constrain the options writers can exercise in a given rhetorical situation. Genre theorists such as Devitt (1993: 576) understand genres as those kinds of responses that, over time, have been shown to be “fitting” for types of rhetorical situations, with the similarities among theses appropriate responses established as “generic conventions.” Bawarshi (1997) goes so far as to assert that genres help determine the way individuals think and act. While mental space theorists find many riches in the individual examples, genre theorists value individual text samples only insofar as they are representative of a whole genre. As suggested above, mental space theory and genre studies have similar difficulties that arise from opposite horns of the methodological dilemma. By looking in detail at diverse sorts of sentential, clausal, phrasal, and lexical and grammatical phenomena, mental spaces theory has established itself as an exceptionally subtle framework for capturing the micro-details of meaning construction (see Fauconnier & Sweetser 1994; Coulson & Oakley 2000, 2005). Nonetheless, in satisfying the criterion of exhaustive description of singular phenomena, mental spaces analysis is methodologically at the mercy of the interpreter’s interests and biases, and thus remains open to the criticism that it is not able to generalize or check its results. Genre studies lacks means of analyzing specific textual artifacts in a manner that captures the recurrent unfolding of meaning construction as readers experience texts of this or that kind. It is at the mercy of the interpreter’s blindness to analytic detail.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.4 (152)
Todd Oakley and David Kaufer
The tension between the exacting interpretive frameworks of mental spaces and the more impressionistic interpretive traditions of genre studies is the unavoidable result of different initial aims and interests, yet it is possible to bring them into productive alignment. That is precisely our aim in this article: to show that mental spaces can be used to advance the aims of genre studies at the same time that it captures the dynamics of meaning construction that genre studies has little ability to do. We argue that mental spaces can function as a viable interpretative framework at three layers of analysis, and we show how this can be accomplished by examining a specific genre of writing: the clinical case study. Our objects if inquiry consist of thirty-five articles on the diagnosis and treatment of diseases published from 1998–2001 in the journal Hospital Practice.2
Three layers of analysis: A brief overview Three layers of analysis capture important facets of the meaning construction process readers bring to bear on a text. At the every least, we consider it a methodological necessity to divide the textual analysis into three layers in order to align these two traditions of scholarship. But methodological convenience is not the only guiding principle, for we intuit that these layers of analysis actually reflect the way readers experience the written word. The Genre Layer refers to the reader’s experience of a text as a kind of event, and specifies what readers expect from such and such a text. At the genre layer, a text is experienced as a member of a “class of communicative events” (Swales 1990: 58). Of the three layers, this one is probably experienced more in the breach, as when a reader feels a norm has been violated. Editors and scholars are often most attuned to this layer of textual experience; nonetheless, these specialized viewpoints should be most profitably viewed as manifestations of readers’ experiences writ large. The Artifact Layer refers to the reader’s functional and “seamless” experience of a single text, or of its “gist.” Readers are using the text to learn about or to do something else, such as assemble a bookshelf, follow a court’s ruling in a particular case, or learn the best means of treating psoriasis. This is perhaps the most important and least understood layer of analysis for researchers interested in investigating the role texts play in culture and cognition, and it is the layer in which both mental spaces theory and genre studies has had little to say. The Grammar Layer refers to the reader’s “seamed” experiences of specific words or constructions experienced as one’s eyes move over the page. More specifically, this layer of analysis corresponds to the readers’ allocation of attention to the . The journal ceased publication after the 2001 volume.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.5 (153)
Chapter 5. Three layers of analysis
formal dimensions of specific morphological, syntactic, lexical, and typographic properties of a text. It also corresponds to semantic and pragmatic dimensions of words and phrases (i.e., evoked mental simulations of content, as for example, when the negative or positive affect associated with specific lexical items). The study of conditionals analogical counterfactuals, adverbial clauses and phrases, referential opacity, and presupposition and other classic topics in the mental spaces literature correspond to this layer of analysis (see Fauconnier 1994, 1997).
Data mining Bringing the mental spaces and genre analysis into productive alignment along these three layers can only begin if we are confident that the examples from our corpus can be defended as representative. In our view of genre, a representative text means that it reflects a discernable rhetorical strategy consistently present within the corpus. We seek to demonstrate how the parsing program Docuscope (developed by Kaufer, Ishizaki et al. 2004) combined with factor analysis can provide independent justification for the selection of texts exhibiting genre features significant to the aims and purpose of the corpus. At the same time, we wish to show how three layers of analysis account for the types of mental spaces used in designing clinical experiences with words and to account for signal differences in the specific features of these mental spaces among the six exemplar texts under analysis. Before we begin the analysis, we wish to be clear about what we are not arguing. We are not claiming a direct causal relationship among three layers, such that minute patterns of variation captured at the grammar layer of analysis will determine what happens at the artifact and genre layers or vice versa. There is considerable variability between the different layers, indicating multiple possible tactics for designing clinical experiences. We are claiming that the analysis of recurrent features whose statistical aggregation forms a “macro” snapshot over the corpus uncovers selectional dependencies among the examples that can loosely constrain the interpretation of discourse structures active in the artifact and grammar layers. These constraints operate as selection filters for lower level objects. Lexical and grammatical phenomena chosen from this filter fill a crossover between microdetail and important macro recurrent regularities. Yet these constraints are also sufficiently weak to insure that, once identified from a macro-filter, the text analyst can retain complete interpretative independence over the kind of analysis assigned to these objects. The value of this approach is in allowing for detailed microanalysis to proceed without losing sight of recurrent instances of similar forms throughout the artifact or genre strata. Another name for these recurrent traces linking specific linguistic constructions to a whole corpus of text are Language Action Types, the presence or absence of such types in a text can yield a statistical profile for a rhetorical strategy operating at the artifact layer of analysis. We iden-
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.6 (154)
Todd Oakley and David Kaufer
tify three distinct strategies distributed among the clinical experience corpus: the challenge strategy; the consultation strategy; and the presentation strategy. For the purpose of this study, we assume that a mental space is an interpretive construct operating at the generic, artifact, and grammar layers of analysis. Following Oakley and Coulson (this volume), we define mental spaces similarly as representations of scenes and situations relevant to a given discourse genre as a means for packaging and construing information about elements pertaining to readers’ “centers of interest” within an interactive context. At the artifact layer, a mental space comprises a whole scene and situations relevant to the global conceptual and interactive aims of the discourse. At the grammar layer, a mental space comprises facets of scenes and situations relevant to specific parts of the discourse. Networks of mental spaces show how they can guide explication of generic and narratological conventions thought to reside at the genre layer of analysis. We likewise use the same mode of analysis adapted by Oakley and Coulson (this volume) and Brandt (this volume) from the protocol developed by Brandt and Brandt (2005).
The genre-layer: Intuition and factor analysis All interpretive endeavors proceed by understanding the relations of parts to whole and whole to parts. All understanding is circular (but not viciously so). Gallagher (2004) describes the process of interpretation: to understand the meaning of a particular passage in a text, one has to understand how it relates to the whole text, and to understand how the whole text functions rhetorically, one needs to see how that whole operates in relation to its parts. Such is the interpretive condition: one understands X only in relation to a context and one understands the context more completely when one understands X. Textual interpretation is conditioned on what one already knows. This circular process is consistent with schema theory, where schemas represent patterns we already know. In rhetorical hermeneutics, schemas provide a means for us to assimilate new information into established frameworks. It is invariably the case that new information can effect change in the schema such that the schema itself accommodates itself to new information (see Gallagher 2004: 4). This brings us to our first strata of analysis. The genre layer comprises cognitive schemas for clinical experience corpus based on a blend of intuition and computational analysis. We will address each in turn.
Intuition Our intuitions bear particular relevance to genre and narrative. Multiple exposures to the texts falling under the heading Clinical Experience will yield the inference
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.7 (155)
Chapter 5. Three layers of analysis
that in order for a text to count as a clinical experience it must 1) discuss medical cases, and 2) discuss the implications of each case for the practitioner, either as it relates to diagnosis or treatment. It is usually the case that the opening paragraphs establish the nature and scope of the problem to be addressed. Lloyd Bitzer (1969) refers to this as “exigence,” the imperfect state of affairs compelling one to speak or write. Writers establish exigence either by showing that instances of X disease or disorder are increasing, or by referring to new treatments for X, or by presenting a problem according to the argumentative form, “X is often misdiagnosed as Y.” In our analysis, exigency constitutes the situational and argumentative relevance governing the construction, maintenance, and differential activation of mental spaces (cf. Brandt, this volume; Oakley and Coulson, this volume). Our second intuition is that the entire clinical experience genre derives its schematic and thematic structure from a narrative of healing. As we define it, narrative is the unfolding of actions through time. Nelson (1996), quoting Polkinghorne (1988), considers narrative to be “the primary form by which human experience is made meaningful,” and narrative meaning “organizes human experience into temporally meaningful episodes” (1996: 186). Narrative structures organize whole domains of human experience. The experience of healthcare and medicine itself has a conceptual structure that is widely understood as the incorporation of different episodes (action and event sequences) into a coherent script of examination→diagnosis→prognosis→treatment→followup→outcome. What is more, the emotional valence of clinical narratives operates along a dysphoric↔euphoric continuum, usually with dysphoria accompanying most episodes, except perhaps outcomes. The physician’s professional and ethical duty is to try to ameliorate suffering, thereby inducing a more euphoric state in the patient. Once again, there is considerable accommodation of the schema to specific scenic realizations of this narrative schema, but the schema does provide a stable design strategy for meaningful clinical experiences. The generic and narratological dimensions provide a rationale for specifying the kinds of mental spaces that limn out the relevant scenes and situations of clinical encounters (equivalents of Nelson’s episodes). We hypothesize that mental spaces comprise the dramaturgical structure of the genre and thus are manifestations of episodes within the narrative and generic schemas. Similar to Oakley and Coulson (this volume), our model includes a concept of grounding, diagrammatically represented as three concentric circles, with the inner circle corresponding to discourse participants, the middle circle corresponding to interpersonal situational characteristics of the interaction, and outer circle corresponding to features of the setting itself. Specifically, our analysis posits a writer (s) – shorthand for all the authors, editors, reviewers, etc. involved in the fabrication and dissemination of the text; and reader(s) – shorthand for the type of reader implied in the structure of the text, namely a practicing physician, who reads these clinical experiences
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.8 (156)
Todd Oakley and David Kaufer
selectively. In summary, grounding depicts key features of the rhetorical situation that are held constant from one text to the next. The rhetorical situation is as follows: a primary care physician discusses clinical practices associated with common diseases/disorders encountered in the practitioner’s office or emergency room. In addition, we posit a professional setting of rhetorical engagement (such as an office or hospital or an otherwise similarly framed space) in which the reader assumes the role of medical practitioner. In our model, four kinds of spaces develop from the ground: presentation, reference, blend, and elaboration spaces. The presentation space typically consists of the knowledge about a particular disease or disorder or class of related diseases and their generalizations across clinical populations. The scene or situation is typological and the focus of attention is on illness: the conceptual domain through which the physician interacts with and understands the patient (i.e., is a sick person). The reference space pertains to particular clinical instances: this patient in front of me. The scene is observational and interactive. The blended space typically corresponds to the scene of diagnosis in which a particular patient is classified as suffering from disease/disorder X. This scene is definitive. This mental space network governing the genre stratum is not an instance of “double scope blending,” as defined by Fauconnier & Turner (2002: 340–345), because each space acquires its structure from the same conceptual domain of medicine with little competition; therefore, we prefer to call it a conceptual integration for depicting a type-token relationship, whereby a single person, X, is understood as having or displaying signs and symptoms of disease type, Y. The last space in this functional network is the elaboration space, a narrative extension of the blended space that typically focuses attention on the treatment of the person diagnosed with disease/disorder X. The scene is procedural. Figure 1 presents a mental space network specifying the conceptual structure for the clinical experience genre. It is a mental space analysis at the genre stratum, which also means that it is highly schematic and capable of assimilating new information to it just as the specific constituents of a mental space at the artifact and grammar layers may require accommodation of the schema to the phenomena being represented. For example, diagnosis does not figure prominently in every text; in some cases, diagnosis is assumed and the focus is on the episodes of treatment and their outcomes, so the schematic network accommodates by composing, completing, and elaborating scenes of treatment instead. The accommodation of the narrative schema may be accounted for by “high point analysis,” as made operational by Labov and Waletsky (1967). Narratives – be they fictional, personal, or (in this case) clinical – unfold in relation to a “point” or “complicating action” in which beginning parts orient toward and ending parts conclude. The dramatic episodes gather around these complicating actions. As
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.9 (157)
Chapter 5. Three layers of analysis
Figure 1. Mental spaces network for the clinical experience genre
suggested above, the dramaturgy of each clinical experience text is scripted around the challenges of either diagnosis or treatment; rarely does there appear to be an even split between the two.
Factor analysis We supplement these intuitions with a computational analysis of 35 examples of this corpus using the text-analysis program Docuscope. This program scans the surface of all the texts in the corpus looking for patterns of consistency and variations in the concentration of language action types, strings of English with a minimal functional interpretation. For instance, the phrase, “An anxious 83-yearold woman” will be tagged as combination of two language action types, negative affect and sense property, whereas the phrase “appears to be” will be tagged as the language action type uncertain thinking. In all, Docuscope consists of a database of 140 language action types tagged as one of three families of representations: Interior Perspectives; Relational Perspectives; and Exterior Perspectives. What does Docuscope do? This program performs intensive data mining, effectively extending the range of analysis by using a device that never transcends the artifact; it continuously and consistently “strip mines” the surface for these patterns. Unlike human readers who must transcend the textual artifact itself in order to make sense of it, Docuscope is the “anti-interpreter,” remaining resolutely
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.10 (158)
Todd Oakley and David Kaufer
immanent within the artifacts themselves. This close accounting of the text is then condensed into something akin to a map for the “interpreter” to read – Docuscope allows analysts to scale up the map to global vistas (i.e., an entire corpus of texts) or to scale it down to map very localized phenomena (i.e., a single text or words and constructions therein). After completing a surface scan of the corpus, we then conducted factor analysis of the results. Factor analysis (hereinafter FA) is a data reduction technique that collapses highly correlated variables into single composite variables, or factors. Docuscope’s strip-mining technique assigns different chunks of the surface text to 140 different variables. The size of the variable space is large and cumbersome for human interpretation. FA is a technique that prunes a large variable space into a smaller number of distinct factors, and these factors, importantly, are not correlated with one another. This means that factors are readily interpretable as independent rhetorical strategies that the authors of the clinical documents have adopted in the composition of their texts. As telegraphed above, the results of our factor analysis revealed the presence of three distinct rhetorical strategies present among these texts. Texts with the statistical profile of factor 1 follow a narrative strategy in which a narrator presents cases about challenges doctors face and the resistances they must overcome in dealing with disease and getting their patients to a positive state. We call this the challenge strategy. Texts with the statistical profile of factor 2 follow a dialogic strategy in which the physician engages in simulated dialogues with other medical professionals and/or medical students, teaching them about accepted procedures and anticipating their questions about what to do and when. We call this the consultation strategy. Texts with the statistical profile of factor 3 follow a descriptive strategy in which an “Olympian” voice presents diseases and research in a non-narrative and non-dialogic format. What is more, factor 3 texts are “presentist,” insofar as they suppress linguistic cues for time orientation (i.e., retrospection and prospection) as well as for interaction. In other words, texts with the statistical profile of this factor 3 suppress constructions that prompt narrative and dialogic representations, while texts with the statistical profiles of challenge and consultation strategies prime readers for these experiences. We call this the presentation strategy. Table 1 provides at a glance the factor analysis of language action type classes in their various concentrations. (A + sign means the class of LATs is highly active relative to the rest of the corpus; a – sign means the class of LATs is suppressed relative to the rest of the corpus.) These three factors represent the variational cleavage between dramaturgies of resistance, dialogue, and description. Whether or not the authors are consciously guided by these strategies is well beyond our powers of discernment. Docuscope brings within our powers of discernment a statistical set of regularities that fall out from the hundreds and thousands of recurrent, small, and implicit rhetorical
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.11 (159)
Chapter 5. Three layers of analysis
Table 1. Factor 1: Challenge
Factor 2: Consultation
Factor 3: Presentation
+ Anticipation +Positive Values +Pronouns (3rd Person) +Oppositional Reasoning +PositiveThinking –Word Picture
+First Person +Addressive +Contingent Reasoning +Dialogue +Interactivity – Negative Thinking
+ Inclusiveness +Directives – Time Orientation retrospection prospection
choices that were required to compose texts in the corpus. Keeping the placement of the micro-event under analysis within this macro family of dramaturgies will help discourse analysts keep in mind the larger significance of the micro-event to the entire text and corpora of texts in ways that are not readily available outside the statistical analysis. More importantly, the statistical analysis can aid in the noticing and selection of passages at the grammar layer analysis that have the highest probability of linking to artifact layer dramaturgies. Our analysis will now focus on six texts, three with the highest concentration in either factor 1, 2, or 3 and three with the lowest concentration in one of the same factors. These six texts account for the greatest range of variability within the corpus, providing independent justification for the choice of texts within the collection. This combination of texts would not have caught the attention of the analyst looking for idiosyncratic examples nor would they have been representative of the usual patterns of reading by the empirical readers – i.e., the physicians, who follow a selectional strategy based on disease/disorder type.
The artifact-layer: Mental space analysis of six clinical texts The following sections explicate these texts at the artifact layer of analysis. We consider this layer to cover the kinds of mental space networks operating in the rhetorical design for these clinical experiences as made manifest in complete artifacts.
Factor +1: This practice is sick The statistical profile of this text places it highest on the scale for factor 1, a narrative of the challenges facing practicing physicians, as evidence by the concentration of language action types for positive values (e.g., “providing effective treatment”) and oppositional reasoning (e.g., “don’t think a luxurious waiting room will compensate for long waiting times”). An examination of the text bears this out, for the reading experience is that of reading a third person account of the working life of Mike Hudson, the managing partner of Wallace and Associates, and the economic
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.12 (160)
Todd Oakley and David Kaufer
decline of the practice since the institution of “Managed Care” despite efforts to improve the practice.3 The most conspicuous feature mental spaces analysts will notice is the title. Quite distinct from other titles that identify specific diseases/disorders, this one is precisely the kind of text that attracted our attention as a bona fide conceptual blend based on metonymic compression. The entire practice is likened to a single sick patient, such that the experience of being sick – the very population this practice serves – comes to stand for the present economic state of the practice, a business enterprise. In this example, the presentation space projects schematic event structure associated with the appearance of an ill patient. (This arrangement diverges from the typical space delegation with the presentation space providing core knowledge of an underlying disease.) The reference space provides topical structures of the entire article: the medical practice of Wallace and Associates. This mental space contributes information about the history of the practice (established in 1963) along with its past working environment and financial profile. Anytime the writer seeks to provide background information about the practice, a version of the reference space comes into focus. The blended space represents the present state of the practice as being a sick patient. The conceptual blend prompted by the title has a direct influence on the lexical and grammatical structure within the text, however. The reader will not find phrases blending the domains of economics and business with sickness and medical practice, such as “anemic practice,” “contagious third-party payers,” “the new business plan is just the right antidote,” and so on. Nevertheless, the conceptual blend itself provides a schematic framework for elaborating on the case history of a practice, and we speculate that the choice in title was a clever means of fitting an economic topic within a genre constrained by the narrative structure of healthcare and medical practice. Once again, we stress that the blended space is most salient when readers think about the title. This in itself is an important analytic point, for it suggests to us that a conceptual blend can function as a structuring heuristic for the whole text, even though it does not in fact influence the lexical and grammatical choices comprising it. The meaning the blend is simply to “diagnose” the source of Wallace and Associates’ economic difficulties, taking the case histories of
. For readers unfamiliar with American Healthcare, “Managed Care” refers to systems that control the flow of health care services. A managed care organization is an institution that manages risk through contracts with health insurance providers, which are typically paid for through employers or patient groups. These insurance providers serve as “brokers” and “gatekeepers” between payers (employers or groups), providers (physicians) and patients. There are thousands of bureaucratically distinct managed care organizations.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.13 (161)
Chapter 5. Three layers of analysis
the physicians, staff, and patients into account, and offering a causal explanation for how the practice came to be so “unhealthy.” The topic of an unhealthy practice finds eager and attentive audience with readers of this journal, and the general implication readers are supposed to take from this case study is that medical practices are businesses operating in increasingly competitive environments: patients are customers who can choose to go elsewhere. More needs to be said about the narrative trajectory of this whole piece. The writer begins third person narration, with the managing partner as the focal character knee deep in paper work and unable to enjoy a Saturday afternoon with his family, as scenario many of readers can easily imagine for themselves. Astute readers can connect the scenario of a physician facing stacks of papers on a Saturday afternoon as a symptom of an unhealthy practice. Shortly thereafter, this opening vignette changes to include other vignettes told from the perspectives of other key players – colleagues, staff members, and patients – as is permissible when constructing third person narrator standing outside the action. That initial viewpoint is nevertheless the most salient, and readers are left with the impression that the multiple perspectives entertained throughout the case study are in fact presented to the managing partner as reports by external consultants hired to assess the practices weaknesses. The managing partner then remains a focal point of the narrative; it is his experience of his colleagues, of his work staff, and of Wallace and Associates’ clientele that comprise the center of interest and complicating action of the unfolding drama. Figure 2 depicts the mental space networks for this clinical experience. The blended space and the network of related spaces are consistent with the schematic structure of clinical experiences outlined in the previous section. Discussing a medical practice’s financial troubles can be easily integrated into a diagnosis and treatment scenario when the topic concerns “economic health,” an entrenched metaphoric blend of the healthcare and economic domains. We take this mental space network to be diagrammatic of the range of scenes and scenarios operating over the whole text, but the “work load” of each mental space is identical. As is also the case with other six texts, this text places a heavy burden on the reference space.
Factor –1: Chronic leg ulcers The statistical profile of this text suggests a low concentration of factor 1. Largely descriptive in its profile, this text is rich in language action types for directives (e.g., “apply multilayer compression bandages”) and sense property (e.g., “swelling in the legs”). Even so, this text does not fit comfortably into the presentation strategy, because it presents at least one instance of patient non-compliance, one of the most conspicuous topics associated with the resistances and challenges facing medical practitioners.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.14 (162)
Todd Oakley and David Kaufer
Figure 2. Mental spaces network for “This practice is sick”
In contrast to the previous text but in concert with the other texts in this sample, the reference space at this level of analysis can be likened to the physical examination and medical history of three patients: a 56-year-old male, a 76-yearold female, and a 53-year-old male. (The grammar layer of analysis zooms in on each case presentation, setting up so-called “daughter” spaces.)4 The presentation space contributes knowledge of the three disorders underlying limb ulcers: disorder of the arteries, disorders of the veins, and disorders of the nerves. (Remember, presentation spaces set up scenes through which one understands and interprets the referents.) The challenge for doctors is to make the proper diagnosis, because the etiology of leg ulcers will indicate contrasting, even contradictory treatment protocols. Thus, the presentation space corresponds to the attending physician’s knowledge of limb ulcers. The “blended” space corresponds in the activity of applying the appropriate diagnosis to the instance, and the elaboration space corresponds to the implications and precise treatments indicated depending on the underlying disorder. . Fauconnier (1994: 17–19) uses the same terminology but for different purposes.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.15 (163)
Chapter 5. Three layers of analysis
We wish to emphasize a point made previously. The space we are calling for convenience) the “blend” in this and most of the other texts are not to be considered conceptual blends, metaphoric or otherwise. Rather, it is more accurate to consider them simply as instances of integrating a type with a token. The person occupying the focal role in the reference space is a token seen through the prism of a type of condition known to affect innumerable others. Everything predicated of person X in the reference space will be set in terms of the conditions or range of conditions represented in the presentation space. In this case, the blend develops conceptual structure relevant to this person’s life as it relates to limb ulcers (all other aspects of person X’s life being incidental). The take-home message of this text (depicted in our diagram as the pragmatic implication running from the elaboration space back to the discourse ground) is that leg and limb ulcers are only superficially the same and that the attending physician has to pay close attention to the underlying causes of ulcers. The drama of the text corresponds most closely to a descriptive case history of each patient’s condition, where the focus of attention is not on the doctors and the challenges facing them, as evident in the previous example, but in the procedural details of treating ulcerations. As is typical of medical presentation, the focus of attention is most saliently placed on the relevant anatomy and physiology of each patient, in this case legs and feet. (In fact, this text provides pictures of patient’s legs and feet, some of which are close-ups with no clear visual orientation with respect to rest of the body.) In this version of the mental space network, the reference space dramatizes the verbal interactions between patient and doctor, evidencing slightly higher concentrations of interactive language action types. As with most case presentations, the doctor is not a focal participant; rather, he or she is presented as being an “offstage” observer. In this particular case, the 76-year-old woman “complained of pain at rest” as well as “denied intermittent claudication and diabetes.” More is to be said about the particulars of this mental space in the next section. For now, we note that the presentation of reported speech figures prominently as a rhetorical strategy in this text, but it neither plays a pervasive structural role beyond the presented case history nor is it specifically dialogue between physicians, a strategy described in the next section. For these reasons, the overall strategic profile fits within neither the challenge or consultation strategies. Figure 3 depicts the mental space networks for this clinical experience.
Factor +2: A practical approach to Atrial Fibrillation The statistical profile of this text places it highest on the scale for factor 2, a strategy based on simulating dialogue between practitioners. The reference space contributes four individual clinical cases (two female, two male) of Atrial Fibrillation (hereinafter AF), a dangerous and difficult to treat cardiac arrhythmia. The presen-
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.16 (164)
Todd Oakley and David Kaufer
Figure 3. Mental spaces network for “Chronic leg ulcers”
tation space presents facts about the nature of this disorder, most pressing of which being that 1) instances of AF are increasing as the population ages and that 2) current therapies are not efficacious, alleviating short-term symptoms may cause long-term complications. The presentation space, then, represents the state of the art knowledge in treating AF. In this space, the electrophysiologist with expertise in AF is willing to share his expert knowledge with non-specialists. The general impression is that the state of the art presents physicians with a “disconcerting” situation to contemplate. One will notice that diagnosis is not a salient part of the representational strategy, for the issue is not the diagnosis of AF but the variegated and confusing treatment options available to the attending physician. The blended space in our network represents case reviews in which a concerned and bewildered physician tries to make sense of a confusing welter of symptoms and treatment choices, prompting him to seek help from an electrophysiologist. The physician from the reference space and the electrophysiologist with expertise in AF in the presentation space meet each other in the blend and engage in a fictive dialogue about the treatment each patient received in the recent past. The elaboration space corresponds to the detailed question and answer review of each case. The time orientation lan-
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.17 (165)
Chapter 5. Three layers of analysis
guage action types, retrospection (thinking back to the case details, as in “I then examined. . . ”) and prospection (thinking ahead or thinking of hypothetical alternatives, “in such cases, you may need to determine if. . . ”) are prevalent throughout the text. Each version of the reference space is explicitly marked as occurring in the past perfect in relation to the blended space. The drama of the text takes place within variously in the reference space (past treatment), blend (present dialogue), and elaboration (possible treatments and procedures) spaces, and the take home point for readers is that they have to be vigilant about consulting experts in cardiology and related disciplines for the latest and best information on the treatment of AF. We posit that there are, in fact, two (daughter) versions of the reference space (which remains in focus much of the time): the space of actual past treatment represented from the Primary Care Physician’s (PCP) perspective; and the counterfactual alternate past reality represented from the Electropysiologist’s perspective. Conceptual structure from the counterfactual reference space projects structure to the elaboration space as hypothetical assessment and treatment scenarios (e.g., “if the patient has X symptoms, then do Y instead of Z”). Interactive questions like “would you have considered immediate cardioversion?” emerge out of the interaction between the elaboration and daughter reference spaces (diagrammed as dashed arrows in Figure 4). The pragmatic implication is that the clinical experience provides readers with the latest information on treatment and,
Figure 4. Mental spaces network for “Atrial Fibrillation”
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.18 (166)
Todd Oakley and David Kaufer
in addition, models for readers their own consultations with specialists in AF. Figure 4 presents the mental space network of the relevant scenes and situation needed to construct this clinical experience.
Factor –2: This practice is sick (Again) This text presents a statistical profile low in the consultation strategy. There are several instances in which the writers refer to arguments and disagreements among physicians, staff, and patients. But the key point is that none of these arguments unfold in real time with direct or free indirect discourse. Dialogue between the managing partner and other interested parties is referred to but not enacted, as it is in the previous text. As stated above, the narrative follows the perspective of Michael Hudson, the protagonist, who must react to the economic and clinical challenges facing Wallace and Associates, without direct representations of those challenging encounters. Factor +3: Outpatient management of community acquired pneumonia The statistical profile of this article fits the presentation strategy by focusing our attention on descriptive details of evaluating patients at the expense of presenting the challenges facing physicians or dialogues with practitioners. The aim of this text is to provide advice to the readers on choosing a “site of treatment” for patients suffering from Community Acquired Pneumonia (hereinafter CAP). Patients overwhelmingly resist hospitalization, and doctors likewise may gain economic incentives by not hospitalizing them. The writers wish to emphasize that both these factors may unduly and inappropriately influence the physician’s decisions. The article focuses on the site of treatment decision physicians have to make and offers guidelines for making those decisions that begin with physical examination, radiology, and blood tests. The mental space network for this article, depicted in Figure 5 below, is as follows. As with the previous examples, the reference space consists of separate cases of CAP: female (45); male (35); and a male (75) who died eight days after being admitted the hospital. The presentation space represents the evaluative task facing the attending physician. The blended space represents a case-by-case integration of the two spaces that dramatizes the proper means by which physicians should base her decisions. The elaboration space zooms in on the particular facets of physician’s decisions. In the initial evaluation, the physician asks himself/herself several questions about the patient and draws implications from each answer. Interestingly, this scene in the elaboration space should count as a dialogic strategy; however, there is no dramatization of actual question and answer sessions, as one would expect had the writer’s pursued a consultation strategy. The linguistic material for designing this clinical experience comes from the sense property language action type, as exemplified by “asymmetric expansion of the chest wall,” and “dull-
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.19 (167)
Chapter 5. Three layers of analysis
Figure 5. Mental space network for “Community Acquired Pneumonia”
ness to percussion.” It should be noted that all the texts in this corpus exhibit a high concentration of word picture language action types (classes related to external descriptions, not interactions). In fact, this text ranked fourth highest in the word picture category, leading us to conclude that factor 3 is defined just as much by the inhibition of narration, interaction, and retrospection classes as it is by the positive activation of any specific set of external description classes. Consistent with the other mental space networks presented above, this diagram presents a network of mental spaces manifesting type-token integration as the primary function of the blend. In the blended space, the physician evaluates “this person suffering from pneumonia.”
Factor –3: Non-invasive diagnosis of Pulmonary Embolism In lieu of an explicit mental spaces analysis of this text, we wish to focus instead on tits statistical profile, as its peculiar profile leads us to conclude that these texts need not follow a dominant rhetorical strategy at all. The statistical profile of this text places it lowest on the scale for the presentation strategy, an odd finding given its overall structural similarity to the previous one. As before, the physician is not functioning as a protagonist (as with factor 1), nor does she engage in dialogue with other physicians (as with factor 2). As before, there is an Olympian voice that
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.20 (168)
Todd Oakley and David Kaufer
relates particular facts of each case and then follows up with general discussion. The drama (or lack thereof) in the blended space focuses on the improvements in the procedures for diagnosing suspected cases of Pulmonary Embolism. The takehome message of this experience is that physicians now have new, non-invasive procedures at their disposal. This point fits with the general ethical imperative that physicians should avoid invasive procedures whenever possible. So, why is this text low in factor 3? The reason becomes apparent when we look closely at the numbers. While this text follows the general pattern found in the article on CAP, it does not suppress retrospection classes, suggesting that the presentation strategy is really defined by the absence of retrospection as much as by the presence of other language action types. This contrasts with the first two factors, in which the strategy is developed by a concentration of the language action types for past events and shifting events are prevalent in texts following the challenge strategy and a concentration of the language action types for interactivity are prevalent in texts following the consultation strategy. A closer look at Docuscope’s findings reveals that this text ranked twelfth among texts for factor 1 and is ranked fifth for factor 2. In some respects, this finding leads to the conclusion that this text does not really follow a consistent strategy. It may be more apt to regard it as a hodgepodge of tactics cutting across all three factors. But absences are often more revealing than presences. Since this text shares many overlapping types, the inhibition of the retrospection, prospection, and interactivity types suggest to us that types one and two are definitive grammatical contributors to the challenge strategy while third is a definitive grammatical contributor to the consultation strategy. Furthermore, the presence of retrospection types in any concentration marks the surface structure of the text as contrary to the third strategy. Writers using this ‘hodgepodge’ seem to be avoiding certain grammatical constructions.
The grammar layer: How to build mental spaces with words and phrases Sampling criteria: Novelty, commonality, and innovation In the previous section, we offered an analysis of each text, with a delegation of mental space types thematically related to the particular content of each clinical encounter as interpreted through the strategic profile of the three factors, thereby aligning artifact analysis and genre analysis. In this section, we wish to conduct analysis of specific sentential examples from these texts, relying, as suggested above, on the distinction between parent and daughter spaces. This nomenclature provides a means of marking the hierarchical relation between two granules of analysis: the granule of whole scenes and situations (parent) functioning across
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.21 (169)
Chapter 5. Three layers of analysis
instances of words and phrases, and the granule of salient facets of scenes and situations (daughter) structured by specific expressions. Our aim is, once again, to bring lexical and grammatical analysis in alignment with the genre and artifact layers of analysis. We consider these five examples from the corpus: (1) But finding out what patients want as they are abandoning the practice is like closing the barn door after the horse is out (Weeks 1998).5 (2) (2) A 24-year-old woman called on her family physician after awakening with a sharp pleuritic pain (Perreir 1998). (3) (3) A 75-year-old woman presented with right foot ulcers. She had had the ulcers for three months and said that they have been precipitated by shoe trauma. She complained of pain at rest, but denied intermittent claudication or diabetes (Bello 2000). (4) A 75-year-old man presented with a dry cough, shortness of breath, and mild confusion of three days’ duration. (Fine et al. 1998). (5) Because the patient had white coat hypertension, she may have seemed to be less well controlled in the office than she actually was (Olshansky et al. 1999).
We chose these five examples on the basis of expressive novelty, expressive commonality, and expressive innovation. These three criteria speak to complementary impulses of mental spaces and genre theorists alike. For mental spaces theorists, novelty reveals the creativity of minds in responses to local conceptual and expressive contingencies. For genre theorists, novelty may reveal the “pliability” of a discourse type, for novelty can be a measure of the extent to which generic schemas can or will accommodate new forms and still be considered the same genre. For mental spaces theorists, commonality signals stability in the means and manner or construing scenes and situations, and likewise for genre theorists, commonality exposes the typified patterns in response to recurring situations. For both mental spaces theorists and genre theorists, innovation reveals those points in which a common form is altered to fit the expressive contingencies of a situation; innovative expression balances assimilation and accommodation. The first example is unique to the corpus. No other instance of such comparisons in the form of “folk wisdom” appear anywhere in the corpus. The second example is also novel to the corpus and represents a shift to an informal register within the discourse. These examples satisfy the criterion of expressive novelty, insofar as their introduction into the text signals a marked shift to an informal register otherwise atypical of this genre. In contrast, examples (3) and (4) are commonly found throughout the corpus. We found 8 instances of this exact wording . Italics added.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.22 (170)
Todd Oakley and David Kaufer
and 14 with some variant thereof, such as presents with or presenting with. These examples satisfy the criterion of expressive commonality, a conventionalized “tactic” for building a mental space type obligatory for the medical case study genre. Example (5) represents a unique construal of a common concept in the medical domain: the White Coat Effect. This example satisfies the selection criterion of expressive innovation, as it alters a common expression among the discourse community for local purposes.
Analysis Sentence (1) closes the “patient perspective” section of the article “This Practice is Sick” by presenting a simile that may, in fact, be offensive to outside readers but intended to highlight through concrete imagery the difficult task of persuading patients to return to practices with the same kind of problems as Wallace and Associates. On analysis, we posit that this simile is a conceptual blend issuing from the reference space at the artifact-layer. This reference space contains struggling medical practices, for which Wallace and Associates is the illustrative example. The daughter reference space contains the relevant facet of a practice in economic decline, namely patients abandoning the practice at the same time that doctors and consultants trying to figure out why they are leaving. The presentation space contains an entirely novel scene from a different domain: farming, ranching, and animal husbandry. In this space, a horse is outside the barn and the farmer/rancher has closed the door. The implication is the task of trying to corral the horse back into the barn is exceedingly difficult if not impossible. The blended space comprises a very loosely layered analogy, such that the patient is being explicitly compared to the horse without being identified with it. The blend projects the following pragmatic implication back to the discourse ground: “patients who are already abandoning your practice are like wild horses, and your practice is like a barn with its doors closed; just as it is too late to get the horse back in the barn, it is too late to get the patient back into the practice.” The moral of the comparison is that you should focus on improving services for existing patients. Figure 6 diagrams the relationship between the parent and daughter spaces in relation to the general mental space structure. This example highlights a possible distinction between strategic blending and tactical blending, with the former exemplified by the text’s title and the latter exemplified by this simile. This is important because it shows that specific local conceptual blends serve expressive functions may have a different function and relevance in relation to how readers consign importance to any given expression. At the artifact-layer, the analyst can see that the strategic conceptual blend provides allegorical structure to the whole text and tactical blends invoked for local expressive purposes, along with the potential tensions, dissonances, or clashes it may produce (especially with non-physician readers). Thus, there appears to be
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.23 (171)
Chapter 5. Three layers of analysis
Figure 6. Mental spaces analysis of the horse-patient analogy
strategic blending and tactical blending. To date, mental spaces theory has only alluded to this distinction. This example at the grammar layer highlights for us the need for systematic investigations of strategic and tactical blending in discourse. Sentences (2)–(5) appear in the case presentation sections of the other texts and correspond to the parent reference spaces in this analytic mode of analysis. Therefore, the focus of attention will be on the variations of daughter reference spaces in each instance. Sentence (2) opens the first case presentation in “Noninvasive Diagnosis of Pulmonary Embolism.” Recall that this text is negative for the presentation strategy and similarly weak in language action types for the challenge and consultation strategies. Sentence (2) frames the examination as a dialogic encounter, whereby the patient “calls on” the physician. This open is distinct from all others in the collection. The effect, though, is to present the doctor/patient dyad as informal and familiar, thus profiling the close and supposedly long-standing relationship between the participants. This homey scene is typically associated with family practices, and the implication is that the patient is being treated in an office rather than hospital setting, this in contrast to other examination episodes with presented with distinctly “institutional” flavor. This presentation strategy takes on rhetorical significance when contrasted with the second case of a 70-year-old man undergoing a post-operative examination in a hospital setting. The writer opens this case presentation thus: A 70-year-old man underwent surgery for total hip replacement and was placed on a standard thromboprophylactic regimen. . .
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.24 (172)
Todd Oakley and David Kaufer
Figure 7. Mental spaces analysis of reported speech
The writer chooses to forego any representation of dialogue between physician and patient; in fact, there is no hint that the patient is even aware of his condition. The contrasting dramatic representation of these two scenes of examination correlates with the contrasting clinical outcomes. The doctor rules out pulmonary embolism using a few noninvasive diagnostic techniques in the first case, whereas the doctor confirms pulmonary embolism via the full complement of diagnostic techniques, including the invasive angiography, in the second case. In contrast to (2), sentences (3) and (4) represent a very common way of introducing case studies in this genre, as can be witnessed with the appearance of the verb-preposition collocation, presented with. In both cases, the perspective aligns with the physician, as would be intuited. The physician perspective can be considered as the “canonical viewing arrangement,” as discussed by Langacker (1987). The physician is an “off-stage” observer of the patient’s physical condition, as depicted in Figure 7 above. This figure represents an experience of patient examination, with the circle labeled “V” (for “Viewer”) corresponding to the semantic role aligned with the physician and the circle labeled “Ex” (for “Experiencer”) corresponding to the semantic role aligned with the patient undergoing examination. The rectangle encompassing Ex corresponds to the “stage.” Imagine if you will a stock dramatic situation in which a human actor stands on stage while an audience member intensely observes her, drawing spe-
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.25 (173)
Chapter 5. Three layers of analysis
Figure 8. Mental spaces analysis of reported speech
cific conclusions about her physical condition. The canonical viewing arrangement does not by necessity a veridical representation of the situation, as many acts of examination entail direct contact between doctor and patient, placing both onstage together. The point here is that the phrase presented with focuses attention on the relevant part of the patient’s body as seen from a distal vantage point. In every instance we examined, this phrase appears with complement noun phrases representing signs and symptoms visible to the physician at some distance (often accompanied by sensory state adjectives, such as painless or aching made available through patient reports). We have not seen any instances of complement noun phrases representing signs elicited from direct, instrumental examination, as in a 70-year-old male presented with hypertension. Such expressions, though grammatically correct, appear to be pragmatically anomalous. Sentence (3) quickly shifts out of this canonical viewing arrangement. In addition, the presentation space for this scene of examination shifts to a dialogic arrangement, as outlined in Figure 8. The appearance of quotative verbs say, complain, and deny presents a dialogic situation as viewed from the past relative to the examination and diagnosis scenes. Notice that the viewing arrangement depicted at the top of Figure 9 places two entities onstage, with the top bolded circle labeled “Sp” (for speaker) and with the bottom unbolded circle labeled “Hr” (for Hearer). In this case, the Sp corresponds to the patient and Hr corresponds to the physician. The bold outline signifies the
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.26 (174)
Todd Oakley and David Kaufer
elicited perspective. In the top facet of this examination scene, the focus of attention is on the patient as the speaker with the reader having little regarding a phrase like complained of pain at rest as an accurate, possibly verbatim account of the patient’s own utterances as indirect reported speech. The same is not the case with denied intermittent claudication, as the wording replicates the examination in medical jargon (claudication being the technical term for pain in the foot an calves induced by walking). In our analysis of sentence (3), the reader constructs different facets of the same reference space in rapid succession. The mental space is first structured by a canonical viewing arrangement of visible signs, followed by a dialogic arrangement in which the physician is onstage but the profiled participant is the patient communicating her own experiences, followed again by the same dialogic arrangement, but this time the profiled participant is the physician speaking in a medical paraphrase. In this example, the perspective shifts from being directly “inside” the reference space to “inside” the discourse ground, with the grounding now serving as the viewpoint space from which mental access to content of the reference space becomes salient in the reader’s imagination (see Fauconnier 1997: 44; and Hougaard & Oakley this volume). This granularity of analysis reveals two things. First, that a particular phrase, common to this genre, is a useful construction for presenting particular kinds of signs and symptoms available at a distance. Second, that this viewing arrangement can easily shift to a participant arrangement, with the physician and patient both appearing onstage. Once this shift occurs, writers can generate clinical experiences profiling the patient’s words and experiences inside the reference space or profiling the physician’s interpretations of the patient’s condition, thereby reestablishing a discourse perspective mirroring the practitioner-to-practitioner relationship established in the discourse ground. We now have identified three distinct tactics for organizing variations of the parent reference spaces in relationship to other mental spaces in the network.6 There is another way to represent the examination scenario exemplified in sentence (5). Appearing midway through the dialogue about the second case of AF, these words correspond to the comments of the electrophysiologist and refer directly back to the examination scene. Figure 9 presents a mental space microanalysis in relation to its parent network.
. We have not yet done so, but we plan to reexamine the corpus to see the precise arrangement of these tactics throughout the collection. Is following a canonical viewing arrangement→ participant arrangement pattern, where a physician views the patient from “offstage” before she or he becomes an onstage participant in the dialogic arrangement, a typical tactic for building up these mental spaces or do we find merely a series of equally good options?
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.27 (175)
Chapter 5. Three layers of analysis
Figure 9. Mental spaces analysis of white coat effect
Sentence (5) introduces a well-known concept in the medical practice, known as the White Coat Effect, the Uncertainty Principle of medical examination. The logic goes something like this: the very act of examination by a medical practitioner in a clinical setting produces anxiety that, in turn, produces the very signs and symptoms in the patient. We get this phrase from metonymic compression (see Fauconnier & Turner 2002: 312–325), where the standard uniform (white coat) associated with physicians captures the causal structure of the situation – seeing the white coat causes anxiety which elevates blood pressure and increases respiration rates, and other apparent signs of coronary distress. In this instance, the noun phrase white coat hypertension refers to the elevated blood pressure in the patient as she undergoes examination. As Figure 10 depicts it, the daughter space construes the doctor/patient examination scene as less observational than causal, depicted by the circles-and-arrow figure to the right of the mental space. In contrast to previous tactic in which the physician either views the patient from offstage vantage point or appears on-stage in a verbal exchange, the physician takes on the semantic role of “Ag” (agent) effecting change in the semantic role of “Pt (for patient, depicted by the sigmoid arrow inside “Pt.” This facet of the examination scene links up with the daughter elaboration space, representing a specific conversational turn in the consultation between the attending physician and electrophysiologist. The final representational tactic deployed at the grammar layer is to represent the physician/patient examination as causal. The selected samples for analysis show how specific linguistic constructions create subtly different scenic arrangements among the same types of mental spaces capturing design options at the artifact layer, which, in turn aligns with generic design options.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.28 (176)
Todd Oakley and David Kaufer
Conclusion While traditional mental spaces theory uses the surface of language as “input” for interpretative analysis, most of the analysis that takes place abstracts from that input in a “destructive” process, cleaving specific surface features from the textual cloth, as happens with all acts of interpretation but especially with linguistic analysis. Therefore, an elemental problem for interpretative frameworks like mental spaces is to understand how the object of interpretation must be deformed, perturbed, or destroyed in order to transform linguistic “inputs” into meaningful “outputs.” The limitation of mental spaces theory is that recurrent patterns immanent in the surface are not placed in productive alignment with the dramatization of scenes and situations enacted at the higher-order strata of analysis. The three-layered approach to text interpretation, wherein each layer captures important facets of reader’s experiences with texts, is our attempt to mitigate textual destruction. In addition, we think this three-layered approach provides a blueprint for combination of mental spaces theory and statistical techniques that harvest the surface of in ways that complement lexical and grammatical analysis. These statistical techniques can benefit linguistic analysis without over-determining it. Taking into account statistical variations across an entire corpus allow genre theorists to show how clinical accounts can be segmented into distinct overall rhetorical dramaturgies at the genre and artifact layer of analysis, and that the these accounts can aid discourse analysts in applying the selection criteria of expressive novelty, commonality, and innovation at the grammar layer by pinpointing without prior bias a motivated relationship to the larger sample of language and discourse under investigation. We propose that the account given of the meaning construction processes involved in each text at each layer of analysis is appropriately exhaustive and satisfies, at least in part, the cognitive linguist’s aim of understanding the concept patterns used in setting up the scenes and situations that bring these clinical experiences to life for their readers. We likewise propose that the statistical profile produced by Docuscope satisfies, at least in part, the genre theorist’s aim of providing proper selection procedures for ensuring that the textual artifacts so analyzed and interpreted are in fact representative of the genre features present throughout the entire collection. Those “odd ball” samples prized by mental spaces theorists become more interesting as their relationship to whole texts and collections of texts comes more sharply into view.
TSL[v.20020404] Prn:11/12/2007; 15:38
F: PB17005.tex / p.29 (177)
Chapter 5. Three layers of analysis
References Bawarshi, A.S. 1997. “Beyond Dichotomy: Toward a Theory of Divergence in Composition Studies.” JAC: Journal of Advanced Composition Theory 17(1): 62–71. Bello, Y and Phillips, T. 2000. “Chronic Leg Ulcers: Types and Treatment.” Hospital Practice 34. Bitzer, L. 1968. “The Rhetorical Situation.” Philosophy and Rhetoric 1 (1): 1–14. Brandt, L. and Brandt, PA. 2005. “Making Sense of a Blend.” In Annual Review of Cognitive Linguistics. R. Mendoza Ibáñez & F. José (eds.), 216–249. Amsterdam: John Benjamins Press. Brandt, L. In Preparation. “Conceptual Integration.” Unpublished Manuscript. Coulson, S. and Oakley. T. 2000. “Blending Basics.” Cognitive Linguistics 11(3/4): 175– 196. Devitt, A.J. 1993. “Generalizing About Genre: New Conceptions of an Old Concept.” College Composition and Communication 44 (4): 573–586. Fauconnier, G. 1994. Mental Spaces: Aspects of Meaning Construction in Natural Language. Cambridge & New York: CambridgeUniversity Press. Fauconnier, G. and Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Fillmore, C. 1992. “Corpus Linguistics or Computer-Aided Armchair Linguistics.” In Directions in Corpus Linguistics, Svartvik, J (ed.), 35–60. Berlin: Mouton de Gruyter. Fine, M.J., Chowdhry, T, and Ketema, A. 1998. “Outpatient Management of CommunityAcquired Pneumonia.” Hospital Practice 32. Gallagher, S. 2004. “Hermeneutics and Cognitive Science.” Journal of Consciousness Studies 11 (10/11): 162–174. Kaufer, D., Ishizaki, S., Butler, B., Collins, J. 2004. The Power of Words: Unveiling the Speaker and Writer’s Hidden Craft. Mahwah, NJ: Laurence Erlbaum. Krishnamurthy, R. 2001, April 27. “Corpora: Chomsky and Corpus Linguistics”[Msg 16]. Message posted to http://torvald.aksis.uib.no/corpora/2001-2/0065.html Labov, W. and Waletsky, J. 1967. “Narrative Analysis: Oral Versions of Personal Experience.” In Essays on the Verbal and Visual Arts, J. Helm (ed.), 12–44. Seattle: University of Washington Press. Langacker, R.W. 1987. Foundations of Cognitive Grammar, Volume 1. Stanford, CA: Stanford University Press. Meyer, C. F. 2002. English Corpus Linguistics. Cambridge, UK: Cambridge University Press. Miller, C. 1984. “Genre as Social Action.” Quarterly Journal of Speech 70: 151–167. Nelson, K. 1996. Language in Cognitive Development: The Emergence of the Mediated Mind. Cambridge & New York: Cambridge University Press. Oakley, T. and Coulson S. This Volume. “Connecting the Dots: Mental Spaces and Metaphoric Language in Discourse.” In Mental Spaces in Discourse and Interaction, T. Oakley and A. Hougaard (eds). Amsterdam and Philadelphia: John Benjamins Press. Olshansky, B. and Sulo, R. 1999. “A practical Approach to Atrial Fibrillation.” Hospital Practice 33. Perreir, A. 1998. “Noninvasive Diagnosis of Pulmonary Embolism.” Hospital Practice 32. Polinghorne, D. 1988. Narrative and the Human Sciences. Albany: State University of New York Press. Weeks, W.B. 1998. “This Practice Is Sick: A Teaching Case.” Hospital Practice 32.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.1 (179)
chapter
Compression in interaction Anders Hougaard Center for Social Practices and Cognition, Institute of Language and Communication, University of Southern, Denmark Keywords: blending, compression, talk-in-interaction, long term working memory and interactional memory
. Introduction Fauconnier and Turner (2002a) claim that compression of connections (“vital relations”) within and between mental spaces is a major goal of conceptual blending, especially since it serves to produce so-called human scale products which allow human beings to comprehend comprehensive or complex matters in terms of simpler representations. It is for example a compression to say that “You are reading what I have written.” It is a massive compression of all the intermediary electronic stages, processes and transformations that the document (or screen image) you have in front of you has undergone. This comprehensive process is construed linguistically in the same way as if I had just written on and then handed to you the pages that you have physically in front of you. The world is a complex place, but human beings are phenomenal at handling this by representing it in simplified ways, by means of compression achieved in conceptual integration networks. That seems in essence to be the claim made by Fauconnier and Turner (2002a). (See also the introduction to this volume for a brief discussion of compression.) This claim has been supported by a wealth of analyses in the blending literature, including Fauconnier (2005), Fauconnier and Turner (2000 & 2002a), Hougaard (2005) and Turner (2006). So far, however, compression has only been dealt with as an analytic or modeling concept (i.e. a type of “mental space mapping”), which nonetheless is claimed to capture purported, hidden, cognitive mechanisms involved in individuals’ understanding or creation of meaningful items across very diverse cognitive activities (from language use to art).1 It still remains to be studied, though, if compression is a real phenomenon to . Fauconnier’s and Turner’s 2002 book has the subtitle “Conceptual Blending and the Mind’s Hidden Complexities”.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.2 (180)
Anders Hougaard
sense-making human beings in actual situations of meaning construction. By employing a method not used before in connection to cognitive semantics (except in prior work presented in Hougaard (2004 & 2005) and similar work in Alaˇc (2005) and Williams (2005) and this volume), I would like to suggest that a sort of compression, understood as a shared, visible, enacted, interactional process – not a hidden process – whereby sense-making human beings achieve a certain type of condensed representation and understanding of their talk, may indeed be a very real, concerted mechanism for sense-making interlocutors engaged in talkin-interaction. The data presented below consists of excerpts of talk recorded from an American call-in show on the San Francisco Bay Area radio station. The excerpts are examples from a collection in progress. The method used is roughly to qualify and constrain “cognitive accounts” on the basis of emic, micro-sociological analyses in a vein that is inspired by conversation analysis (hereafter CA), although no full-blown CA phenomena are presented.
.
Data & method
Conversation analysis (CA) developed in close connection to ethnomethodology (EM) in California in the 1960’s. The goal of EM founder Harold Garfinkel (1967, 2002) was to study the methods and common sense (“public knowledge”) that ordinary human beings apply when making sense in ordinary-for-the-participants settings. EM itself is not a particular method, but rather a particular way of thinking about social organization and everyday, social sense-making. Studies in EM represent a great diversity in terms of topic and method. Yet all are concerned with the orderliness achieved in everyday social life. Crucially, EM does not consider the details of everyday social action to be coincidental or chaotic, nor is the order of social life seen as encoded in members’ minds or pre-established by situational contexts. Instead order and sense is turned into a procedural phenomenon which is studied in terms of how they are achieved in and through temporally contingent actions. In the same vein, Harvey Sacks (1992) set out to study how interlocutors achieve intersubjective sense during the specific but seminal social activity of talk-in-interaction and what common sense knowledge they rely on. By talk-in-interaction is meant only naturally occurring conversation – that is, conversations in real life that would have taken place whether or not the researcher had filmed or taped them. Importantly, CA research focuses not on what interlocutors may mean, think, believe or intend when they say or do something during interaction; it focuses on the sense which the interlocutors themselves orient to and accomplish in and through their concerted, interactional actions. Conversational conduct is viewed as methodical and therefore as analyzable in terms of the participants’ own achieved actions and senses. In this sense CA is strictly obser-
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.3 (181)
Chapter 6. Compression in interaction
vational and emic. Thus by employing CA assumptions and method for studying the sense-making which interlocutors orient to, this paper takes on an approach which differs radically from most research in cognitive semantics. This does not just lead me to constrain my studies to include only public aspects of cognition while disregarding what may be considered the most important processes taking place below the horizon of consciousness, far beyond the public scene. Instead this paper takes issue with the notion of “unconscious processes” as a ghost that has inhibited progress within cognitive semantics by rendering the most essential parts of its theories as hypothetical modeling of an inaccessible locus of thought. What this paper proposes is a rethinking of the concept of cognition within conceptual integration theory and cognitive semantics at large that does not ascribe its most essential parts to ghosts. It rejects the traditional “backstage” (Fauconnier 1994) focus of blending theory and replaces it for a focus on the work that participants can be observed to do. One might say that it endorses an “onstage” focus. For an elaborate introduction to the notion of cognition as a phenomenon of interacting bodies, I direct the reader to Gitte R. Hougaard & Anders Hougaard (forthcoming). With this study I also take a path that is different from another, influential path in present-day cognitive semantics and cognitive science at large: the neural or neurally informed study of cognitive processes as for example carried out within the discipline of cognitive neuroscience. Before moving on to the study itself, I will therefore briefly delve at the prospects of a neural cognitive semantics in order to specify further how this paper positions itself in the field of communication and cognition. Perhaps expressed most forcefully by George Lakoff (in for instance Gallese and Lakoff (2005) and Lakoff (2007)), neural studies of cognition are believed by certain contemporary cognitive semanticists to be the only appropriate way to certify theories that have been developed over the past two-three decades and to give them an interpretation as “real,” biological processes. Furthermore, according to researchers such as Lakoff, the linguistics theories must first and foremost answer to knowledge of the brain accumulated within the neurosciences. Following such trends in cognitive semantics, one might perhaps argue that the best way to approach the issue of the reality to people of purported phenomena like compression is via for instance cognitive neuroscience methods (however such investigations would be designed to shed light on compression), such as fMRI and ERP (cf. Coulson and van Petten (2002), Rohrer (2001b), Lakoff and Johnson (1999) and Lakoff (2007)) or neural computation (Narayanan 1997).2 These methods seem to bear the promise of providing hard biological evidence for or . For a review of ERP studies of cognitive semantics theories, see for instance Coulson (2007). For a study of Johnson and Lakoff ’s embodiment theory using ERP and fMRI, see for instance Rohrer (2001b)
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.4 (182)
Anders Hougaard
indicating the neural plausibility of cognitive semantics theorizes. Yet, there are at least two important caveats here: 1) Neuroscientific and neurologically inspired studies do not provide definite results. There are substantial limits to the insights we can achieve. For instance, cognitive neuroscience methods such as fMRI (which is often mentioned as complementary to ERP measures) can only give rough, labbased, superficial impressions of the individual brain at work; for one thing it does not produce actual images of neural activity but of blood concentrations. Neural computation does not provide any direct evidence, and since computations remain “models” or “imitations” we can never be certain that what an artificial neural network can do and cannot do is the same as what the brain can do and cannot do. Besides as Werner (manuscript) observes, the concept of neural “computation” is not an uncontroversial concept in neuroscience. Furthermore, in an age where phrenologist thinking is still with us (Uttal 2003), alternative views on brain function and cognition are not firmly established. Neuroscientists are still struggling to figure out how neurons process cognition and behavior organized in circuits, groups, systems and levels (Damasio 1994, Bates & Dick 2000). Finally, a too straightforward coupling of linguistic/psychology/behavioral/analytic studies and insights from neuroscience must be resisted. One of the major goals of Jerome Feldman and George Lakoff ’s Neural Theory of Language Project is to combine the “level of ” linguistic analysis and a number of intermediary levels all the way down to the “level of ” neurophysiology (for an overview go to http://www.icsi.berkeley.edu). And already now, neural interpretations of analytic linguistics terminology have been proposed by the NTL (for instance blending as “neural binding”, Lakoff 2007). Yet, thorough, methodological, theoretical and terminological discussions are warranted in that connection (see for instance Rohrer (1999) and Adolphs (2003)). 2) In its present form, the neuroscience endeavor in cognitive semantics carries on the tradition of focusing on the cognitive work of individuals in relation to isolated sentences. Therefore in terms of their current design, cognitive neuroscience studies have little to do with how humans construct meaning in natural, full-fledged, social contexts. And if neuroscience based/inspired studies in cognitive semantics continue to merely focus on testing insights from analytic studies of individuals’ cognitive work in relation to isolated sentences, the endeavor risks becoming simply self-fulfilling: There is a great risk for instance that the analytic results will determine the interpretation of neural findings and thus define their own neural verification. Consequently, this paper can also be read as making a statement of caution with regard to the kind of neuro-optimism – strongly represented in for example the NTL – which is currently strong within certain areas of cognitive semantics, relying on neuroscience or imitations of neural activity to give definite, hardwired substance to analytic theories. The study of cognitive processes of sense-making
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.5 (183)
Chapter 6. Compression in interaction
needs alternative, socially anchored, observational methods to challenge both its analytic and its individualistic, neuroscientific focus. Another approach that may be undertaken to test the psychological or social reality of compression is through controlled experimental tests in specialized lab settings. Such tests might allow us to get a pure and close look at mechanisms which either would not or only rarely would appear in naturally occurring interaction or which would simply be unobservable. Furthermore, they could probably also generate the unfolding of human capacities that are not at play in naturally occurring interaction. Yet, such studies would thus not document actual, naturally occurring behavior, even though experiments may indeed be very reallife-like. Hence while experiments can give us insights that cannot be obtained from naturally occurring data, they cannot give us direct insight into the focus of this paper: the role of compression in the real lives of sense-making, social human beings. In the following, I will first walk the reader through interactional analyses of the Bay Area Radio data, then discuss the “cognitive implications” of the interactional analyses with respect to 1) the notion of compression and 2) memory and psychology in interaction. The analytic terminology used in the following paragraphs is mostly ordinary CA terminology, which will be explained along the way. An appendix of transcript conventions is appended at the back of the article.
. Packing up turns at talk The excerpts focused on in this section are recorded from a call-in show on the San Francisco Bay Area Radio on the eve of coalition air strikes, lead by US forces, on Baghdad in early 1991 (The First Gulf War) in response to Iraq’s invasion of Kuwait the year before.3 Occasioned by these attacks, anti-war demonstrations took place in several major American cities, some of which escalated into violence and rioting. Host on the Giant Sixty Eight KNBR Leo Laporte is taking calls from listeners who wish to express their view on the situation, from tactics of war in Iraq to fear of encountering terrorism at home in the USA.
. The First Gulf War was a conflict between Iraq and a coalition force of about 30 nations, led by the United States, which was mandated by the United Nations to liberate Kuwait. The lead up to the war began with an Iraqi invasion of Kuwait on August 2, 1990. The invasion was met with economic sanctions by the United Nations against Iraq. Sanctions were replaced by coalition attacks on Iraqi troops in Kuwait and Iraq and on “strategic targets” in for instance Baghdad, the capital of Iraq. The result was a decisive victory for the coalition forces, which drove Iraqi forces out of Kuwait with minimal coalition losses.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.6 (184)
Anders Hougaard
A particular interactional action often seen in these calls is that Laporte responds to turns at talk by callers with a phrase that sums up, or packs up, as I call it, these turns and precloses the call–that is, it is understood as being an action that can be followed by a closure sequence. 4,5 Item C#11 from the collection constitutes this general structure. In lines 1–17 the caller, GE, reports his experience of the anti-war demonstrations, expressing his regrets about the way in which the few demonstrators who riot also give a bad name to the majority of demonstrators who are peaceful (ll. 1–17). In lines 18 and 20 Laporte sums up this report, receives agreement by GE (in lines 19 and 21) and concludes the talk in line 22: (Excerpt 1, C#11)6 1GE: 2GE: 3GE: 4GE: 5GE: 6GE: 7GE: 8LL: 9GE: 10GE: 11GE: 12GE: 13GE: 14GE: 15GE: 16LL: 17GE: 18LL: 19GE:
now i wanna proTEST; (-) people who MARCH a::nd uh crEate vIolence;= <
and i think you’re exACtly rIght;> .hh uh y your rePORTS <<monotonous> on your stAtion say that there are about ten thousand marching and a fEw hUndred [(-) broke off and did those things so; [yeah; .hh i ↑HOPE that we don’t uh lUmp everybody together and say that; (.) PROtesters A:re violent people;= they (.) they DON’T WANT the KILLing to contInue just like myself; .hh and we’re VEry much against people that uh; STOP TRAffic an an START FIRES. [yeah [< and do other things.> a lot of BRIGHT intElligent people of CONscience who a:re; yeah;
←
. CA is in not interested in language per se. It is interested in whatever resources interlocutors rely on to construct shared meaning. Hence language is interesting from a CA perspective with regard to how it is used, as a resource, to achieve interactional accomplishments together with other resources. The focus of CA studies is thus on interactional actions. . The relevant unit in terms of which actions can be analyzed is the sequence, not single actions. It is through sequences consisting of combined actions that single actions acquire an interactional meaning. . The data used in this paper has been collected and transcribed by Elizabeth Couper-Kuhlen. The transcript conventions have been developed by Selting et al. (1998) on the basis of a system developed by Gail Jefferson (http://talkbank.talkbank.org/ca/gail.doc). A glossary of the transcript conventions is appended at the back of the article.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.7 (185)
Chapter 6. Compression in interaction
20LL: OUt there uh mArching because they feel that they HAVE to. 21GE: i am GLAD you brought that up. 22LL: .hh i THANK y’ for the cAll. (9Gulfwar.ca, ll. 48–69)
←
It should be noted that even though the examples presented in this paper have not been examined specifically for their resemblance to established interactional phenomena, they do bear resemblance to, and may all turn out to be, a subclass of a phenomenon studied by Heritage and Watson in a famous article from 1979. Heritage and Watson’s work is in turn based on Garfinkel and Sack’s (1970: 350) more general observation that A member may treat some part of the conversation as an occasion to describe that conversation, to explain it, or characterize it, or explicate, or translate, or summarize, or furnish the gist of it, or take note of its accordance with rules, or remark on its departure from rules. That is to say, a member may use some part of the conversation as an occasion to formulate the conversation.
Garfinkel and Sacks (ibid.: 351) speak of such practices of “saying-in-so-manywords-what-we-are-doing” as “formulating,” and Heritage and Watson (ibid.: 124) define the phenomenon that they study as a “specific subclass of formulations.” Their primary concern is with “formulations of conversational gists and, to a lesser extent, with formulations of upshots” (ibid.: 136), such as when a recipient or speaker of a story, report, announcement, or simply “news,” as Heritage and Watson (ibid.: 124) call it, formulates the sense or gist of the matters presented to him or by him. Heritage and Watson discuss formulations in terms of three “orders” (ibid.: 139) of conversational organization: 1) utterance-by-utterance organization, in relation to which Heritage and Watson discuss an adjacency pair organization7 involved in formulating, 2) organization in terms of topic order, where formulations “may serve to demonstrate understandings of the cumulative import of a previous string . Adjacency pairs are sequences in which a particular type of response is conditionally relevant from the initial action. If no response or no fitting response is given to the initial action, it is noticeably absent or inappropriate. A typical example of an adjacency pair is a question-answer sequence. In the technical jargon, a question “projects” an answer. Projections between the first pair part and the second pair part of an adjacency pair are strong. Consider this example from Hucthby and Woffitt (1998): 1A: Can I have a drink? 2B: Are you over 21? 3A: No 4B: No Notice that even though other necessary business is attended to in lines 2–3, the answer is still conditionally relevant – it is “hanging in the air,” so to speak.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.8 (186)
Anders Hougaard
of utterances” (ibid.: 150); and 3) the overall structural organization of the unit “a single conversation,” where formulations can serve as “preclosings,” that is as an action which brings the whole conversation as such into a possible closing sequence. As Heritage and Watson (ibid.: 139) point out, there is one important caveat to this division into organizational orders: that while such analytic isolations of conversational orders or general issues relating to conversational construction may prove useful from an analytic point of view, the production and monitoring of the orderliness of conversation is characteristically achieved through the simultaneous and integrated treatment of utterances in terms of all orders and with respect to the solution of all problems in maintaining conversational coherence.
The current study mainly focuses on issues pertaining to conversational organization in terms of the topic order (#2 above), but it also discusses important points in relation to the utterance-by-utterance order of organization (#1 above) and the overall structural organization of the unit “a single conversation” (#3 above). Furthermore, with its focus on what is considered a type of formulations, turnpacking, the present study additionally focuses on a quality of formulations which is only mentioned en passant by Heritage and Watson, namely the way in which they interpret preceding utterances instead of just “neutrally” summing them up. Moving on, there are a number of issues which need to be addressed in order to establish whether Laporte’s turn in lines 18 and 20 constitutes a particular type, of recognizable action which the interlocuters orient to as such a particular action – among these are the placement of the action and whether the interlocutors orient to it as in fact “packing up” turns at talk. We start by considering whether its sequential position in the flow of talk is particular. At two places during GE’s report Laporte contributes a “yeah” (ll. 8 and 16). However, these are not treated as turns; that is, they are not responded to by GE as requiring interactional attention other than a mere continuation of his report. Furthermore, Laporte does not make any indications that the yeah’s should be given any particular treatment. What the two yeah’s do are to mark recipiency and to encourage continuation of the report by GE. Elsewhere (Schegloff 1982) such receipt talk elements have been described as “continuers”. In principle, however, there are many places at which Laporte could have responded with a full turn to GE’s report. According to a classic study by Sacks, Schegloff and Jefferson (1974), there are three general rules for turn-transition in interaction. Initially, they argue, a speaker, is entitled to one turn-constructional unit (hereafter TCU, typically a clause or phrase which can be heard as somehow an interactional “unit”–potentially as a turn by itself) and the first possible completion of that unit constitutes a ‘turn transition-relevance place.’ At such a place, they propose, the following set of turn-transition rules hold:
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.9 (187)
Chapter 6. Compression in interaction
(1) For any turn, at the initial transition-relevance place of an initial turn constructional unit: a. If the turn-so-far is so constructed as to involve the use of a “current speaker selects next” technique, then the party so selected has the right and is obliged to take next turn to speak; no others have such rights or obligations, and transfer occurs at that place. b. If the turn-so-far is so constructed as not to involve the use of a “current speaker selects next” technique, then self-selection for next speakership may, but need not, be instituted; first starter acquires rights to a turn, and transfer occurs at that place. c. If the turn-so-far is so constructed as not to involve the use of a “current speaker selects next” technique, then current speaker may, but need not continue, unless another self-selects. (2) If, at the initial transition-relevance place of an initial turn constructional unit, neither 1a or 1b has operated, then the rule set a-c reapplies at the next transition-relevance place, and recursively at each next transition-relevance place, until transfer is effected.
It should be emphasized that we are not here talking about causal rules, which explain human behavior as being conducted just so because of the rule. We are talking about a particular social orientation. The rules then explicate observed patterns of behavior and expectancies, not assumed generative schemas. According to these rules, Laporte could not take a next turn after, for instance, l. 1, since GE in and through the initial TCU – proclaiming “I want to protest” – selects at least one other turn-constructional unit for himself. Furthermore, the talk following from l. 2 does not seem to reach a conceivable TCU completion until l. 11. However, Laporte might have taken a next turn after l. 11, l. 13 or, as indeed he does, after l. 18, since the end of all three can be heard as the completion of a turn-constructional unit. Thus whether or not the assumed turn-packing utterance is simply produced as any next utterance produced at a turn transitionalrelevant place, or whether this action comes at a particular transitional-relevance place we cannot say yet. Notice though that Laporte’s response in l. 18 makes an assessment of the topic of the caller’s turn – the demonstrators – and this follows the caller’s attribution of standpoint to the demonstrators. Laporte’s response then follows a TCU which is somehow descriptive of its topic. Other data suggests that a turn-packing utterance (and perhaps a range of other relevant actions) may be invited by a shift in the reporting or telling done by current speaker. Consider the following excerpt which has the caller (D) reporting her ambivalence about how to feel about the US presence in Iraq (ll. 1–12) and Laporte taking the next turn in line 13:
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.10 (188)
Anders Hougaard
(Excerpt 2, C#1) 1D: you know I went down on monday to the uh the demonstration 2D: because i wAsn’t sure how I felt. 3D: uh i have uh FRIENDS whose BROthers are in saudi arAbia no:w, 4D: i have FRIENDS that i went in hIgh school with (.) that are 5D: thEre, (.) 6D: a:nd (.) uh i wanna supPORT them, 7D: < but on the Other hand> I don’t wanna see them 8D: come back in a BOdy bAg. (.) 9D: and uh it was REAlly NICE to go down and be with people who 10D: FELT the same way that I did. 11D: and uh (.) what i’m seeing today REAlly upsEts me; 12D: [ 13LL: [it’s hArd to know HOW to feel now isn’t it. 14D: it YEAH exACtly. 15LL: dEbbie we’re gonna go back to bob LAzec now. (6Gulfwar.ca, ll. 37–51)
← ← ←
Notice how D makes a shift in his turn from reporting on a topic to assessing it in line 9. First a pre-positioned assessment (Goodwin & Goodwin 1987: 14), “it was really nice,” is produced before the assessable, “to go down and be with people who felt the same way that I did” (l. 10), has been made available. A second assessment, “really upsets me,” follows (in l. 11) after its assessable, “what I’m seeing today” (l.11), has been made available. Notice that according to general turn transition – relevance rules, Laporte could have taken a turn after line 10 which can be heard as the completion of a TCU (notice the falling intonation) – but he doesn’t. Instead, D goes on; not to produce further reporting, but to produce a second assessment, in principle a repetition of his prior TCU (in line 9), though with a different content. This may indicate that the prior turn-constructional unit was an appropriate place for turn-transition since the turn does not progress thematically (it continues to do assessment). Notice also that in both assessment utterances the central components (“really nice” (intensifier + adj) and “really upsets (me)” (intensifier + verb)) are produced with more intensity (higher sound) than their surroundings, which makes these assessment components stand out from the assessables (“to go down and be with the people who felt the same way that I did”, ll. 9–10, and “what I’m seeing today,” l. 11) as the components which define the current activity. Thus there is a clearly signaled shift from report detailing to assessment activity. The turn has moved from reporting (ll. 1–8) to an intensively produced assessment (ll. 9) followed by a second intensively produced assessment (ll. 10–11).
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.11 (189)
Chapter 6. Compression in interaction
It is in immediate response to the second unit of assessment utterances (l. 11) and directly following an intensively produced assessment component that recipient, Laporte, takes a next turn and produces a turn packing utterance, which in response adds an impression of the situation reported. I do however still hesitate to postulate a general, recognizable pattern where turn-packing utterances are invited at particular places in the flow of talk (e.g. being invited by assessments, descriptions or impressions which may all have an evaluative quality to the coparticipants, which may then be taken as a place to relevantly engage with a turn-packing utterance). It remains a hypothesis to be studied further. Yet other things indicate that turn-packing utterances are indeed within a scope of relevant and appropriate actions following one or more turns of reporting/news/accounting, etc.; they are just not projected by current speaker’s actions (see Note 7). First, it is clear that there is a recognizable action which does something specific in the interaction. As we have already seen, a turn-packing utterance not only may close a topic, it may also preclose the whole conversation. In both excerpt 1 and 2 the turn-packing utterance is followed by recipients’ agreement and then a closure of the conversation. (Excerpt 1, C#11) 18LL: a lot of BRIGHT intElligent people of CONscience who a:re; 19GE: yeah; 20LL: OUt there uh mArching because they feel that they HAVE to. 21GE: i am GLAD you brought that up. 22LL: .hh i THANK y’ for the cAll. (9Gulfwar.ca, ll. 65–69)
← ←
(Excerpt 2, C#1) 13LL: [it’s hArd to know HOW to feel now isn’t it. 14D: it YEAH exACtly. 15LL: dEbbie we’re gonna go back to bob LAzec now. (6Gulfwar.ca, ll. 49–51)
← ←
In the following example the same pattern is potentially at play, but potential closure gets cancelled by recipient’s treatment in lines 11–12 of the turnpacking/agreement sequence (ll. 8–10) as closing a topic which is not the reason for the call, which is instead announced to follow: (Excerpt 3, C#9) 1GE: 2GE: 3GE: 4GE:
hey thAt gUY that said that he’s sick of Terrorists. (.) should rEAlly pay more attention to what’s goin’ ON in the world. .hh BUSH JUST okayed thIrty million DOllars for el SALvador –
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.12 (190)
Anders Hougaard
5GE: .hh and it been a HUNDred thousand people mUrdered 6GE: < there.> .hh (-) 7GE: a:nd uh 8LL: we have a HIStory of supporting terror in ONE form or 9LL: another. 10GE: we DO ALL over the whole wOrld; 11GE: i am callin’ to proTEST the: (.) Entry into uh the WAR in 12GE: irAQ; .h (9Gulfwar.ca, ll. 15–26)
← ←
In light of other excerpts which show the same pattern as in excerpt 1 and 2, excerpt 3 may be seen as providing negative evidence for the recognizable pattern turnpacking/agreement/closure of conversation, since GE at a place where a topic has been closed in and though a turn-packing/agreement sequence announces that the reason for the call is only then to be introduced. It thus seems fairly well established that turn-packing utterances, when recognized as such, establish a possible preclosure of a conversation, which may be cancelled as such by making a new topic relevant. Thus the general interactional point to be noticed is that when turns are packed up, a conversation is potentially brought to an end. But how are the turn – packing utterances recognized as turn-packing utterances? Above I argued that turn-packing utterances are probably not projected by current speaker’s turn. But the action may still be recognizable in terms of its composition and placement. First, it is at least noticeable that typically turn-packing utterances are composed as THERE-constructions or IT-constructions: (Excerpt 1, C#11) [There are] a lot of intelligent people of conscience who are out there marching because they feel that they have to (Excerpt 2, C#1) It’s hard to know how to feel now, isn’t it (Excerpt 5, C#10) It’s pretty amazing what we were able to do last night There are, however, exceptions to this tendency in excerpt 3: (Excerpt 3, C#9) We have a history of supporting terror in one form or another Yet, despite the overwhelming occurrence of particular constructions which seem to be recognized as turn-packing utterances, the linguistic material itself cannot account for the action’s recognizability. Thus it is not the linguistic construction that determines the activity; rather the linguistic construction is used as an appropriate linguistic resource for doing an interactionally relevant job. The same
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.13 (191)
Chapter 6. Compression in interaction
linguistic constructions do lots of different interactional jobs. However, while perhaps not specifically projected at particular places and while not being recognizable as a specific linguistic structure, turn-packing utterances may still have a general relevance: At a turn transition-relevant place, that is at a conceivable completion of the delivery (following perhaps a TCU which describes, assesses or expresses an impression of the current topic) – spanning one or more turns by one speaker or more – of information regarding a specific topic (doing reporting, giving news, doing accounting etc.), a relevant responsive action is, unless other actions have been projected/made relevant, to pack up the turn/series of turns in and through a turn-packing utterance which in turn may serve as initiating a closure of the conversation. And the typical linguistic resources used to achieve turn packing may be THERE-constructions and IT-constructions. Yet, other constructions can be used too, because of the general relevance of turn-packing at such places. However, it may indeed be the case that the typical occurrence of THERE- and IT-constructions has contributed to making turn packing utterances recognizable through the recurrence of particular linguistic forms at such places. One major issue has not yet been discussed: the claim that turn-packing utterances are understood by the interlocutors as packing up turns and not just as responding to turns. How can it be seen that turn-packing utterances in fact pack up? In this paper I present two interactionally based arguments which support the intuition that turn-packing utterances pack up turns. First, typically in interaction a next turn orients to an immediately preceding turn, and if the previous turn consists of several turn constructional units, the next turn typically orients to the final – or a late – turn constructional unit. That however, is not the case with excerpts with what I call turn-packing utterances. Contrast the following two excerpts: (Excerpt 6, Hyp#7) 1LL: 2LL: 3LL: 4LL. 5LL: 6LL: 7LL: 8LL: 9LL: 10LL: 11LL: 12LL: 13LL:
all right it’s about ah: what 9 am in bagdad iraque and the second wave of u.s. bombers is continuing to bomb (.) iraque and i guess we’re just waiting to hear from saddam hussein the only report we got (.) uh via service report that uh said the iraqui radio uh said that the united states and the allies would regret our decision uhm but they didn’t say how or why (.) and we have yet to see any response from iraque at all (.) [(any) retaliation]
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.14 (192)
Anders Hougaard
14XX: [it would be ] a long time before we see what their 15XX: idea of retaliation is (1Gulfwar.ca, ll. 11–42)
← ←
(Excerpt 1, C#11) 1GE: .hh and we’re VEry much against people that uh; 2GE: STOP TRAffic an an START FIRES. 3LL: [yeah 4GE: [< and do other things.> 5LL: a lot of BRIGHT intElligent people of CONscience who a:re; 6GE: yeah; 7LL: OUt there uh mArching because they feel that they HAVE to. 8GE: i am GLAD you brought that up. 9LL: .hh i THANK y’ for the cAll. (9Gulfwar.ca, ll. 61–69)
← ←
Notice in excerpt 6, how XX’s turn (l. 14–15) responds directly to the last bit of LL’s long turn (ll. 1–13) repeating the word “retaliation” (l. 15). In contrast, in excerpt 1 it takes more than the last sentence of GE’s (ll. 1–2 and 4) turn to understand the meaning of LL’s response (l. 5 and 7). The meaning of LL’s response does however become clear if we include most of GE’s preceding turn, from line 2 to line 17: (Excerpt 1, C#11) 1GE: 2GE: 3GE: 4GE: 5GE: 6GE: 7GE: 8LL: 9GE: 10GE: 11GE: 12GE: 13GE: 14GE: 15GE: 16LL: 17GE:
now i wanna proTEST; (-) people who MARCH a::nd uh crEate vIolence;= ← From < and i think you’re exACtly rIght;> .hh uh y your rePORTS <<monotonous> on your stAtion say that there are about ten thousand marching and a fEw hUndred [(-) broke off and did those things so; [yeah; .hh i ↑HOPE that we don’t uh lUmp everybody together and say that; (.) PROtesters A:re violent people;= they (.) they DON’T WANT the KILLing to contInue just like myself; .hh and we’re VEry much against people that uh; STOP TRAffic an an START FIRES. [yeah [< and do other things.> ← To
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.15 (193)
Chapter 6. Compression in interaction
a lot of BRIGHT intElligent people of CONscience who ←Turn-packing a:re; 19GE: yeah; 20LL: OUt there uh mArching because they feel that they HAVE ←Turn-packing to. 21GE: i am GLAD you brought that up. 22LL: .hh i THANK y’ for the cAll. (9Gulfwar.ca, ll. 48–69) 18LL:
Hence in and through their responses producers of turn-packing utterances show orientation to more of the preceding talk than the last TCU(s). But how much of the preceding talk is oriented to? Another type of pattern shows how interlocutors may orient to particular lumps of talk. Consider again excerpt 3 and 1: (Excerpt 3, C#9) hey thAt gUY that said that he’s sick of Terrorists. (.) ← Source should rEAlly pay more attention to what’s goin’ ON in the 3GE: world. .hh 4GE: BUSH JUST okayed thIrty million DOllars for el SALvador – 5GE: .hh and it been a HUNDred thousand people mUrdered 6GE: < there.> .hh (-) 7GE: a:nd uh 8LL: we have a HIStory of supporting terror in ONE form or ←Anchor, Turnpacking 9LL: another. ← Turnpacking 10GE: we DO ALL over the whole wOrld; (9Gulfwar.ca, ll. 15–24) 1GE: 2GE:
(Excerpt 1, C#11) 1GE: 2GE: 3GE: 4GE:
now i wanna proTEST; (-) people who MARCH a::nd uh crEate vIolence;= ← Source < and i think you’re exACtly rIght;> .hh uh y your rePORTS <<monotonous> on your stAtion say that 5GE: there are about ten thousand 6GE: marching and a fEw hUndred [(-) broke off and did those 7GE: things so; 8LL: [yeah; 9GE: .hh i ↑HOPE that we don’t uh lUmp everybody together and say 10GE: that; (.)
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.16 (194)
Anders Hougaard
11GE: PROtesters A:re violent people;= 12GE: they (.) they DON’T WANT the KILLing to contInue just like 13GE: myself; 14GE: .hh and we’re VEry much against people that uh; 15GE: STOP TRAffic an an START FIRES. 16LL: [yeah 17GE: [< and do other things.> 18LL: a lot of BRIGHT intElligent people of CONscience who ←Turnpacking a:re; 19GE: yeah; 20LL: OUt there uh mArching because they feel that they HAVE ←Anchor, to. Turnpacking 21GE: i am GLAD you brought that up. 22LL: .hh i THANK y’ for the cAll. (9Gulfwar.ca, ll. 48–69)
In excerpt 3, recipient uses the word “terror” (l. 8) in his turn-packing utterance, which connects the turn-packing utterance back to the beginning of previous speaker’s story where “terrorists” (l. 1) specifies the topic for the following talk. Next speaker thus anchors his turn-packing utterance at a specific place in previous speaker’s turn indicating an encapsulation of the portion of talk between the source, “terrorists” (l.1), and the anchor term, “terror” (l.8). In excerpt 1, first speaker’s story is packed up and assessed in lines 18 and 20. In line 20, the anchor term “marching” connects the turn-packing utterance back to the source “march” (l. 2), which specifies the topic for the following talk. Thus in and through recycling a term, which may be construed as setting the topic for a portion of talk, recipients may anchor their turn-packing utterance at a specific place in the previous talk and in that way specify which part of that talk is packed up. In sum, producers of turn-packing utterances do not only relate their turnpacking utterances to the last bit of a previous turn; and this together with the fact that producers of turn-packing utterances may anchor their response at particular places in the previous talk which may be understood as specifying the topic for the following talk shows that the interlocutors indeed orient to the turn-packing utterance as summing up, encapsulating or packing up a particular portion of talk beyond the normal last bit of a turn. Another point that I wish to make here and which is demonstrably significant to the interlocutors is that a turn-packing utterance is not just a “neutral” summary of previous talk. The utterance does the packing-up in and through particular linguistic elements which get to constitute a particular interpretation of the thereby packed up material. Yet, turn-packing utterances may differ with respect to how much content they are understood to add to the packed up material. Three
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.17 (195)
Chapter 6. Compression in interaction
examples from my collection will illustrate this point in forming a continuum from more neutral turn-packing to turn-packing which simultaneously is clearly understood to contribute an interpretation to the packed up TCUs: (Excerpt 2, C#1) 1D: 2D:
you know I went down on monday to the uh the demonstration because i wAsn’t sure how I felt.
[Lines 3–12 which contain a report of the demonstration have been cut out] 13LL: [it’s hArd to know HOW to feel now isn’t it. 14D: it YEAH exACtly. (6Gulfwar.ca, ll. 37–50)
← ←
(Excerpt 5, C#10) [DO: talking about the effectiveness of the allied attack on Iraq] 1LL: it’s prEtty aMAZing <<monotonous + l> what we were able to 2LL: do last night.> 3DO: i thInk uh it’s < gOnna tAke a lIttle lOnger to really> 4DO: you know sOrt out WHAT Happened. 5DO: but i REAlly fEEl like that uh; (.) 6DO: they’ve made a MAjor DEBT by what they dId last night;= 7DO: an and in ↑MAKing this very short an an nOt drAgging it on. 8LL: well if youre gonna FIGHT this is the way to DO it. 9DO: i i aGREE. (10Gulfwar.ca, ll. 138–146)
← ←
← ←
(Excerpt 1, C#11) 1LL: a lot of BRIGHT intElligent people of CONscience who a:re; 2GE: yeah; 3LL: OUt there uh mArching because they feel that they HAVE to. 4GE: i am GLAD you brought that up. 5LL: .hh i THANK y’ for the cAll. (9Gulfwar.ca, ll. 65–69)
← ←
In excerpt 2, recipient (Laporte) packs up the contents of first speaker’s turn with the utterance “it’s hard to know how to feel, isn’t it?” (l.13). The turn-packing utterance recycles source terms with which first speaker begins her account of her mixed feelings: “How I felt” (l. 2) → “how to feel” (l. 13). No new material is added. Thereby, the turn-packing utterance becomes merely an abbreviated reproduction of the previous turn. In excerpt 5, a strongly evaluative story packing utterance is at first produced (ll. 1–2) which is then disqualified and followed by another story packing utterance (l. 8). The excerpt then shows how the turnpacking utterance imposes a particular understanding which may be subject to
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.18 (196)
Anders Hougaard
interactional negotiation. Finally, in excerpt 1, recipient orients – in his wording – to the turn packing utterance (ll. 3–4) as contributing a new point to the story while packing it up (l. 4). These observations find support in Heritage and Watson’s (ibid.: 137) study of formulations: . . .multiple readings may, on occasion, prove extractable from stretches of conversation and, by the same token, members may thus be oriented to the occasioned multifaceted quality of their conversational productions with a view to establishing preferences among available readings. Thus, given that it is not the case that members can invariably treat stretches of conversation as automatically unproblematic from a descriptive point of view, it is clear that a concerted arrival at a “determinate, for-all-practical-purposes” reading of a stretch of conversation may, on occasion, be treated as problematic and hence in need of repair.
Excerpt 5 displays a case of the latter. The final point I wish to consider in this paper is the possibility that a turn-packing utterance may constitute a first pair part of an adjacency pair (see Note 7), that it projects a response in terms of an acceptance or rejection of the turn-packing utterance, what Heritage and Watson (ibid.: 141) call a “decision” from next speaker. In their data, Heritage and Watson find that formulations strongly constrain “the items which may fill subsequent slots” (ibid.: 140) and that the absence of a decision following a formulation is noticeable. Heritage and Watson thus consider formulations and their decisions a type of adjacency pair: the formulation projects a decision as next relevant action. In this paper we have only seen cases in which a turn-packing utterance is followed by a decision. Therefore it is certainly a possibility that turn-packing utterances project decisions as well. Let us now, before we evaluate the interactional findings in cognitive terms, list the central points made about turn-packing utterances in the first part of the paper: 1. Speaker A talks about some issue – doing reporting, accounting, giving news, etc. 2. The topic may extend more turns. 3. Speaker A may make a shift to evaluate the topic of the turn(s) by assessing or describing it. 4. Speaker A’s turn may reach a culmination in evaluating its content, which may mark a departure from “reporting” to a “concluding mode.” 5. At this possible culmination, speaker B may engage with a turn-packing utterance. 6. Alternatively there is no shift to assessments in A’s turn and speaker B just produces a turn-packing utterance at a turn transition-relevant place.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.19 (197)
Chapter 6. Compression in interaction
7. The turn-packing utterance presents a particular interpretation of the packed up turns. 8. Speaker B’s turn-packing utterance may be the first pair part of a topic-closing sequences and of a sequence that precloses the conversation. 9. The turn-packing utterance projects a decision. 10. A decision is given by speaker A, which either agrees with the turn-packing utterances or disagrees, in which case a “loop” may be initiated leading to a new turn-packing utterance and another projection of a decision. 11. The topic is closed. The conversation may now get closed.
. Turn packing utterances as a type of conceptual compression
i) A behavioral cognitive account A central part of my discussion of the cognitive implications of what we have seen in the data above is to clarify my position with respect to the relation between behavior and cognition. A typical blending analysis would at this point proceed to consider what mental structures – hidden below the surface, in the mind of each participant – are being composed in the talk that we have analyzed. It would consider what mental spaces are set up, what conceptual structures they contain, and how connections between mental spaces and within mental spaces may be compressed in the turn-packing utterance. However, such an interpretation of the data would completely miss the lessons we learn from studying interactional behavior. What we have seen is examples of remarkable shared processes of meaning construction where interlocutors in and through close sequentially organized coordination act out interactional patterns and achieve meaningful interactional products. To represent this as hidden networks of mental spaces, as is common practise in blending theory, would be deeply unfaithful to the data. One might propose that the interlocutors individually produce similar or comparable networks of mental spaces and that the matching of these networks in some intricate mediating process accounts for the shared meaning between separate minds. But why impose such an awkward dualism on the data? The only answer seems to be that as theorists of cognitive processes of language and discourse we always tend – in keeping with the tradition of cognitive science – to take the individual mind as the proper or primary locus of cognition. However, the fact that can be observed in the data above is simply that there is no need for speculation about complex, “internal”, cognitive processes in order to account for the shared cognitive work that takes place. No matter how different each individual’s personal experience and biography might be, the shared work that is going on is exactly characterized by being shared and therefore speculations about
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.20 (198)
Anders Hougaard
individuals’ individual experiences are at best irrelevant. In this paper cognitive processing in interaction is consequently not viewed as something which goes on “below the surface” or “behind the scenes”; it is viewed as enacted and composed by the interlocutors jointly. Now, this is not a rejection of anti-behaviorist ideas entertained since the late 1950s and the conviction that psychology cannot be reduced to mere behavioral patterns. It is a rejection of an unfortunate side-effect of the anti-behaviorist movement that started with or was reawakened (cf. Sinha 2007) by such works as Chomsky’s Syntactic Structures (1957), his review (1959) of B. F. Skinner’s Verbal Behaviour (1957) and Neisser’s (1967) founding work Cognitive Psychology. The side-effect is that in rejecting Behaviorism as proper psychology, cognitive science has come to take on a “hiddenmind-view”, which grossly underestimates or neglects what is actually achieved interactionally when people construct meaning face-to-face. Instead, to protect psychology from behaviorist impoverishment, it seems, psychological processes have been located to often obscure, inaccessible inner spheres, which are constantly at risk of leading cognitive science into a sheer speculative abyss of unverifiable hypotheses. The present approach is anti-behaviorist in many ways including the following: 1. It assumes that interlocutors are conscious and interpreting beings, but it focuses on shared, enacted interpretations 2. It assumes the phenomenon of mind, but as a shared activity in which sense emerges between people 3. It acknowledges and is interested in the possibility that behavior may be generated by psychological capacities, but not as a default explanation of the first resort or as the only explanation. In other words, the approach undertaken here does not replace psychology or cognition for social eliminativism, but it brings social action to the center of attention, which causes a reconsideration of the often implied notion that social behavior somehow emerges from individual psychological or social capacities. This point will spin off a socio-cognitive hypothesis of turn-packing utterances later in the paper.
ii) Compression as the achievement of an interactional pattern In the case of turn-packing, compression is not something people do for themselves on the basis of individually interpreted communicative resources. It is something they achieve in and through concerted, sequentially organized interaction. On the basis of the analyses above we can idealize this pattern in the following way:
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.21 (199)
Chapter 6. Compression in interaction
turn 1
turn 2
turn n…
turn-packing
decision
Figure 1.
The full horizontal arrow represents the interactional trajectory composed of turnby-turn sequences. N turns are produced which are followed by a turn-packing utterance. The latter has the quality of bringing previous turns within a specific, limited range to bear on it, such that a response to the turn-packing utterance is a response to those turns in some particularly interpreted sum. This is represented by the dotted arrow going from the turn-packing utterance to turn 1 and back to the turn-packing utterance. The turn-packing utterance projects a decision to that interpreted sum. Finally, the curved arrows represent the way in which each turn sits in and is defined by the ongoing interaction. Each turn is context-shaped in that it is oriented to the prior turn, which is represented by the lower curved arrows, and each turn is context renewing in that it sets up a new context for the following turn, which is represented by the upper arrows. This structure is cognitive and dynamic in the sense that it is “processed” and reconstructed intersubjectively in and through the enactment of it, visibly to both (or all) interlocutors. Recognition of the structure provides local expectations and local possibilities for action, and it offers a shared sense of completion.
iii) Compression in interactional working memory What then about the very information which the members have shared and which is compressed, interpreted and evaluated? How are we to think of that? We can suggest a typical network analysis in which the information shared gets interpreted and compressed into a blended space. More technically, outer space vital relations (e.g. identity) would be compressed between discursively connected spaces (see Figure 2). This mental space notation captures in an idealized way the product which the turn-packing utterance brings forth for evaluation. A particular amount of information (encapsulated by the dotted circle) is made available through a blended space which receives its governing structure from the concept evoked by the turnpacking utterance. We can certainly think of the interactionally achieved treatment of the discursively shared information in this way. But how are we to understand this idealized model then? First of all, as argued above, it must be understood as shared. Secondly, it must be understood as an abstraction which may not be real at
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.22 (200)
Anders Hougaard
Turnpacking schema
Discourse content spaces connected by discourse connectors, identity connectors, etc.
Blend
Figure 2.
any actual interactional point. Thirdly, the model has disintegrated the very discursively shared information from the interactionally achieved actions in and through which it is established. Fourthly, the model is concept or information based; it is not based on the possible psychological processes that take place. Finally, the model presupposes a mental partitioning of the flow of information in interaction, which may however, seem quite qualified. Consider for example how the members orient to the turn-packing utterance as turning around upon the foregoing shared information. Furthermore, the turn-taking system – which was described above – shows that members indeed orient to chunks within the interaction which may be interpreted “cognitively” in terms of some sort of “mental spaces.” Yet, the interactional analyses conducted above demand more than the idealized, product information-structure represented in the typical mental space analysis. Mental spaces, Fauconnier (1998: 1–2) writes, “proliferate in the unfolding of discourse, map onto each other in intricate ways, and provide abstract mental structure for shifting anchoring, viewpoint, and focus, allowing us to direct attention at any time onto very partial and simple structures, while maintaining an elaborate web of connections in working memory and in long term memory.” Mental spaces are here connected to a psychological research tradition – the field of memory reserach – and this seems like a promising way to go when considering how to understand what, in terms of information that is shared and worked on, it is that the interlocutors are achieving. Research into memory has a long history. Here I shall be concerned with a line of research established with Baddeley’s (1986) introduction of the concept of “working memory” as a specification of the term “short term memory”. The notion of working memory sparked off an interest in very local memory tasks such as remembering telephone numbers, lists of words,
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.23 (201)
Chapter 6. Compression in interaction
etc. A common finding has been that the maximum number of items which people can keep immediately available is around seven. This may seem to fit Fauconnier and Turner’s (2002b) “anecdotal” (as they call it) hypothesis that humans can keep 5–7 mental spaces active at the same time. However, what an “item” is may differ greatly. Hence, if the maximum number of items that can be kept active in working memory is seven in cases where items are constituted by unrelated numbers, most mental space “items” (shapes, causal relations, words, events, etc.) will typically seem to belong to a more “heavy” type than numbers. Thus, if we look at the type and amount of information commonly ascribed to a mental space (a walking monk, a businessman with attributes, a boat that sails from Boston to San Francisco, former President Bill Clinton, etc.) it would seem that 5–7 mental spaces go far beyond the limit of seven items as these are typically understood – notwithstanding gestalt “chunkings” which aid working memory (cf. Driscoll 1993). If we count the features of concepts that are held to be active, hardly more than one or two mental spaces could be held active in working memory. Yet, as Kintsch et al. (1999) point out, real life memory tasks are quite different from the superficial situation of trying to remember a list of words in a lab. Tasks such as text comprehension, demand a different type memory, they argue, dubbed “Long Term Working Memory” (LTWM), which can roughly be described as a local combination of working memory and long term memory. The principle in forming an LTWM is that each new feature of for instance a text establishes a new node in a network of knowledge. This node is held active in working memory, but it is also connected to previous nodes activated by the text which have passed out of working memory and – through stable and strong connections – to nodes in long term memory. The process is illustrated as follows:
Figure 3.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.24 (202)
Anders Hougaard
I will not go further into an account of Kintsch et al.’s model here. I will only observe that this notion of memory seems more fitting for mental spaces than working memory since many of the meaning constructs which are described in blending litteratue are such episodic processes which seem to go beyond working memory capacity (see below). Kintsch et al. also make several points which make their model interesting from an interactional point of view. Firstly, the integration of active processes and “stored” but available knowledge in real life memory tasks seems fit as a psychological interpretation of Alfred Schutz’ (1974) socio-phenomenological understanding of social “acts” versus social “actions.” The action, Schutz argues, is the conscious, developing, motivated behavior and the act is the remembered, completed action. This can be interpreted to mean that interlocutors’ acts are temporally contingent achievements which span an in-themaking-phase which is not but becomes an interpreted social act. In psychological terms, we could propose that the action concerns things that are simultaneously active whereas the act has passed out of working memory. Secondly, Kintsch et al. argue that retrieval of long term memory is fast and automatic, and not intentional. This is much in line with what we observe in naturally occurring interaction. Most of the time members act swiftly, automatically and without careful preceding thought, while of course still sensefully. Yet, in many classic blending analyses the process is depicted as a careful and intentional process: thinking about riddles, solving math problems, studying art, reading literature, etc. Unfortunately, the LTWM model is what we can characterize as an “inside-out” model. First of all it assumes that knowledge that is used in a situation is kept at store, ready for activation when needed. This, however, goes against not only recent (e.g. Rummelhart et al., 1986) and less recent (e.g. Bartlett 1932) notions in cognitive science and psychology of memory as being reconstructive, it also does not provide a good account of the constant reworking of socially established knowledge which can be observed in interaction. For instance, outside the idealized, abstracted structure depicted in Figure 1, there are no two turn-packing sequences which are alike. If the pattern was merely activated from its place in memory, we would expect much more homogeneity across its enactments. But instead it seems that knowledge of this interactional structure is being re-composed on every occasion of its occurrence. Yet, the most essential thing which makes the LTWM model inadequate for interaction is that it only focuses on how psychological processes afford meaningful activities, not how meaningful activities may define psychology. What I would like to propose as my suggestion for how we may understand the cognitive significance of the turn-packing sequence is that it creates for the members an interactional memory chunk, extending an interactionally appropriate time span and compressing (“chunking”) an interactionally relevant amount of information. We know that turn-packing utterances serve as closures of topics and as preclosures of conversations. And there is also some indication that turn-packing
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.25 (203)
Chapter 6. Compression in interaction
utterances project a decision. Thus by establishing an interactional memory chunk at such places in the interaction a concerted piece of memory has been created which of course may be of great importance to the involved members: “What did we say?” or “What was the conclusion we ended with?”. And it seems quite logical from this perspective that a turn-packing utterance requires a decision: consent from or rejection by the other interlocutor(s). Hence what blending theory’s “compression” becomes here is the interactional construction of a gestalt-like piece of shared memory. Consequently if we can say that the turn-packing utterance compresses in the sense that it encapsulates a chunk of turns (or a chunk of utterances in one turn) which are thereby connected to it and come to bear on it, the “goal” of the compression here is not (only) to achieve “human scale” for purposes of mentally or socially grasping information but for remembering it as somethings. As a final analytic point we must consider that the turn-packing also involved a particular interpretation which was imposed on the compressed content. This can be seen to support Fauconnier’s and Turner’s hypothesis that blending and compression go hand-in-hand, if indeed the interactional processes described here are accepted as involving compression and blending as described in authoritative sources such as Fauconnier and Turner (2002a). But again the interactional analysis questions the cognitive theory. Do the members do what they do because they can blend and compress, or do blending and compression happen because the members do what they do? This question is too comprehensive to attempt to answer here, and instead I once more direct the reader to Hougaard and Hougaard (forthcoming), where such issues are also discussed. Summing up, then, the “results” which are suggested to follow from the interactional analyses in the first part of this paper are: 1. During talk, interlocutors may produce turn-packing utterances, which: – sum up and interpret turns at talk, – serve as closures of topics or preclosures of conversations – project decisions 2. The interactional work done by turn-packing utterances involves a particular type of “blending” and “compression”, but these are understood in certain specific ways that are not a part of Fauconnier and Turner’s theoretical machinery 3. The cognitive processes are enacted and visible, not prompted and hidden 4. And in general cognition in interaction is “onstage”, not “backstage” 5. Compression and blending involved in turn-packing utterances are not ends in themselves; they are aspects of the construction of a shared interactional memory chunk, which may come about because of the interactional activity
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.26 (204)
Anders Hougaard
6. The construction of this chunk goes beyond working memory. Furthermore, it is not dictated by “inner”/pre-existing psychological capacities; it creates a piece of psychology. Notice that these “results” are not solutions to anything, they are indications of a different way of considering issues that have emerged from blending theory and cognitive semantics as well as from cognitive science in general.
. Some concluding remarks A couple of times I have referred to a forthcoming article by myself and Gitte R. Hougaard which introduces the first ideas and work on the way to a research paradigm for interactionally enacted and constructed sense by entire knowledgeable bodies. This paper along with A. Hougaard (2005) can be seen to exercise aspects of this new research focus specifically in connection to blending theory and cognitive semantics. Yet, in general the paradigm proposed in Hougaard & Hougaard (forthcoming) is not committed specifically to any existing theory. 8 The title of this book is Mental Spaces in Discourse and Interaction. The proposal made in this paper is to turn the typical perspective around and instead discuss mental spaces and blending from the perspective of empirical studies of naturally occurring talk-in-interaction. And this is to be seen as a proposal made to cognitive semantics in general. Much advance in cognitive semantic concerns the ways in which theories can be applied to new fields of inquiry, and not at all to the same extent on how now fields of inquiry can contribute to the (critical) evaluation of cognitive semantics. A case in point is Turner’s Cognitive Dimensions of Social Science (2001) where Turner proposes a recasting of Social Science analyses in terms of blending theory. However, if cognitive semantics is to merge fruitfully with the Social Sciences it must also be considered how the latter may force us to rethink concepts and ideas developed within cognitive semantics.
. Even though the approach undertaken by Gitte R. Hougaard and myself has not been influenced by other attempts at combining conversation analysis and ethnomethodology with cognitive studies, it should be noted that a whole wave within conversation and discourse studies seems under way, which is joined by leading conversation and discourse analysts (cf. Molder te, H. & Potter, J. (2005). Conversation and Cognition. Cambridge: Cambridge University Press & Dijk, Teun A. van. 2006. Discourse Studies, February 1, Volume 8, No. 1)
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.27 (205)
Chapter 6. Compression in interaction
References Morana Alaˇc. (2005). “From Trash to Treasure: Learning about the Brain Images through Multimodality”, Semiotica, 156–1/4, 177–202. Adolphs, Ralph. (2003). “Investigating the cognitive neuroscience of social behaviour”, Neuropsychologia 41, 119–126. Baddeley, A. D. (1986). Working Memory. New York: Oxford University Press. Bartlett, F. (1932). Remembering. Cambridge, England: Cambridge University Press. Bates, Elizabeth and Frederic Dick. (2000). “Beyond Phrenology: Brain and Language in the Next Millennium.” Brain and Language, 71, 18–21. Chomsky, Noam. (1957). Syntactic Structures. The Hague: Mouton. Chomsky, Noam. (1959). “A Review of B. F. Skinner’s Verbal Behavior” in Language, 35, No. 1, 26–58. Coulson, Seana. (2007). “Electrifying results: ERP data and cognitive linguistics”. In M. Gonzalez-Marquez, I. Mittelberg, S. Coulson, & M. Spivey (Eds.), Methods in Cognitive Linguistics. Amsterdam: John Benjamins. Coulson, S. & C. Van Petten. (2002). “Conceptual integration and metaphor: An event-related potential study.” Memory & Cognition 30, 958–968. Damasio, Antonio. (1994). Descartes’ Error. New York: G.P. Putnam. Dijk, Teun A. van. (2006) Discourse Studies, February 1, Volume 8, No. 1. Driscoll, M. P. (1993). Psychology of Learning for Instruction. Needham Heights, MA: Allyn & Bacon. Fauconnier, Gilles. (1994). Mental Spaces. Aspects of Meaning Construction in Natural Language. New York: Cambridge University Press. Fauconnier, Gilles. (1998). “Mental Spaces, Language Modalities, and Conceptual Integration.” In Michael Tomasello (ed.), The New Psychology of Language: Cognitive and Functional Approaches to Language Structure. Mahwah, NJ: Lawrence Erlbaum. Fauconnier, Gilles. (2005). “Compression and Emergent Structure.” In S. Huang, (ed.) Language and Linguistics. 4/6, 523–538. Fauconnier, Gilles and Mark Turner. (2000). “Compression and global insight.” Cognitive Linguistics 3–4/11, 283–304. Fauconnier, Gilles and Mark Turner. (2002a). The Way We Think. Conceptual Integration and the Mind’s Hidden Complexities. New York: Basic Books. Fauconnier, Gilles and Mark Turner. (2002b). “The Past, Present, and Future of Blending Research.” Key note address at The Way We Think: A Research Symposium on Conceptual Integration and the Nature and Origin of Cognitively Modern Human Beings, University of Southern Denmark, Odense, 19–23 August, 2002. Gallese Vittorio & Lakoff, G. (2005). “The Brain’s Concepts: The Role of the Sensory-Motor System in Reason and Language.” Cognitive Neuropsychology 22, 455–479. Garfinkel, Harold. (1967). Studies in ethnomethodology. Englewood Cliffs, N.J.: Prentice-Hall. Garfinkel, Harold. (2002). Ethnomethodology’s Program. Working Out Durkheim’s Aphorism, edited by Anne Warfield Rawls. New York: Rowman & Littlefield Publishers, Inc. Garfinkel, Harold and Harvey Sacks. (1970). “On Formal Structures of Practical Actions.” In: McKinney and Tiryakian (eds.), Theoretical Sociology. Perspectives and Developments. New York: Appleton-Century-Crofts, 337–366.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.28 (206)
Anders Hougaard
Goodwin, Charles and Marjorie Harness Goodwin. (1987). “Concurrent Operations on Talk: Notes on the Interactive Organization of Assessments.” In: IPRA Papers in Pragmatics 1, No. 1, 1–54. Heritage, John and Watson D. R. (1979). “Formulations as Conversational Objects.” In: Psathas, 123–162. Hougaard, Anders (2004). How’re we doing: An Interactional approach to cognitive processes of online meaning construction. PhD Dissertation., University of Southern Denmark. Hougaard, Anders. (2005). “Conceptual Disintegration and Blending in Interactional Sequences: A Discussion of New Phenomena, Processes vs. Products, and Methodology.” In Coulson; Seana & Todd Oakley (eds.). Journal of Pragmatics 10/37, 1653–1685. Hougaard, Gitte R. & Anders Hougaard (forthcoming). “Fused bodies: sense making as a phenomenon of knowledgeable bodies”. In Hanna Pishwa (ed.), Social Cognition and Language. Berlin: Mouton de Gruyter. Hougaard, Gitte R. (forthcoming). “Legitimate Peripheral participations as a Framework for CA-informed Work within SLA. The Problems of Bringing together a Top-down theory and a Bottom-up Approach.” Manuscript. Hutchby, Ian and Robin Wooffitt. (1998). Conversation Analysis. Principles, Practises and Applications. Cambridge, UK: Polity Press. Kintsch, Walter, Vimla L. Patel & K. Anders Ericsson. (1999). “The role of long-term working memory in text comprehension”, Psychologia, 42, 186–198. Lakoff, George. (2007). “The Neural Theory of Metaphor.” Paper distributed on the CoglingList, 14th January. Lakoff, George and Mark Johnson. (1999). Philosophy in the Flesh. The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books Molder te, H. & Potter, J. (2005). Conversation and Cognition. Cambridge: Cambridge University Press. Narayanan, S. (1997). “Karma: Knowledge-based Action Representations for Metaphor and Aspect.” PhD Dissertation University of California, Berkeley. Neisser, Uric. (1967). Cognitive Psychology. New York: Appleton-Century-Croft. Rohrer, Tim. (2001a). “Pragmatism, Ideology and Embodiment: William James and the Philosophical Foundations of Cognitive Linguistics.” In Sandriklogou and Dirven (eds.), Language and Ideology: Cognitive Theoretical Approaches. Amsterdam: John Benjamins, 49–82. Rohrer, Tim. (2001b). “Understanding through the Body: fMRI and ERP investigations into the neurophysiology of cognitive semantics”, paper presented at ICLC 2001, UCSB, Santa Barbara, July 2001. Rumelhart, D.E. and McClelland, J.L. (eds.). (1986). Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Vol. 1–2: Foundations. Cambridge, MA: MIT Press. Sacks, Harvey. (1992). Lectures on conversation, vol. 1–2, edited by Gail Jefferson. Oxford: Basil Blackwell Sacks, Harvey, Emmanuel Schegloff and Gail Jeferson. (1974). “A simplest systematics for the organization of turn-taking for conversation.” Language, 50, 696–735. Schegloff, Emanuel A. (1982). “Discourse as an interactional achievement: Some uses of ‘uh huh’ and other things that come between sentences.” In D. Tannen (ed.) Analyzing Discourse: Text and Talk.Georgetown University Press, 71–93.
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.29 (207)
Chapter 6. Compression in interaction
Schutz, Alfred. (1974). The Phenomenology of the Social World. Northwestern University Press. English translation (Walsh, Lehnert, Evanston), [German version 1932]. Sinha, Chris. (2007). “Cognitive linguistics, psychology and cognitive science.” In D. Geeraerts and H. Cuyckens (eds.), Handbook of Cognitive Linguistics. Oxford, Oxford University Press, 1266–1294. Selting, Margret et al. (1998).“Gesprächsanalytisches Transkriptionssystem (GAT).” In Linguistische Berichte 173, 91–122. [Together with Peter Auer, Birgit Barden, Jörg Bergmann, Elizabeth Couper-Kuhlen, Susanne Günthner, Uta Quasthoff, Christoph Meier, Peter Schlobinski and Susanne Uhmann]. Skinner, B.F. (1959). Verbal Behavior. New York: Appleton-Century-Crofts. Turner, Mark. (2001). Cognitive Dimensions in Social Science. Oxford: Oxford University Press Turner, Mark. (2006). “The Art of Compression”. In Turner, Mark (ed.), The Artful Mind: Cognitive Science and the Riddle of Human Creativity, 93–113. Uttal, William R. (2003). The New Phrenology. The Limits of Localizing Cognitive Processes in the Brain. Cambridge, MA: MIT Press. Werner, Gerhard. “Perspectives on the Neuroscience of Cognition and Consciousness.” Manuscript. Williams, Robert. (2005). “Material anchors and conceptual blends in time-telling.” UCSD Dissertation.
Appendix: Transcript glossary The GAT System (Selting et al. 1998) [] = (.) (-), (–) (0.8) :, ::, ::: Á haha hehe so(h)o hm,yes hm=hm,ja=a akZENT ak!ZENT! ? , -
simultaneous talk, overlapping utterances latching micropause longer pauses, with one ‘-’ for each quarter of a second, i.e. 0.25, 0.5, 0.75 sec. pause timed in tenths of a second prolongation or stretching of the sound just preceding them. The more colons, the longer the stretching. glottal stop laughter laughter inside the boundaries of a word receipt tokens, one syllable receipt tokens, two syllables primary accented syllable of a unit extra strong/loud accent final pitch of intonation unit/contour: rise to high final pitch of intonation unit/contour: rise to mid final pitch of intonation unit/contour: level
TSL[v.20020404] Prn:30/01/2008; 12:14
F: PB17006.tex / p.30 (208)
Anders Hougaard
; . ((coughs)) () (solche)
final pitch of intonation unit/contour: fall to mid final pitch of intonation unit/contour: fall to low para- and/or non-linguistic events something is being said, but no hearing can be achieved uncertainty on the transcriber’s part, but represents a likely possibility
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.1 (209)
chapter
Guided conceptualization Mental spaces in instructional discourse Robert F. Williams Lawrence University
Introduction Originally devised to address issues of reference in the philosophy of language, the theory of mental spaces (Fauconnier 1994, 1997) has proven to be such a powerful tool for linguistic analysis that it is considered one of the foundational theories of cognitive linguistics. The present chapter, like others in this volume, aims to demonstrate the utility of mental space theory for analyzing the construction of meaning in real discourse and interaction. This research is important for two reasons. First, mental space theory shows promise for illuminating aspects of discourse which have not been adequately explained by studies using other approaches. Chief among these is the relation between observable behaviors – talk, gestures, and physical manipulation of the environment – and the conceptual operations used to construct meaning. This link between the external and internal is crucial to our understanding of how human cognition and communication function in real-world activity. Second, studies of actual discourse provide a way to test and refine mental space theory itself. Since its inception, mental space analysis has been fruitfully applied to “problem” sentences and linguistic constructions of many kinds, including both invented and attested examples of language use. The time has come to push mental space analysis into the messy world of real human interaction, both to see what insights might be gleaned there and to see how far the theory can take us. With this in mind, the present chapter presses mental space theory into the analysis of instructional discourse. Instruction is a means through which one participant guides the performance and shapes the understanding of another. Prototypically, instruction is directed from an expert (teacher) toward a novice (learner), although interludes of instruction can surface in many kinds of dis-
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.2 (210)
Robert F. Williams
course among co-participants with different conceptions or levels of understanding. How does instruction actually function? What can mental space analysis reveal about the process? These questions will be addressed using samples of instructional discourse drawn from a study of time-telling instruction (Williams 2004). Before we turn to the data, it will be important to introduce three key ideas, one having to do with the unit of analysis and two others with the conceptual framework used to analyze the data.
Three key ideas The first key idea is to define the unit of analysis, i.e., what constitutes the discourse, in a way that does not do violence to the phenomena we wish to understand. In the research presented here, the discourse is taken to include the spoken language of the participants, their gestures, and their physical interaction with objects in the setting. As we will see, this broad unit of analysis is needed to reveal the means through which the teacher shapes the conceptualization of the learners and to illuminate how talk, gesture, and manipulations of objects interact to craft the meanings essential to performing the task (see Hutchins and Palen [1997] for related discussion). The second key idea relates to the conceptual framework used in the analysis. This framework is an elaboration of mental space theory called “conceptual integration theory” or the theory of “conceptual blending” (Fauconnier & Turner 1998, 2002). Conceptual integration theory describes how mental spaces are linked with one another to form conceptual integration networks. These networks produce blended mental spaces that integrate content from diverse inputs, often in novel ways. The creative power of conceptual blending provides a dynamic mechanism for constructing meaning moment-to-moment in specific contexts; it is exactly that power that will be illustrated in the examples that follow. To the theory of conceptual blending we add the third key idea: the use of the physical world to “anchor” blended mental spaces (Hutchins 2005). In an anchored blend, the physical world fixes a constellation of conceptual elements so that we can reason about them without losing track of their intervening relations. A simple example discussed by Hutchins (2005) is the cultural practice of queuing or standing in line. Say we happen upon a place where some people are standing in single file. Materially, this is an arrangement of bodies in the spatial environment. To understand this as an instance of queuing or standing in line, we construct a blended mental space (Figure 1). One input to this blend is the perceptual scene – bodies in space – while the other is a conceptual model of service, a “cultural model” intersubjectively shared by members of our society. This conceptual model is based on the principle that people receive service in the or-
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.3 (211)
Chapter 7. Guided conceptualization
…
2
3
1
Arrangement of Bodies 12
3…
…
Model of Service
Emergent property: The arrangement of bodies encodes the order of service.
Queuing or Standing in Line
Figure 1. The Queuing or Standing in Line blend. The arrangement of bodies anchors the conceptual blend used to determine order of service. (Adapted from Hutchins 2005.)
der in which they arrive, as stated in the maxim “first come, first served.” The model includes ordered slots and a progression toward the experience of receiving service. “Cross-space” mappings link the perceptual scene with the conceptual model: each body is mapped to an ordered slot, while the path of motion of the bodies is mapped to the path of progression toward service. In the blended space, the arrangement of bodies encodes the order of arrival and thus the prescribed order in which people are to be served. Because the bodies anchor the blended space, we are able to reason about who is to receive service when simply by shifting our gaze from person to person, perhaps even counting aloud as we compute our own place in the queue. With the anchored blend in mind, we can act directly on the world to carry out such a computation. Let us elaborate on Hutchins’ example. Say that a person from a foreign culture happens upon the scene and this person’s culture does not incorporate the practice of standing in line. The newcomer needs some guidance. An experienced participant acts as teacher and uses talk, gestures, and other actions to help the newcomer interpret the scene in an appropriate way so that the newcomer can succeed in being served. To understand and behave appropriately, the newcomer must activate – or develop – the requisite conceptual model, map it appropriately to the present scene, and use the blended space and its associated network to generate correct inferences of the relevant form. The expert acting as teacher tries to prompt and shape these conceptual operations through specific talk and actions. What this hypothetical example shows is that instructional discourse is likely to reveal the processes of conceptual mapping and blending. These processes are rapid, automatic, and generally invisible in expert performance, but they become slower,
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.4 (212)
Robert F. Williams
more sequential and controlled, and more directly observable in the interactions between teacher and learner. Instructional discourse opens conceptual mapping and blending to scrutiny – but only if we use a unit of analysis broad enough to capture all of the relevant components.
Analyses of instructional discourse With these three key ideas in mind, let us turn our attention to specific examples of instructional discourse. The examples presented here are drawn from a recent study of time-telling artifacts and practices and how these practices are perpetuated through instruction (Williams 2004). A centerpiece of that study was close analysis of specific lessons in clock-reading. We will use excerpts from two lessons taught on consecutive days in a 1st-grade class to illustrate the power of mental space analysis in uncovering how instruction works. In the first excerpt, the teacher helps the students construct a conceptual blend they will use to read “quarter past” times. In the second, the teacher elaborates that blend with paths of motion and present and future states as she prepares the students to read “quarter till” times.
Building the Clock Quarters blend The first excerpt is shown in Transcript 1 (in the Appendix). Because we are including gesture and manipulation of objects in the unit of analysis, the transcript includes both speech and annotated images of activity. The format, adapted from Goodwin (2003), presents speech in the conventional form used in conversation analysis: text with indications of vocal emphasis, pause length, overlapped speech, and so on (details of the transcription conventions are described in the Appendix). Boxes around speech are linked to images of co-occurring activity. In the images, gestures are annotated in red and manipulations of objects in blue; these appear gray in the printed chapter. Looking at the structure of the discourse, we see several distinguishing characteristics. The teacher does nearly all of the talking, while the learners simply indicate agreement or recognition; these responses provide clues to their understanding. The teacher also glances at the students frequently; we might infer that she is checking for signs of understanding or confusion, an inference supported by her verbal prompts (e.g., “right?” in lines 14 and 25). Looking at the teacher’s speech, we see that she speaks in short, distinctly separated utterances; each utterance is a simple clause or phrase ending with emphasis on a key word (clock, shape, divide, and so on) followed by a brief pause. It seems that she is introducing one new piece of information at a time and waiting for it to be processed. We see that even though the teacher controls turn-taking and takes most of the turns herself, the discourse is constrained by the processing capacities of the six- and seven-
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.5 (213)
Chapter 7. Guided conceptualization
year-old learners. The teacher’s segmented utterances are accompanied by gestures and manipulations of the clock hands and other objects. These play important roles in guiding conceptual mappings onto the clock face and anchoring conceptual elements used to generate time readings. The discourse is also shaped by the teacher’s and students’ shared history of interaction (this lesson was recorded midway through the school year) and the intersubjective awareness that such a history creates. As an example, this particular lesson draws on knowledge recently developed in the class – the dividing of a circle into halves and then into fourths – as well as other knowledge shared by the 1st graders, including a basic understanding of shapes, numbers, and counting, and a rudimentary understanding of time concepts (days/hours/minutes). These resources become the conceptual inputs to the blends the teacher is constructing. To illuminate how the teacher guides the construction of meaning, we proceed through the transcript step by step, detailing how each utterance, gesture, and action contributes to the construction and elaboration of mental spaces and conceptual integration networks. To do this, we diagram the inputs, cross-space mappings, and blends for each line of the transcript, and we identify which elements are being profiled (foregrounded or highlighted) at each moment in the discourse. Space limitations preclude us from walking through the complete analysis here, but an abbreviated account will provide the basic insights (more detail can be found in Chapter 5 of Williams [2004], while a focused analysis of the role of gesture in guiding conceptual mapping can be found in Williams [in press]). In line 2 of the transcript, the teacher lifts an object and places it in her lap while saying “if I take my clock” (Figure 2). The co-timed speech and action map
11
12
11
1
9 4 7
6
1
2
9
3 8
12
10
2
10
3 8
5
4 7
Actual Clock
6
5
Teaching Clock 11
12
1
10
2
9
3 8
4 7
6
5
“Clock”
if I: take my clock
Figure 2. The Clock blend. The teaching artifact is seen as a clock.
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.6 (214)
Robert F. Williams
11
12
1
10
2
9
3 8
4 7
Circle
6
5
“Clock” 11
12
1
10
2
9
3 8
4 7
6
5
Clock-Circle
it’s~the same circle shape (1.7)
Figure 3. Building the Clock Quarters blend (1). Mapping a circle onto the clock face.
“clock” onto the object the teacher is holding. This object is not a clock: it does not keep time. Instead, it is a specially crafted teaching tool made to look like a clock face with hands the teacher can easily manipulate into different configurations. The teaching tool (hereafter called a “teaching clock”) anchors a “Clock” blend, a mental space in which the students see the object as a clock and make inferences accordingly. If they fail to do this, they will miss the point of the entire lesson. In line 6, while saying “same circle shape,” the teacher traces a circle around the perimeter of the teaching clock (Figure 3). The tracing gesture starts at the top of the teaching clock, proceeds in steady motion in a clockwise direction (continuing after the end of the utterance), and ends when it reaches the top of the clock again. Goodwin (2003) points out that a trace has both indexical and iconic components. From our perspective, the indexical component highlights the physical structure that is to anchor the conceptual element, while the iconic component outlines the conceptual element that is to be mapped to that anchor. The gesture superimposes the “circle shape” profiled in the teacher’s speech onto the band around the perimeter of the clock face. The gesture guides the mapping process, but it is the interrelation of that gesture with the speech and with the environmental structure that actually sets up the mapping. For the remainder of the transcribed excerpt, the clock band anchors a conceptual circle.1 . The circular tracing gesture also enacts a path of motion important to time-telling, namely the path of the long hand through a single clock hour. This conceptual content is not described in the accompanying speech. Other analyses show similar instances of gestures adding path struc-
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.7 (215)
Chapter 7. Guided conceptualization
11
12
1 2
10 9
3 8
4 7
11
12
6
5
1 2
10 9
3 8
4 7
6
5
and I divi:de it(h) (1.2) up~and~down he:re (1.0)
Figure 4. Building the Clock Quarters blend (2). Preparing a material anchor and mapping the vertical dividing line.
Next, the teacher manipulates the teaching clock into a non-clock state, forcing the short hand straight up while holding the long hand straight down (Figure 4). Why form a configuration that disrupts the Clock blend? The speech and subsequent tracing gesture (“divide it. . .up and down here” [lines 8–10]) clarify what is happening. By aligning the hands in this way, the teacher prepares a material anchor for a conceptual element: a vertical dividing line. She then uses a mapping gesture (a trace coordinated with speech) to superimpose the conceptual dividing line over its material anchor. The hands fuse into a single, stationary element in the blended space. The teacher’s next utterance, “divide it into halves, right?” (line 14), calls attention to the shapes outlined by the conceptual circle and dividing line: halves of a circle. The teaching clock with its unusual hand configuration anchors the part-whole structure of this Clock Halves blend (Figure 5). In lines 18–20, the teacher moves closer to her conceptual destination, introducing the idea of dividing the clock face into quarters as she picks up a pointing stick and places it horizontally across the clock face (Figure 6). Again she has prepared a material anchor for a conceptual element, and we expect her to trace a horizontal dividing line over the anchor she has prepared. In fact, she does this, but in a very interesting way. Because her hands are occupied, she traces the line with her eyes, fixing her gaze on the point where the stick crosses the 9 and then tracing with her gaze along the stick to the point where it crosses the 3 (Figure ture not profiled in speech. In some cases, the structure enacted solely in gesture is incidental; in others, it is obligatory to correct understanding. For more discussion, see Williams (in press).
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.8 (216)
Robert F. Williams
11
12
1 2
10
½
½
9
3 8
4 7
Divided Circle
6
5
“Clock” 11
12
1 2
10
½
9
½
8
3 4
7
6
5
Clock Halves
divide it into ha:lves, ri:ght?=
Figure 5. Building the Clock Quarters blend (3). Profiling the regions bounded by the conceptual circle and dividing line.
11
¼
12
1 2
10
¼
9
¼
¼
3 8
4 7
11
12
6
5
1 2
10 9
3 8
4 7
6
5
now if I wanted to divide it into: (0.5) quar:ters:
Figure 6. Building the Clock Quarters blend (4). Preparing a material anchor for the horizontal dividing line.
7); this eye-trace is accompanied by the utterance “we go from the nine to the three, right?” (lines 22–27), which describes the path she is tracing. The indexical component of the trace picks out the portion of the pointing stick that is to act as material anchor, while the iconic component of the trace superimposes the horizontal dividing line onto this anchoring structure.
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.9 (217)
Chapter 7. Guided conceptualization
12
11
1 2
10 9
3 8
4 7
11
12
5
6
1 2
10
9
3
8
4 7
5
6
we go from: the:: (0.5) ni:ne to the three:,=ri:ght?=
Figure 7. Building the Clock Quarters blend (5). Mapping the horizontal dividing line onto the pointing stick anchor.
11
12
1
10
2
9
3 8
4 7
11 10 1 8
12
2
7
6
1
4
2
3
4
6
5
5
from~nine~to~three~we~have fou:r pa:rts fou:r equal pa:rts
Figure 8. Building the Clock Quarters blend (6). Profiling the parts bounded by the conceptual dividing lines while pointing to each part with the eyes.
The teacher has already cued the idea of quarters (line 20), and her next pair of utterances (lines 30–31, analogous to line 14) calls attention to the shapes defined by the circle and dividing lines: the four quarters of a circle (Figure 8). Here the teacher executes an eye-point, glancing at each part in turn as she says “four equal parts.” Finally, the utterance in line 33 (“on our clock”) brings the original
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.10 (218)
Robert F. Williams
11
12
1
10
2
9
3 8
4 7
Divided Circle
6
5
“Clock” 11
12
1
10
2
9
3 8
4 7
6
Emergent property: Canonical clock quarters
5
Clock Quarters
on our clock.
Figure 9. Building the Clock Quarters blend (7). Construing the mapped quarter circles as quarters of the clock face.
clock input from line 2 back into profile, completing the Clock Quarters blend (Figure 9). The Clock Quarters blend has an important emergent property. The quarters of the clock face support a specific way to name times: as a quarter past or a quarter till the hour. In the next part of the lesson (not included here), the teacher associates the label “a quarter past” with the upper right clock-quarter. She then has the students practice naming times such as “a quarter past nine.” If the teacher has succeeded in guiding the students’ conceptualization, then when the long hand points to the 3 the students will “see” the Clock Quarters blend, even though some of the anchoring structures – the pointing stick and the vertically aligned clock hands – are absent or otherwise employed. In order for the students to use the blend to generate the “quarter past” time component, the Clock Quarters blend must be stable enough that the whole structure can be anchored solely by the clock face with the long hand pointing to the 3. With repeated use, the blend becomes so well entrenched that it is automatic. In short, it becomes a conceptual model.2 The example described above details how the teacher orchestrates the construction of a materially anchored conceptual blend used in one form of time. Shortcuts to reading the time as “a quarter past” are possible and may be likely: the students can succeed simply by associating the label “quarter past” with the long hand pointing to the right or at the 3 rather than seeing the clock face in terms of the Clock Quarters blend. This impoverished understanding limits the inferences that can be generated. For more discussion, see Chapter 6 of Williams (2004).
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.11 (219)
Chapter 7. Guided conceptualization
telling. The orchestration involves speech activating conceptual models and profiling elements while co-timed gestures couple these elements with structures in the environment (see Goodwin [2007] for a relevant discussion of environmentally coupled gestures). If the students have succeeded in following these thirty seconds of guided conceptualization, they should have an active Clock Quarters blended mental space anchored by the face of the teaching clock and linked to divided-circle and clock inputs. This conceptual integration network can be used to generate quarter-hour components of relative time readings, including “a quarter past,” “a quarter till,” and “half past.” At its present state of development, the Clock Quarters blend is static. It has part-whole structure: a clock-circle made up of four quarters arranged in a canonical configuration. A verbal label (“a quarter past”) has been associated with the upper-right quarter. The next day’s lesson elaborates this blend by adding motion to the conceptualization.
Building the Clock Quarters Motion blend The goal of the second lesson is to read times as “a quarter till.” The lesson begins by recalling the Clock Quarters blend, but then the teacher adds mental spaces for present and future states to the conceptual integration network. Why she does so is not immediately obvious. Why not simply proceed by analogy and associate the label “a quarter till” with the upper-left clock quarter? The reason for the change in conceptualization emerges toward the end of the excerpt, which is shown in Transcript 2 (in the Appendix). In lines 1–6 the teacher re-establishes the Clock Quarters blend by verbally reminding the students (“remember the quarters were. . .”) and resetting the material anchors. Her initial placement of the pointing stick across the face of the teaching clock recalls the previous day’s lesson but sets up a disanalogy between the misaligned clock hands and the vertical dividing line they anchored previously. The teacher solves this mapping problem by removing the pointing stick and realigning the hands vertically, this time with the long hand straight up and the short hand straight down, a reversal that is irrelevant since the hands become a single vertical divider in the blend. The teacher then replaces the pointing stick across the clock face, completing the anchoring of the Clock Quarters blend. As she resets the anchors, the teacher calls attention to where the dividing lines intersect the clock dial: the 12 and the 6 (line 4), and the 9 and the 3 (line 6). These are important because they can be used as landmarks to anchor the dividing lines after the pointing stick is removed and the clock hands are shifted to a time display. Next, the teacher begins a new phase of instruction. The first part of line 8 (“now what if. . .”) cues a mental space for a hypothetical situation (Figure 10). Lines 8–12 describe that situation, filling in the elements of the new mental space. “The big hand” (line 8) profiles the object the teacher touches: the long hand on
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.12 (220)
Robert F. Williams
11
12
1
10
2
9
3 8 7
6
5
4
11
12
1 2
10 9
3 8 7
6
5
4
Hypothetical
now what if th’ big ha:nd
Figure 10. Building a conceptual integration network to read ‘a quarter till’ (1). Cueing a space for a hypothetical situation and profiling the long hand in that space.
11
5 seconds
12
Representation
1 2
10 9
6
1 2 3
8
4 7
12
9
3 8
11 10
4 7
5
Teaching Clock
12
6
5
Actual Clock
S 11
45 minutes
1 2
10
X 9
3
8
4 7
6
5
Hypothetical
moved a:::ll the way arou:n:d (1.3) to the ni::ne.
Figure 11. Building a conceptual integration network to read ‘a quarter till’ (2). Defining a path of motion and compressing time in the blended space.
the teaching clock. “Moved all the way around to the nine” (lines 10–12) describes motion along a path toward a destination. The teacher enacts this motion by moving the long hand from its present position (pointing straight up with its tip over the 12) steadily around the teaching clock in a clockwise direction until the tip
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.13 (221)
Chapter 7. Guided conceptualization
of the long hand is directly over the 9, where she releases the hand (Figure 11). This motion description and its enactment add another structuring element to the ongoing conceptualization: a source-path-goal image schema (Johnson 1987). The basic elements of a source-path-goal schema are a moving object (trajector), a starting point (source), a series of contiguous locations occupied by the trajector as it moves (path), and a destination or endpoint (goal). The trajector moves along the path from source to goal, and at any given moment it occupies some position along the path. In this instance, the trajector is the tip of the long hand – its most salient and freely moving part. The path begins at the 12 and proceeds clockwise around the clock dial. The 9 has been profiled as the goal, but we will see this profile shift in a moment, construing the 9 as the present position and some other location as the goal. When the focus shifts from the Clock Quarters blend to the Hypothetical space, the teaching clock hands change anchoring functions. They are no longer to be seen as a vertical divider; now they are to be seen as clock hands moving with the passage of time. This is the familiar Clock blend from the start of the previous excerpt, the one in which the teaching clock represents an actual clock. The passage of forty-five minutes on an actual clock has been enacted on the teaching clock in a span of five seconds. This significant scaling of time is an example of the kinds of compressions that occur in blended spaces.3 Next, the teacher repeats her description of the hypothetical clock hand motion (“moved all the way around to the nine” [lines 13–14]) as she sets down the pointing stick and grasps the teaching clock at the 9, the present position of the long hand. She then asks “how far is it until it gets up here?” while tracing a path from the 9 to the 12 (line 16). The first part of the question (“how far is it”) asks about distance, not time, continuing the path description from the previous utterance. The second part of the question (“until it gets up here?”) cues a Future space in which the long hand points again at the 12 (Figure 12). This second part overlaps with a tracing gesture that mimics the pace, direction, and steadiness of the clock-hand motion she has just enacted. These commonalities promote a construal of the gesture as continuing that motion. While the teaching clock anchors the present state, the tracing gesture leads to the future state. The path has been redefined as extending from the 12 around the clock clockwise to the 12 again, with 9 being the present position. This path corresponds to the passage of a single clock hour, a correspondence which the teacher does not mention. . Blends compress diffuse elements from diverse inputs into an integrated scene at human scale – the scale at which we are used to interacting with the world. See Fauconnier and Turner (2002) for discussion of common patterns of compression and the principles that govern blending operations.
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.14 (222)
Robert F. Williams
11
12
Identity
1
11
2
10 9
4 7
6
1 2
9
3 8
12
10
3 8
5
4 7
Present (Hypothetical)
11
5
Future
G 12
6
1 2
10
X 9
3 8
4 7
6
5
Counterfactual
how far is it until~it~gets~up here? (2.0)
Figure 12. Building a conceptual integration network to read ‘a quarter till’ (3). Asking about the portion of the path yet to be traversed and indicating the goal.
Both the Present-Hypothetical space in which the long hand points at the 9 and the Future space in which the long hand arrives at the 12 are inputs to the blend shown in Figure 12. In the blended space, the tip of the clock hand is a trajector moving along a continuous circular path of motion clockwise from 12 to 12. The trajector has traversed the portion of the path from the source (12) to its present location at the 9; it still needs to traverse the portion from the 9 to the goal (12). Time is compressed in the blend, but here it is syncopated rather than scaled: the present and future positions of the clock hand are co-present in the blend. For this reason, the blend is labeled Counterfactual. In this Counterfactual blend, these two profiled elements – the present and future positions of the long hand – share the common conceptual base of the clock dial, and they stand in relation to one another in a way that will support reading times as “a quarter till.” This Counterfactual blend has both material and imagistic anchors. The common conceptual base, the clock dial, is anchored by the face of the teaching clock; this was established at the start of the first lesson. The first profiled figure, the present position of the clock hand, is anchored by the hand on the teaching clock. The second profiled figure, the future position of the clock hand, has no material anchor. Instead, it is anchored by the mental image that resulted from imagining the clock hand moving from the 9 to the 12 as guided by the teacher’s gesture. This superimposition of an imagined structure onto a visually perceived structure is an example of “situated seeing,” a way of seeing in the present situation that supports performance of the task at hand. We have already encountered one example
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.15 (223)
Chapter 7. Guided conceptualization Analogy 11
12
1 2
10 9
3 8
4 7
Divided Circle
12
1 2
10
X
5
Counterfactual
G 11
6
9
3 8
4 7
6
5
Clock Quarters Motion X
G
(moves) from he:re (1.3) tuh he:re
Figure 13. Building a conceptual integration network to read ‘a quarter till’ (4). Switching anchors from the clock face to the felt board display and tracing a path along the upper-left quarter circle.
of this: seeing the Clock Quarters blend even after the pointing stick has been removed and the clock hands have been moved to another configuration. Of course, an imagistic anchor lacks the durable presence of the material anchor from which it is derived, but it serves a similar function: temporarily fixing an element in the blended space while inferences are generated. Returning to the excerpt, we see that no students respond to the teacher’s question (“how far is it until it gets up here?”) during the next two seconds (line 17). The teacher then redescribes the path from the present to the future state (“moves from here to here” [lines 19–21]), but this time instead of gesturing over the face of the teaching clock, she leans over and gestures over four quarter circles displayed on a felt board (Figure 13). She touches the upper-left quarter circle at the point that corresponds to the location of the 9 on the clock face, and then she traces a path along the upper edge of this quarter circle to the point that corresponds to the location of the 12. A student raises her hand, is called on by the teacher (line 24), and replies “one quarter or one fourth” (line 26), an answer which the teacher accepts (line 27). What is happening here? When the students fail to respond to the teacher’s question, she repeats her description but gestures over a different object. In other words, she switches anchors. The felt board display anchors the conception of quarter circles; this “divided circle” was one of the inputs to the original Clock Quarters blend. To instantiate the Clock Quarters blend using this new material anchor, the students must map in the opposite direction: from the clock face to
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.16 (224)
Robert F. Williams
the felt board display. The teacher guides this mapping by pointing with her pinky to the location on the felt board display that is the counterpart to the present position of the clock hand. She then executes a tracing gesture exactly analogous to the one she has just performed on the teaching clock, stopping at the location on the felt board display that is the counterpart to the future position of the clock hand. The strategy of shifting anchors to highlight a different conceptual input makes sense here because the construction of a more complex blend with motion from a present to a future state has drawn focus away from the Clock Quarters blend that was reactivated at the start of the excerpt. Gesturing over the felt board display shifts attention to the divided circle input and brings the quarter circles back into profile, restoring the Clock Quarters blend but with the addition of a path of motion and a present position along that path. The result is a more complex Clock Quarters Motion blend, one that incorporates the divided circle, the clock face, a path of motion around the clock from 12 to 12, and present and future positions of the trajector moving along that path. With the addition of the source-path-goal schema defining the path of motion, the static part-whole conceptualization of the previous lesson has been transformed into a dynamic, motion-based conceptualization. In this richer conceptualization, the quarters demarcate portions of the path traveled by the long hand as it moves around the clock. At the time when the student is responding to the teacher’s question, she has two distinct material anchors available to her in the environment: the teaching clock and the felt board display (Figure 14). The video-recording does not provide
11
12
11
1
10
2
9
3 8 7
6
5
12
1 2
10 9
3 8 7
6
5
4
4
G 11
12
1 2
10
X 9
3 8
4 7
6
5
T: Brianne? (0.9) B: one quarter or one fourth.
Figure 14. Building a conceptual integration network to read ‘a quarter till’ (5). Material anchors are available for both inputs to the blended space, but it remains for the student to do the integration.
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.17 (225)
Chapter 7. Guided conceptualization
Analogy 11
12
1
10
2
9
3 8
4 7
6
5
G 11
12
1 2
10
X 9
3 8
4 7
6
5
It’s a quarter, right?=
Figure 15. Building a conceptual integration network to read ‘a quarter till’ (6). Profiling the portion of the clock face to be traversed as the long hand moves from its present position to the goal.
evidence of where she looks, and the teacher simply looks at the student without interacting with any objects during the student’s response. Once the student has responded, the teacher repeats “it’s a quarter” (line 27) while re-executing the tracing gesture she just performed on the felt board display but now again performing it on the teaching clock (Figure 15). This gesture can be construed in two ways at once: as indicating the present-to-future movement of the long hand from the 9 to the 12, and as tracing the edge of the upper-left clock quarter. In the blended space, the long hand has one quarter of its path left to traverse before it reaches its goal. In line 28, the teacher repeats the gesture quickly and only partially while saying “it’s a quarter of the clock,” explicitly profiling the Clock Quarters input to the more complex conceptual blend she has constructed: the Clock Quarters Motion blend (Figure 16). At this point, the discourse shifts from constructing the blend to using the blend to read the time. In lines 29–33, the teacher says, “So we say it’s a quarter till seven. It’s a quarter till seven.” Each time she says “quarter till seven” she executes the same series of gestures: she traces from the 9 to the 12 as she says “quarter,” pauses with her finger on the 12 as she says “till,” and then jumps her pointing finger to the 7 on the word “seven” (Figure 17). This series of gestures provides a clue as to why the teacher has gone to the trouble of adding motion to the conceptualization: to motivate correct reading of the hour component (the 7).
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.18 (226)
Robert F. Williams
11
12
1 2
10 9
3 8
4 7
6
5
G 11
12
1 2
10
X
9
3 8
4 7
6
5
it’s a quarter~of~the~clock=
Figure 16. Building a conceptual integration network to read ‘a quarter till’ (7). Profiling the Clock Quarters input to the new Clock Quarters Motion blend. hold 11
12
1
10
2
9
3 8
4 7
6
5
GL 11
12
1 2
10
XL 9
3 8
7
4 6
5
GS
T: so we say it’s a qua:rter: . . ti::ll:: (1.2) S: se ven! [seven. it’s a quarter ti:ll: (0.3) seven.= T:
Figure 17. Using the conceptual integration network to read ‘a quarter till.’ Tracing the path of the long hand to its goal and then pointing to where the short hand will be in the goal state.
In the previous day’s lesson, when students read times in both absolute and relative forms as “three fifteen” and “a quarter past three,” there was no difference in the hour portion of the readings. Indeed, the teacher said nothing at all about how to read the hour, and the students appeared to have little difficulty. In the cur-
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.19 (227)
Chapter 7. Guided conceptualization
rent day’s lesson, students are reading times as both “six forty-five” and “a quarter till seven.” Here the hour portion of the readings differs. In fact, the hour portions of absolute and relative times are read using different image schemas (Williams 2004). For absolute times like “six forty-five,” the hour is properly read using a container schema (Johnson 1982): if the long hand is contained anywhere in the space between the 6 and the 7, then the hour is “six,” no matter how close the hand is to the 7, even if it appears to be touching the 7. Six is the current hour, and it does not change until the long hand actually reaches or crosses the 7. For relative times like “a quarter past six” and “a quarter till seven,” the hour is properly read using a source-path-goal schema: “past” times are referred to the previous hour (the source), as in “a quarter past six,” while “till” times are referred to the upcoming hour (the goal), as in “a quarter till seven.” The hour component of a relative time reading is the reference hour, a past or future “o’clock” time. For “past” times the current hour and reference hour happen to coincide; for “till” times they do not. From this we see that adding a path of motion to the conceptualization is important because it motivates the use of the proper schema to read the reference hour correctly. The path of motion of the long hand once around the clock from 12 to 12 coincides with the path of motion of the short hand from one number to the next. If we refer to the future (goal) state of the long hand, as the teacher does in her gestures for “a quarter till,” then we should refer to the corresponding future (goal) state of the short hand, as the teacher does when she touches the 7. If we were to use the static part-whole conceptualization to read “quarter till” times, say by associating the label “a quarter till” with the upper-left clock quarter, then there would be no conceptual connection to the reading of the hour portion, no linking of goal to goal. Of course, we could learn to perform correctly by simply remembering to name the upcoming hour whenever we say “a quarter till,” but there would be no particular motivation to do so – in other words, no real understanding. A proper understanding of “a quarter past” should also include the source-path-goal schema, but without a noticeable difference in the output or propensity for error, there is no compelling reason for the teacher to bring motion into the discourse when teaching “quarter past” times.
Comparing the conceptual integration networks The conceptual integration network for the Clock Quarters blend and the elaboration of that network to produce the Clock Quarters Motion blend are depicted in Figure 18. When we compare the first lesson excerpt with the second lesson excerpt, we see that adding motion to the conceptualization invokes mental spaces for different moments in time, including past (trajector leaving the source), present (trajector at some position along the path), and future (trajector arriving at the goal). Identity mappings link the counterparts in these different spaces: trajector to trajector, path to path, and so on. Different time spaces can be blended
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.20 (228)
Robert F. Williams
Felt Board Display
Teaching Clock
Divided Circle
Actual Clock
“Clock”
Clock Quarters HypotheticalPresent
Future
Counterfactual
Clock Quarters Motion
Figure 18. The conceptual integration network used to build the Clock Quarters Motion blend. Material anchors are shown in boxes; the sequence of conceptual blends is in bold.
in a syncopated way, so that present and future states co-exist in the blend, or they can be blended in a scaled way, so that slow processes like clock-hand movement unfold rapidly in the blend. A static blend like Clock Quarters can be a mental image, but a dynamic blend like Clock Quarters Motion is likely to involve mental simulation of unfolding processes. For this reason, motion-based conceptualizations may call upon greater cognitive resources. In terms of cognitive economy, a static conceptualization may suffice to generate relevant inferences in many situations, but a motion-based conceptualization may be called upon when the static one fails to support reasoning in a particular situation. This is, at present, only conjecture, but Figure 18 does demonstrate greater complexity in the conceptual integration network when motion is added.
Discussion Classroom lessons in time-telling consist of practice reading clock times punctuated by micro-episodes of explicit instruction. The teacher sets times on the clock, calls on students to read them, and provides support (verbal cues, pointing, leading questions, partial responses, etc.) as needed to scaffold student performance. Embedded in this practice are brief interludes of instructional discourse like the two samples presented here. What happens when the discourse becomes overtly instructional? The balance of participation shifts so that the teacher does most of the talking and acting while the students watch and signal agreement or confusion. The teacher directs the students’ attention to particular structures in the environ-
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.21 (229)
Chapter 7. Guided conceptualization
ment and provides information as to how those structures are to be construed or acted upon. This is a fine high-level explanation, but we would like to deepen our understanding of what is happening; the mental space analysis allows us to do that. In the analyses presented in this chapter, we see how the teacher’s talk, gestures, and manipulations of objects prompt the systematic construction of conceptual integration networks that are used to generate portions of time readings. Talk serves primarily to cue space-building (including spaces for alternate times and hypothetical or counterfactual states), to activate conceptual models that frame those spaces, to profile elements, and to prompt other mapping and blending operations. Manipulations of objects and gestures over those objects come to the fore when the blended space being constructed is related to interpreting the state of the environment, as in reading a clock. The teacher uses pointing and tracing gestures to highlight environmental structures at moments in the discourse when related conceptual elements are being profiled in speech. These coordinated gestures and utterances prompt mappings that link conceptual elements with material anchors. The teacher also manipulates objects to prepare anchors just before executing gestures that guide the mappings. Through these means the teacher builds anchored blends, linking mental spaces with the world. Once an anchored blend has been constructed, the teacher can gesture over the anchoring structures to generate an inference, as she did when she traced a path on the clock face to read “a quarter till seven,” or she can manipulate the anchoring structures directly to enact a simulation in the blended space, as she did when she moved the long hand around the clock from 12 to 9 to simulate the elapsing of three quarters of an hour of clock time. As these examples shows anchored blends are used to do cognitive work. Finally, among other things, the mental space analysis presented in this chapter helps us distinguish phases in the instructional discourse when conceptual blends are being constructed or elaborated and when those blends and their associated networks are being used to perform tasks such as reading the time. In the construction phase, the teacher’s talk, gestures, and actions guide conceptual mapping and blending; in the performance phase, they guide the generation of task-relevant inferences. In both phases, the teacher actively guides conceptualization.
Conclusion Much can be learned from acting on the environment and experiencing the outcomes of our actions. Much can be learned from observing and imitating the behaviors of others. Such forms of learning fall short when it comes to sustaining practices like time-telling. Because reading a clock is so complex and because ex-
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.22 (230)
Robert F. Williams
pert clock-reading unfolds invisibly, trial-and-error and imitation are unlikely to produce the conceptual integration networks the novice needs to succeed. For this kind of learning, we humans rely upon guided conceptualization: talk, gestures, and actions that prompt and shape the conceptual operations of others. Guided conceptualization is a hallmark of instruction, a central means of sustaining the cognitive sophistication of our species. Mental space analysis provides insight into this distinctly human process.
Acknowledgements I am deeply indebted to Gilles Fauconnier, Edwin Hutchins, and Charles Goodwin for inspiration, advice, and encouragement of this research, and to the Spencer Foundation for a dissertation fellowship that helped support the initial study.
References Fauconnier, G. 1994. Mental Spaces: Aspects of Meaning Construction in Natural Language. Cambridge: Cambridge University Press. (First published in 1985 by the MIT Press, Cambridge, MA.) Fauconnier, G. 1997. Mappings in Thought and Language. Cambridge: Cambridge University Press. Fauconnier, G. and Turner, M. 1998. “Conceptual integration networks.” Cognitive Science 22 (2): 133–187. Fauconnier, G. and Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Goodwin, C. 2003. “Pointing as situated practice.” In Pointing: Where Language, Culture, and Cognition Meet, S. Kita (ed), 217–242. Mahwah, NJ: Lawrence Erlbaum. Goodwin, C. 2007. “Environmentally coupled gestures.” In Gesture and the Dynamic Dimension of Language, S.D. Duncan and E.T. Levy (eds.), 195–212. Amsterdam: John Benjamins. Hutchins, E. 2005. “Material anchors for conceptual blends.” Journal of Pragmatics 37 (10): 1555–1577. Hutchins, E. and Palen, L. 1997. “Constructing meaning from space, gesture, and speech.” In Discourse, Tools, and Reasoning: Essays on Situated Cognition, L.B. Resnick, R. Säljö, C. Pontecorvo and B. Burge (eds), 23–40. New York: Springer-Verlag. Johnson, M. 1987. The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason. Chicago: University of Chicago Press. Williams, R.F. 2004. Making Meaning from a Clock: Material Artifacts and Conceptual Blending in Time-Telling Instruction. Ph.D. diss., University of California, San Diego. Williams, R.F. In press. “Gesture as a conceptual mapping tool.” In Metaphor and Gesture, A. Cienki and C. Müller (eds). Amsterdam: John Benjamins.
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.23 (231)
Chapter 7. Guided conceptualization
Appendix Transcript conventions: bold italics italics ∼ .. o: r:: s::: (h) = [ CAPS ◦
! ? . , (0.5) [boxed speech] (?) (( inaudible ))
Transcript 1
vocal emphasis slight emphasis rapid speech slow speech lengthening of sound aspiration abrupt cut-off conjoined speech overlapped speech loud speech soft speech excited speech rising intonation falling intonation rising-falling intonation pause length in seconds co-occurrence with gesture/action in linked image uncertain transcription transcriber’s comment
TSL[v.20020404] Prn:12/12/2007; 13:48
Robert F. Williams
F: PB17007.tex / p.24 (232)
TSL[v.20020404] Prn:12/12/2007; 13:48
F: PB17007.tex / p.25 (233)
Chapter 7. Guided conceptualization
Transcript 2
TSL[v.20020404] Prn:12/12/2007; 13:48
Robert F. Williams
F: PB17007.tex / p.26 (234)
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.1 (235)
chapter
Looking at analyses of mental spaces and blending / Looking at and experiencing discourse in interaction Alan Cienki
Since it started gaining wider recognition, mental spaces and conceptual integration theory (MSCI), more commonly known as blending theory, has been subject to the criticism that “anything goes,” that there are few limitations on blending analyses (see, for example, the critique in Gibbs 2000). Fauconnier and Turner (1998, 2002) countered some of these claims with their “optimality principles” which describe a number of constraints under which blends work most effectively. Nevertheless, even with the developments in this approach over the years, certain questions continue to arise when new analyses of examples are presented. One such question concerns whose blends are being represented in a given analysis; another pertains to whether conceptual blending remains an invisible process that must be conjured up by the researcher. As will be discussed below, both of these concerns have to do with looking at blends, but in different respects. Both also come to the fore in research on extended contexts of discourse, and particularly on spoken discourse in interaction. In addition to these issues, this commentary will include consideration of other ways in which blends may be sensed, through proprioceptive processes of enactment. A note on the terminology to be used below already gives an indication of points which will play an important role. Henceforth blends here will mean “conceptual blends” unless indicated otherwise. Reference to mental spaces and conceptual integration per se, as opposed to the theory about them, will be noted as MS and CI. Both the word speaker and the more general term producer will be used to indicate the person not just talking but also gesturing, looking, etc., while talking; similarly, hearer(s) and addressee(s) will be used interchangeably, but also imply that this person/these persons have the speaker in view. Collectively the people in these roles will be referred to as interlocutors. The researcher, however, will represent one with a particular outsider perspective on the interaction,
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.2 (236)
Alan Cienki
one who can see and hear the data, but it is not the one to whom the interaction is addressed.
Whose blend is it that we are looking at on paper? Any blending analysis is really an hypothesis about conceptual operations that may be involved in the formulation or interpretation of an expression of some type. Yet in many past analyses it has remained unclear whose conceptual operations are claimed to be represented. One could get the impression that an ideal speaker-hearer (or writer-reader) is assumed, a possible carry-over from the tenets of generative grammar, despite the fact that cognitive linguistics (including MSCI) arose largely in reaction against those tenets. Note similar critiques levelled in the past against conceptual metaphor research, e.g., Gibbs and Perlman’s (2007) cautions regarding claims made about decontextualized examples. Perhaps the analysis is really about the researcher’s interpretation, which may or may not also apply to the producer or addressee of the given expression? In such cases in which one is describing the possible ways in which one can make sense out an expression (be it verbal, imagistic, or otherwise, or some multi-modal combination), rather than saying the expression “has” a certain mental space structure, it might be more prudent to argue that what is being analyzed “could be explained in terms of ” MSCI. See, for example, Dancygier’s approach in this volume to blending analysis as a way of capturing some of the processes which can explain how a reader could construct a complete, sequential story out of an incomplete set of narrative events. Other papers in the volume use MSCI as a way of analyzing excerpts of spoken interaction as they transpired, and for blending analysis this is an important distinction on several levels. One is that the producer and the addressee (here: speaker and intended hearer) of the expressions analyzed are known and specific, and some background information is known about them; in this way, hypotheses about the blends proposed are specific in scope. They are grounded in the physical realities of the situation in which expressions were used. So in Pascual’s analysis of courtroom interaction, it is clear who the speaker was, and when and where he was speaking. From the unique aspects of the roles which speakers in the courtroom play, the researcher has a lot of information about the intentions motivating speakers’ utterances in that setting. Even though there are multiple addressees for any utterance, leaving open the possibility for multiple interpretations, we know in the context of the courtroom some of whose interpretations those are. In Williams’ analysis, based on videorecorded spoken interaction, the hypotheses about the blends have even more physical grounding. First, we know that the blends being analyzed are intentionally being constructed, and this is taking place as part of a lesson using a physical artefact (the teaching clock). Second, the researcher (and reader of the
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.3 (237)
Chapter 8. Looking at analyses of mental spaces and blending
chapter) can see other explicit cues beyond the words used which lend support to the hypotheses made about the MS and CI being used by the speaker, including the speaker’s manual gestures and eye gaze direction. In addition, the researcher had access to the cues of recognition and understanding on the part of the students for the points which the teacher was making. A final factor important in analyzing blends involved in the production and comprehension of spoken discourse is that of time. The sequential nature of how an expression or sequence of expressions was produced and therefore available for reception, locked to the specific time frame of its production, has particular implications for the analysis of blends in realtime. The researcher can review the audio or video recorded material countless times, but for the participants themselves, the interaction was a fleeting, one-time event, a fact which also constrains the claims that can be made about blends which individuals may have constructed in real time. A novel approach to blending presented in this volume is one based on microsociological analysis, as described by Hougaard. Here the claim made is that the proposed blends are based on shared information, and not on possible psychological processes. As a consequence, the model is “an abstraction which may not be real at any actual interactional point” (Hougaard, this volume) – thus providing yet another answer to the question of whose blend is being represented, namely: possibly no one individual’s. As is clear from the chapters in this volume, MSCI has developed in various ways, not just in terms of the ways in which it is being applied, but even in the theoretical claims being made, such that we now have different blending theories. Given that, it is more important than ever to be clear about what one is referring to (including whose MS and CI are being depicted) in one’s analyses.
Seeing speakers setting up their mental spaces and blends We can see a move (not necessarily step by step, but generally) to increasing degrees of contextualization if we think about the study of blending in individual phrases or sentences, to narrative texts written for mass audiences, to images in print where spatial composition plays an important role, to audio recordings of speech in which the manner of utterance is inherently available, to videorecorded material, especially videorecordings of naturally occurring discourse and interaction. For the study of spoken discourse, we encounter increasing complexity as we take more modalities and media of expression into account – not just the words spoken, but also the prosody, positions of the speaker and addressee, eye gaze during the interaction, gestures (particularly of the hands), and the artifacts at the interlocutors’ disposal. Let us consider these factors, some which were raised above, in
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.4 (238)
Alan Cienki
a little more detail. We will see that the question of whose mental spaces we are looking at has another angle with respect to the kinds of data one is analyzing. If we take mental spaces as the “small conceptual packets constructed as we think and talk, for purposes of local understanding and action” (Fauconnier & Turner 2002: 40), then the study of what speakers and addressees are doing in various modalities of behavior while speaking and listening can provide us with even more detailed clues as to how speakers are setting up their mental spaces while they are talking. This argument applies to the study of speakers more so than addressees, given listeners’ relative silence and much less frequent gesturing while in this role, but addressees are also not stone-still monoliths, and they provide certain bodily cues of reception, especially if they have a more active style of listening. If we focus just on spontaneous manual gesture with speech, we know that speakers’ gestures may correlate with the expression of ideas as objects or as spaces (Sweetser 1998). Indeed, the ideas which a speaker may refer to gesturally as entities can be as broad as conversation topics. In this regard, “[d]iscourse-structuring gestures which highlight different parts of a logical argument can be seen as representing the speaker’s mental spaces (about the organization of the argument) in the form of physical spaces” (Cienki, in press). Thus in video-recorded conversations I have elicited between pairs of U.S. American students, discussing how they take exams at the university, one student concludes: “It depends on the student, but it also depends on the teacher.”
with the comma indicating the end of a continuing intonation unit, and the period/full stop – a final intonation unit (see Du Bois et al. 1993 for details of this convention for the transcription of speech). Thus we have grammatical evidence (in the use of separate verb phrases, the second of which begins with a contrastive “but”), in addition to prosodic evidence (given previous research on intonation units as reflections of a speaker’s idea units [Chafe 1994]), with which to argue that the speaker has distinguished these ideas as separate “small conceptual packets,” similar to what Oakley and Coulson argue in this volume. But looking at the speaker in the original video data, we also have additional clues. When uttering “depends” in the first intonation unit, she sets her two hands, palms down, into a space on her right side. She then lifts them and places them back down, but on her left side, when she utters “also” near the start of the second intonation unit. The two gestures lay out the two conditions in her argument as separate spaces in front of her. This can be claimed as additional evidence that the speaker’s distinction of these ideas can be helpfully thought of in terms of distinct mental spaces. (See also Sweetser 2007 on the use of gesture data as visual evidence for supporting a mental spaces analysis.)
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.5 (239)
Chapter 8. Looking at analyses of mental spaces and blending
In other examples we find speakers not only indicating different physical spaces for different ideas, but also pointing back and forth between them, relating elements from one to elements of the other. We see an example of how this can be very physically anchored in Williams’ chapter where the teacher uses pointing between the divided circle on the felt board and the teaching clock in order to tie the formal integration on the clock face to the conceptual integration which one needs to use the clock to tell time (see Turner & Fauconnier 1995 on integration on the formal versus conceptual levels). We see an instance of this with abstract reference in an example analyzed in Cienki (1998: 197–198) in which a student compares a test given in a course in the past (a “back-test”) with a hypothetical test to be given in the future. The student argues about the ethics of using the back-test one might have obtained in order to study for the upcoming test, which might be the same test if the teacher re-uses old materials from when the course was offered in the past. He consistently refers to the back-test by pointing to a space on his left, and to the imagined future test by pointing to an area on his right. What is interesting here are the pointing gestures which go back and forth between the two spaces several times; these could be interpreted as indicating the kinds of identity relations (connectors, vital relations) between the two ideas which have been set up as physical spaces. (See Parrill & Sweetser 2004 for a detailed analysis of gesture in a study on conceptual integration.) In these examples based on videorecorded spoken interaction the study of MS and CI is no longer hypothesizing about processes as if they are happening in a black box. Taking the strong position about gesture in relation to thought, one could say that, “Gestures (and words etc. as well) are themselves thinking in one of its many forms – not only expressions but thought, i.e., cognitive being, itself ” (McNeill & Duncan 2000: 156, emphasis in original). Even if one does not want to go so far, one can at least cautiously say that what seemed like a black box actually gives off a lot of light. There are multiple forms of data, some of which we can see, which can provide converging evidence for arguing that a given speaker in a particular setting at a given moment is setting up distinct “conceptual packets” for purposes of local understanding (that is, mental spaces) and may be using them as the basis for conceptual integration. Taking the multimodal nature of in-person interaction into account, we (as researchers) can not only see cues of speakers setting up their own mental spaces, we can also see how the addressee may engage with them. This evidence may take several forms. One is in terms of signs of “pick-up” by the addressee, cues that those listening to and viewing the speaker have understood the point just made (such as the students’ responses to the teacher in Williams’ chapter). Another is gestures performed within the addressee’s line of sight, particularly if their motion is large enough that it engages the addressee’s gaze to move with them. (See Müller [in press] on cues of “activation” of a metaphor on the part of a gesturing
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.6 (240)
Alan Cienki
speaker which may, in turn, attract the gaze, and thus attention, of the addressee.) A third is mirroring by the addressee, performing a return gesture to one that the speaker has made which copies important aspects of its form and movement (Fornel 1992). This process can go back and forth between interlocutors as they alternate roles as speaker, possibly with further developments of the gesture or the gesture scene which one of them set up but which the other picked up on. For example (as discussed in Cienki and Müller 2006) one student talking about having a “better base” of knowledge now that she is a fourth-year student makes a horizontal sweeping gesture with her flat left hand (palm down) when she says “better base,” and her interlocutor agrees saying, “Yeah. It builds,” during which she moves her right hand upward in steps, palm down, as if patting increasingly higher surfaces. The first speaker replies, “Yeah, it all builds on each other,” and moves her flat right hand slowly upward, palm down. In this way, the two speakers, who are sitting facing each other, collaboratively build an abstract object in space with their gestures, each showing different aspects of the same scenario which integrates metaphors for knowledge as a structure. One could argue that this provides evidence of how the speaker and addressee may be setting up and running the same (or comparable versions of a) blend, and elaborating on it, such that it goes back and forth and develops in the process of interaction. These kinds of examples argue against the strong version of the claim that the locus of thought which is being described via blending analyses is inaccessible to research. I think the death knell sounded for cognitive science of being “constantly at risk” of being led into “a sheer speculative abyss of unverifiable hypotheses” (Hougaard, this volume) is premature; there are methods of researching various kinds of cognitive processes, including conceptual integration, which can be justified based on fine-grained analysis of interlocutors’ behaviors beyond just the words that they utter. Cognition in interaction is multiple things happening at once, both onstage and backstage. As a side note the study of written narratives provides an interesting point of contrast. As Dancygier observes in her chapter, certain forms of information can more easily be withheld in a written narrative, such as which of two characters is the lover of a third one. However this would be more difficult to hold off in a film rendition of the same story in which either the emotional expression, or the lack of it, on a character’s face would provide a visual cue to the audience to integrate certain concepts in certain ways and not others. The modality/ies of communication affect not only whether mental spaces can be cued for addressees, but also when and how.
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.7 (241)
Chapter 8. Looking at analyses of mental spaces and blending
Feeling mental spaces One emphasis in Brandt’s chapter is “the pragmatic in the semantic,” a recurring theme in the volume. Making mental contact with a referent in a way which is affectively engaging is more effective in many ways than doing so without such engagement. As Brandt expresses it, “Vivifying concepts by having them be played out in the imagination, animating them, rather than just having them be pointed to as meanings abstracted from experience makes these discourse referents more present experientially which appears to increase the rhetorical effect.” In terms of spoken interaction, the process of referring to something (someone, someplace, some action, etc.) is always animated in that the human (or even virtual) speaker’s degree of use of prosodic and gestural expression is inherently tied to the referential words being uttered. This raises interesting questions for the types of fictive interaction discussed by Pascual and Brandt, and much can still be researched about how it relates to such embodied forms of expression as voice quality (particularly quotation quality), marked intonation contours, facial expression, body position (torso shift to indicate other-character mode), and gesture directedness, all of which might be considered types of presentational performance. It is also interesting to consider the words discussed in Pascual’s and Brandt’s work which speakers, but also writers, use to introduce real or generic quotations (in English, for example, so, like, and all). In speech these may be accompanied by additional embodied cues of spoken utterance to indicate quotation style, as noted above, contributing to the formal expression of the conceptual blend of speaker-being-quoted-as-speaker-present-now. For written language, it is interesting to follow up on their observations that authors resort to other devices to try to prompt the same blend in the reader, such as marked font use (e.g., all capital letters) or mid-phrase punctuation (as in the example of “I Do! Dishes” cited in Brandt’s chapter) – elements which evoke marked prosody or stress which invoke simulation of the words in aural (sonic) form. As a case in point, consider the “I do! Dishes” example cited by Brandt in connection with the phenomenon of fictive interaction within a clause (developing on Pascual 2006). In addition to the analysis provided, one can note additional layers of interpretation which can be explained via the invocation of other input spaces which afford additional blends. In American English, this brand name recalls the formulaic phrase I don’t do dishes! and the related family of constructions based on I don’t do ___ ending with windows, laundry, floors, and potentially other household entities that require occasional cleaning. (Note that to do the dishes in American English corresponds to British English to do the washing up, usually after dinner/the evening meal.) But the phrase I don’t do dishes is not a simple statement of a habit that one doesn’t engage in; it plays on the assumption that certain people are expected to do/clean the dishes (or laundry or windows, etc.), but that the
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.8 (242)
Alan Cienki
speaker feigns being too superior to engage in such manual labor. The utterance presumes a self-conscious “diva” attitude of glamor, and is often uttered with a corresponding looking-down-one’s-nose delivery. It is therefore not surprising that a Google search of the phrase I don’t do dishes to test its use on the internet turned up as one of its first hits the web page [http://www.opi.com/Whirlpool.asp] (30 July 2007) for the nail polish company OPI which announces, “OPI teams up with Whirlpool Brand to offer the I Don’t Do Dishes™ line of Nail Lacquers – perfect for any woman who hates washing dishes, but loves fabulous nails!” (Whirlpool is a company well-known in the United States for making kitchen appliances, including dish washers.) To someone familiar with the uses of the negative expression I don’t do dishes which are laden with theatrical self-aggrandizement, the brand name “I do! dishes” can be read with an added layer of humor, perhaps even perceived as naively cute for the intended audience of newlyweds – the vision of traditional gender roles being self-consciously reinforced, according to which the wife would obviously (exclamation point) do the dishes . . . and presumably not worry about her nail polish. In terms of MSCI, this background knowledge which some readers may have can provide them with an additional input space available for potential interpretation of the ad on yet another level via a new blend which engages a heightened affective response; invoking the performative aspects of the more common expression I don’t do dishes!, this response may be humorous, cynical, or both. We see that blending theory provides an important tool for the semantic analysis of spoken discourse (and of written texts which incorporate devices from spoken language) particularly because it makes tractable these elements which some would label “non-verbal” that are part of the multi-modal nature of spoken language usage. These include elements of context not explicitly mentioned in words, such as referents indicated gesturally, and the speaker’s, or the quoted speaker’s, affective stance in the process of utterance. In this sense, I agree with Brandt’s claim in her chapter that “an essential component in fictive verbal interaction is the performative staging of an enunciation – what we could call ‘dramatized enunciation’.” The significance of this becomes even more apparent if we think of the “experienced animation” of concepts, discussed in that chapter, in connection with theories from cognitive psychology of perceptual symbols, the construction of situation models, and the running of mental simulations of actions during language comprehension (e.g., Barsalou 1999; Glenberg & Robertson 2000; Zwaan 1999). There is great potential for integrating these approaches with MSCI as a means of analyzing the “pragmatic in the semantic” and getting at the basis of the felt, enacted experience of meaning-making. In addition, there are interesting possibilities for blending theory to make links not only to the study of how performative speech acts and rituals function and achieve the effects that they do, cognitively and socially (explored, for example, in
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.9 (243)
Chapter 8. Looking at analyses of mental spaces and blending
Sweetser 2000), but also to performativity as it has been engaged theoretically in certain branches of social theory, such as in the kinds of discourse analysis which examine how social structures and power relations are reproduced through certain practices of representation, positioning, image creation, and so on. Thus blending theory could engage with performativity as it has played a role in fields such as gender (Butler 1990) and queer (Sedgwick 1993) studies, as Sweetser suggests, but also other fields, such as economics (e.g., Mackenzie et al. 2007), and science and technology studies (e.g., Pickering 1995).
Conclusions As MSCI is developing, greater scrutiny is duly being paid to how we (should? can?) look at blends. As analyses become formalized into diagrams, and the contexts which they are describing become abstracted away into circles and lines, we should not lose sight of who is being claimed to be constructing the illustrated blends (if indeed this is being claimed about any individuals). Particularly if the data come from spoken language and interaction, then actually looking at, as well as listening to, the participants in the interaction can provide important cues about the mental spaces and blends which they may or may not have set up. Noticing the expressive quality with which each utterance was performed can have particular significance in the researcher’s interpretation of the affective nature of the concepts being expressed and in the interpretation and analysis of the speaker’s viewpoint. For written discourse, the cues by which quotation quality and fictive interaction can be marked, as discussed in this volume, similarly provide the researcher with evidence of the MS and CI which the writer likely employed in composing the relevant passage. Given the argument that such devices are likely to facilitate relevant sensory and affective simulation on the part of the reader (including the researcher), these cues provide further grounds for justifying blending analyses pertaining to the reader of such a text. The way in which a MS and CI analysis can bring this out suggests potential for applying this theoretical approach in the study of performativity in social science research on discourse analysis. One methodological challenge which blending researchers are still confronted with is that the conceptual integrations being described are dynamic processes, a point which becomes even harder to ignore when one is analyzing data like spoken discourse which transpires in real time. Perhaps in the future it may become commonplace to use dynamic media (such as animated graphics) in order to better render blending analyses as characterizations of the active processes that they are. Overall we see that taking on the complexities of real sequences of discourse in blending analyses can be revealing for understanding how we make sense out
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.10 (244)
Alan Cienki
of language use in more extended contexts, and can also suggest new methods for working with blending theory itself.
References Barsalou, L. 1999. “Perceptual symbol systems.” Behavioral and Brain Sciences, 22:577–660. Butler, J. 1990. Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge. Chafe, W. 1994. Discourse, Consciousness, and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: University of Chicago Press. Cienki, A. 1998. “Metaphoric gestures and some of their relations to verbal metaphoric expressions.” In Discourse and Cognition: Bridging the Gap, J.-P. Koenig (ed.), 189–204. Stanford, CA: Center for the Study of Language and Information. Cienki, A. In press. “Why study metaphor and gesture?” In Metaphor and Gesture, A. Cienki & C. Müller (eds.). Amsterdam: John Benjamins. Cienki, A. & C. Müller. 2006. “How metonymic are metaphoric gestures?” Paper presented at the Second International Conference of the German Association for Cognitive Linguistics. Munich, Germany, October 5–7, 2006. Du Bois, J., S. Schuetz-Coburn, S. Cumming & D. Paolino. 1993. “Outline of discourse transcription.” In Talking Data: Transcription and Coding in Discourse Research, J. A. Edwards & M. D. Lampert (eds.), 45–87. Hillsdale, NJ: Lawrence Erlbaum Associates. Fauconnier, G. & M. Turner. 1998. “Conceptual integration networks.” Cognitive Science, 22.2:133–187. Fauconnier, G. & M. Turner. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Fornel, M. de. 1992. “The return gesture: Some remarks on context, inference and iconic gesture.” In The Contxtualization of Language, P. Auer & A. di Luzio (eds.), 159–176. Amsterdam: John Benjamins. Gibbs, R. 2000. “Making good psychology out of blending theory.” Cognitive Linguistics, 11:347– 358. Gibbs, R. & M. Perlman. 2007. “The contested impact of cognitive linguistic research on the psycholinguistics of metaphor understanding.” In Cognitive Linguistics: Current Applications and Future Perspectives, G. Kristiansen, M. Achard, R. Dirven, & F. Ruiz de Mendoza (eds.), 211–229. Berlin: Mouton de Gruyter. Glenberg, A. & D. Robertson. 2000. “Symbol grounding and meaning: A comparison of highdimensional and embodied theories of meaning.” Journal of Memory and Language, 43:379– 401. Mackenzie, D., F. Muniesa, & L. Siu (eds.). 2007. Do Economists Make Markets? On the Performativity of Economics. Princeton, NJ: Princeton University Press McNeill, D. & S. Duncan. 2000. “Growth points in thinking-for-speaking.” In Language and Gesture, D. McNeill (ed.), 141–161. Chicago: University of Chicago Press. Müller, C. In press. “What gestures reveal about the nature of metaphor.” In Metaphor and Gesture, A. Cienki & C. Müller (eds.). Amsterdam: John Benjamins. OPI Products, Inc. . Consulted 30 July 2007. Parrill, F. & E. Sweetser. 2004. “What we mean by meaning: Conceptual integration in gesture analysis and transcription.” Gesture 4:197–219.
TSL[v.20020404] Prn:12/12/2007; 12:10
F: PB17008.tex / p.11 (245)
Chapter 8. Looking at analyses of mental spaces and blending
Pascual, E. 2006. “Fictive interaction within the sentence: A communicative type of fictivity in grammar.” Cognitive Linguistics 17:245–267. Pickering, A. 1995. The Mangle of Practice: Time, Agency and Science. Chicago: University of Chicago Press. Sedgwick, E. 1993. “Queer performativity: Henry James’s The Art of the Novel.” GLQ 1:1–16. Sweetser, E. 1998. “Regular metaphoricity in gesture: Bodily-based models of speech interaction.” Actes du 16e Congrès International des Linguistes (CD-ROM), Elsevier. Sweetser, E. 2000. “Blended spaces and performativity.” Cognitive Linguistics 11:305–333. Sweetser, E. 2007. “Looking at space to study mental spaces: Co-speech gesture as a crucial data source in cognitive linguistics.” In M. Gonzalez-Marquez, I. Mittelberg, S. Coulson, and M. Spivey (eds.), Methods in Cognitive Linguistics. Amsterdam: John Benjamins. Turner, M. & G. Fauconnier. 1995. “Conceptual integration and formal expression.” Metaphor and Symbolic Activity 10:183–203. Zwaan, R. 1999. “Embodied cognition, perceptual symbols, and situation models.” Discourse Processes 28:81–88.
TSL[v.20020404] Prn:12/12/2007; 12:13
F: PB17009.tex / p.1 (247)
chapter
“Mental spaces” and “blending” in discourse and interaction A response Gitte R. Hougaard
Through their work (1994, 1996, 1997, 1998 and 2002) Fauconnier and Turner have made very important contributions to the way research can (best) understand, that is conceptualize how human beings go about constructing meaning or, to use a term that I prefer, how they make sense of their life-world. The present volume testifies to the importance of Fauconnier’s and Turner’s work as it takes issue with the theory of mental spaces and addresses some of the empirical questions that theories always give rise to, in this case for instance: Can empirical research document that a) blending processes are the way human beings construct meaning and b) if so, that these processes are central to meaning construction in all contexts? The contributions attempt to answer these questions by investigating blending in the context of discourse. Further, and in some respects more importantly, the analyses in this volume are characterized by using different notions of discourse, context, intersubjectivity, mind and cognition as well as different methods of analyzing blending. All this work is carried out in order to find empirical evidence of mental spaces and of blending. This goal makes the volume very important to mental space and blending theory: depending on assumptions and methods the empirical investigations are not just further applications of the theory, instead they set out to test theoretical hypotheses, and may as a consequence make as well minor as major theoretical adjustments relevant, strengthen or weaken the theory or as a matter of fact occasion that the theory falls. The papers can roughly be divided into two groups: one (Oakley & Coulson, Dancygier, Oakley & Kaufer, Pascual and Brandt) that take the individual as the starting point of analysis and papers that take the shared communicative situation as the starting point. With the first group, the individual is not considered to be isolated from the social world but the social world is external to the individual and it is a world that the individual can choose to interact with or not. In
TSL[v.20020404] Prn:12/12/2007; 12:13
F: PB17009.tex / p.2 (248)
Gitte R. Hougaard
discourse, the individuals of course interact with the world and the analyses aim at describing how meaning between individuals is constructed ’intersubjectively’ in that specific context. Meaning is embedded and constructed in the individual mind and it is through the exchange of meaningful actions, utterances and words that an intersubjective meaning is achieved. Thus, when participants in discourse construct a blend in this tradition, they typically do so individually in their ’isolated’ minds prior to expressing the end-product, the blend. In the discourse, then, they express the blend and in and through it they strive to make the co-participant construct the same blend so that they in the end have an intersubjective understanding. This group of papers documents that blends can be found, described and analyzed in discourse. So, for instance, Pascual’s contribution describes how speakers construct blends as they set up mental scenes in which they talk or interact with for instance dead persons as if they were alive. And Coulson & Oakley’s paper analyzes elegantly how the construction of blends not only can be found in discourse but that speakers provide cues for blending in terms of intonational contours for the listener to determine whether structure from an existing mental space or a new space must be constructed. However, the question still remains whether the conceptualization of the participants’ actions as expressions of blending processes can only be ascribed to the analysts’ understanding and not to the participants’ own understanding of the discourse. Does this matter, one could ask? If other approaches are interested in how the participants themselves understand their actions is that then not just an alternative research interest that complements other interests like the ones mentioned above? No, one has to answer, not if we take the results of the other group of papers in this volume seriously, namely that blending is something that participants can do but do not necessarily do. That raises the question whether participants can turn everything into a blend, because, if so, then so can the analysts. Analysts that are capable of constructing the blend demonstrate how they understand the theory but do not, strictly speaking, support empirically that blending is central to meaning construction. This empirical question, I believe, is answered by the other group of papers (A. Hougaard and Williams). These papers rely on ethnomethodology and conversation analysis which aim at understanding how people make themselves understandable to each other and in doing so operate with notions that differ dramatically from the ones mentioned above. So for instance, discourse is not understood as a context for sense-making but as the product of concerted, sensible, recognizable, analyzable and understandable actions. Context itself is thus constructed by the participants in discourse, or rather interaction, and made meaningful by them. Context is not only something that they act in but something that they do as they act (in it). In other words: every action is done in the context of a previous action and at the same time it renews that context. In this understanding, understanding/
TSL[v.20020404] Prn:12/12/2007; 12:13
F: PB17009.tex / p.3 (249)
Chapter 9. “Mental spaces” and “blending” in discourse and interaction
sense/meaning is not “there” before the action and the action itself does not carry a self-evident pre-constructed meaning that just needs to be decoded: understanding is worked at constantly. It is something that participants in interaction achieve socially in the here-and-now of the actions that they carry out. Meaning occurs (or is done) as participants do the actions and it is defined or established subsequently in the next turn at talk – that is interactionally. Unterstanding/meaning/sense is thus not achieved until some next turn has been produced, and when done the meaning is done/established outwardly, in the social open. Intersubjectivity is not the result of the exchange of individual meanings but a co-construction of meaning, and mind (if talked about at all) is thus social as well. With the paradigms of ethnomethodology and conversation analysis A. Hougaard and Williams address the empirical question of whether people do blending processes when they construct meaning. Hence, they aim at investigating the hypothesis of blending by using a method that is meant to capture how people understand what they do when they interact. Interestingly, the authors find two different things which however together indicate the same crucial thing, namely that blending is not the way we think. Hougaard shows how participants in a call-in show compress the sense (gist or upshot) of prior sequences of talk in formulations that are subsequently responded to ((dis)confirmed) by next speaker. Hougaard thus describes how, the participants in the interaction, not only make use of compression as a technique for constructing meaning but also how they show one another that they understand one another to make use of that technique. Now in the blending literature, ‘compression’ is considered to be part of the blending processes. However, Hougaard argues that ‘compression’ described in terms of ‘formulations’ should not simply be treated as a blend on a par with other ‘blends’ as these differ from interactionally done blending processes (i.e., blending as an interactional phenomenon, as described in Hougaard 2005). Williams demonstrates such a blending process. His paper studies how a teacher instructs 1st graders how to read the time from an analogue clock. In and through talk, gestures and actions she guides the conceptualizations of the learners during the course of the activity. In other words Williams describes how a blending process is directly observable and thus analyzable in the interaction between teacher and learner. William’s work differs from Hougaard’s with regard to the analysis of the interactional support that co-participants in the interaction give to their (the analyst’s) analysis. The teacher in William’s paper seeks the responses of the learners before she goes on but the quality of the data does not allow the author to describe what these responses are. Instead he relies on the teacher’s response to the learners’ responses. In summary, like the other papers in this volume, interactional studies (Hougaard 2005 and this volume and Williams, this volume) document that blending can be found in discourse. Even with all the interactional notions which differ
TSL[v.20020404] Prn:12/12/2007; 12:13
F: PB17009.tex / p.4 (250)
Gitte R. Hougaard
from those within ”classic” mental space theory, these studies too document that blending is something that people do in interaction; that blending is accomplished socially; that blending is recognized as a specific process. In sum they argue that blending and blends are recognizable to the participants themselves. Concluding this commentary, with Hougaard as an exception (see Hougaard 2005 though) the contributions in this volume seem to support mental space theory as they all find empirical support for the theory. However, the analysis of blending as a technique (Williams) and as an interactional phenomenon (Hougaard 2005) goes much further. The consequences of those analyses, which find support in Hougaard’s analysis of ‘formulations’/‘compressions’ in this volume are as a matter of fact not likely to support mental space theory, i.e. blending, as a theory of the way we think. Blending in interaction is one alternative to numerous other socio-interactional phenomena (with ‘compression’/‘formulations’ as one example) and techniques. Interactional analysis of blending suggests thus that people can choose to construct a blend (or rather that they can do a blend) and that they can choose not to. People thus treat blending as a way to think (ourwardly or do) not the way we think. In this way, Williams’ and Hougaard’s (this volume and 2005) analyses are fatal to the hypothesis of blending as ’the way we think’.
References Fauconnier, G. 1997. Mappings in Though and Language. Cambridge: Cambridge University Press. Fauconnier, G. and M. Turner. 1998. “Conceptual Integration Networks.” In Cognitive Science, 22:2 (April–June), pp. 133–187. Fauconnier, G. and M. Turner. 1994. Conceptual Projection and Middle Spaces. an Diego: University of California, Department of Cognitive Science Technical Report 9401. Fauconnier, G. & M. Turner. 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books Hougaard, A. 2005. ‘Conceptual disintegration and blending in interactional sequences: A discussion of new phenomena, processes vs. products, and methodology. In Journal of Pragmatics, 37/10. Special issue: 1653–1685. Turner, M. 1996. The Literary Mind. New York: Oxford University Press.
TSL[v.20020404] Prn:12/12/2007; 12:23
F: PB17010.tex / p.1 (251)
chapter
Reflections on blends and discourse Paul Chilton Lancaster University
One of the theoretical innovations to emerge in this collection is the addition of grounding to the conceptual integration network, which, though it was originally conceived in order to model on-line discourse processing, is in itself purely cognitive. The term “grounding”, incorporated in several diagrams in the book and introduced explicitly by Oakley and Coulson, is used broadly in the sense of Langacker (1987: 126–129). The advantage of introducing grounding is that it helps to motivate conceptual integration theory’s (CIT) claim to be offering a theory of discourse processing. More specifically, it helps to constrain the kind of processing that can be carried out in the blend. In de-contextualised blending diagrams and their verbal summaries it is often not clear why, which or how many inferences and other cognitive operations contribute to the “running” of the blend. In principle, these questions may simply not be answerable; CIT is interested in what is possible within this category of cognitive phenomena. Discourse processors can, within limits, do what they like with linguistic input susceptible of prompting blending – as Brandt points out, they may choose not to run the blend at all. However, if the blend is run there are constraints, and it is these we should be interested in. If we want to develop a theoretical framework, we at least need to know what kinds of constraints may structure any possible blend. The question that this volume of essays prompts us to pursue is: what is the nature of these discourse constraints? Inevitably, this question leads to a consideration of “context” – surely one of the most over-used and least well defined of pragmatic notions (but note the developments in Duranti & Goodwin 1993). Oakley and Coulson say, albeit fleetingly, that integration networks “must be understood as operating in dialogue with grounding” and their blending diagrams incorporate an innovation initiated by Brandt and Brandt (2005) that is intended to model this dimension of the blending process. The concentric circles they adopt in order to model the variable grounding context seem a fair iconic representation of the kinds of knowledge involved, so long as the circles are understood to be
TSL[v.20020404] Prn:12/12/2007; 12:23
F: PB17010.tex / p.2 (252)
Paul Chilton
porous. Discourse processors need the kind of information arising in the immediate face-to-face situation, but they may need also information stored in long-term memory relating to previous conversations, cultural knowledge of varying degrees and complexity, extending out into what Oakley and Coulson label “setting” and what Brandt labels “the pheno-world.” These are suggestive notions, and provide a hypothesis against which to investigate the precise nature of the interface between cognitive discourse processes and the communicative environment. What is important, and abundantly clear in this volume, is that context is both a cognitive and an interactively produced construct. That is to say, constraints on conceptual integration come from context only in the sense that discourse processors (i.e. both speakers and hearers) have shared knowledge of contexts, including the co-constructed contexts in short-term memory. Knowledge of context will, for example, include knowledge of typical participants in a medical consultation context, a TV interview context, a speech to a jury in a courtroom, how to read a novel, etc., but, as Hougaard makes clear, the application of knowledge of contexts is guided by action in a sequence of local linguistic contexts constituted in ongoing discourse. Further, from an analytical point of view we need to recognise that context is inseparable from genre, since interactants themselves appear to know how to relate types of text and talk to particular contexts. “Genre” is a way of categorising discourse types and their contexts. For some time it has been clear that genres, understood as text- or utterance-types, are integrated with social structure (Swales 1990). For example, judicial structures are defined intrinsically in terms of participants and the way they interrelate through differentiated but interlinked types of discourse. These kinds of points have been made by discourse analysts without any reference to what is in participants’ minds and the present volume is important in taking us some considerable way towards understanding how at least one kind of mental activity, blending, is related to genres. In this connection there are two related results that arise from the analyses in this volume. One is that genres themselves might be defined through the kinds of conceptual integration that they stimulate in interactants. Oakley and Kaufer propose three different “rhetorical strategies” in a professional health setting (challenge, consultation and presentation), strategies which we can perhaps think of as yielding sub-genres of an over-arching “clinical experience genre”, which in turn can be characterised as a “healing” narrative, a deep-seated schema that may be found in many cultures. The other result that pervades the analyses is the crucial role of speech (prototypically conversational dialogue) in direct form or varieties of reporting. I want to say a little more about this. Perhaps the most fundamental empirical issue concerns the psychological and neural substrate of blending operations – an issue outlined clearly in Hougaard’s paper. Hougaard argues that compression integration can be observed objec-
TSL[v.20020404] Prn:12/12/2007; 12:23
F: PB17010.tex / p.3 (253)
Chapter 10. Reflections on blends and discourse
tively by ethnomethodological techniques. Most of his paper presents evidence for recognising what he calls “packing up” turns, in which a speaker “formulates” (to use the term current in CA) the preceding gist of a conversation and may simultaneously use that turn as a preclosure of the conversation. Such conversational moves are found in interviews and particularly in radio-phone-in dialogue. It is surely correct to say that “packing up” turns involve conceptual blending of some kind that participants must engage in. However, there remain at least three serious questions that Hougaard’s insightful analysis throws up. One is the question of the precise relationship between the content of the packing turn and the preceding dialogue. Is it a paraphrase relationship and indeed what kind of relationship is a paraphrase? A similar and equally important question arises in the case of Pascual’s analysis of the summing up phase executed by lawyers in jury trials, where hearers are expected to accept that the speaker’s reporting of another speaker (for example the prosecution’s projection of the defence lawyer’s arguments) is a true rendering. This raises a second question, apparently overlooked in CA analyses. The data analysed can lead one to suspect that the speaker doing the formulating or packing is not simply co-constructing the talk, as conversation analysts are wont to claim, but is actually manipulating closure on their own terms, that is on the basis of their own representation of someone’ else’s words: for example, radio phone-in hosts and others are in possession of a certain controlling discourse power. (Exceptions do arise, as in Hougaard’s example where a caller rejects a packing turn which is subsequently reformulated by the radio show host.) Williams’s meticulous analysis of teacher-child discourse also shows how one speaker can (here for entirely benign purposes) control the discourse in order to produce blends in recipients. The celebrated Kant blend too is open to this kind of analysis – the professor’s particular representation of Kant is under his control. Even the ethnomethodologist’s claim that participants indicate assent to a summative compression in the packing turn should not be take at face value, as indicating a participant’s acceptance or their having the same conception in mind or as “constructing memory.” These considerations are not, of course wholly a cognitive matter, but the point does show that compressions of specific kinds in specific kinds of discourse genres will not be able to be accounted for in a straightforward manner. The third point is, however, of direct concern to any serious cognitive theory of discourse of the kind CIT claims to be. Hougaard is sceptical of the relevance of neuro-imaging techniques, and while in my view the potential of such methods should not be discounted in advance, Hougaard is certainly right to demand a more realistic psychological explanation for blending theory. It does not appear to be the case that blending theorists have any clear hypotheses about what the brain might be doing when it is setting up mental spaces or blending such spaces, despite references in the literature to neural co-activation and neural binding. In the absence of detailed neuro-science findings, Hougaard turns to psychological models
TSL[v.20020404] Prn:12/12/2007; 12:23
F: PB17010.tex / p.4 (254)
Paul Chilton
of memory, in particular the long-term-working memory model of Kintsch et al. (1999) – which is also found wanting because of its failure to represent interactive cognition of the kind that must account for conversational dialogue building, in which interactants have got to be reciprocally aware of expectations and interpretations and in which the interaction itself provides input to the individual brain engaged in discourse processing. However, it may be that a more appropriate place to look for relevant psychological theory is the psychology of joint attention. Baron-Cohen (2001) and associates have developed the notion of “theory of mind” to deal precisely with the problem of understanding what normal human interaction, including most crucially what we call discourse, involves. Theory of mind is postulated as a mental module that enables normal individuals to recognise that other people also have minds, have intentions and have their own representations of reality – their own systems of mental spaces, their own viewpoints and construals. It is thought that people on the autistic spectrum do not have a normal theory of mind. All this may seem a long way from blending compression, but it is close to the concerns of mental space theory, for it points precisely to the intrinsic ability of the normal human brain to set up mental spaces and keep them distinct while maintaining or constructing appropriate cross-space mappings. It is worth perhaps noting that while CIT has rightly pointed to the largely ignored activity of blending, blending activity does not occur indiscriminately or freely in the normal human brain. What would it be like to blend spaces that were not “meant” to be blended? One may speculate that unconstrained blending underlies pathological brain states, conceivably, for example, schizophrenia. One of the recurrent concerns of this volume centres on what in other literatures is called “metarepresentation” – a speaker’s representation of the states of mind or thoughts of another person. Cosmides and Tooby (2000) present a model of human representation that in some ways is complementary to the notion of blending – arguing that human cognition, under evolutionary pressure, developed a recursive system that kept apart what are in effect what CIT would call mental spaces. The human cognitive architecture has the survival advantage that it does not simply allow itself to be flooded indiscriminately with information, but rather tags it for source and reliability – it assigns an epistemic value to information embedded within its reality space. CIT has the merit of explaining further how counterpart mappings must and do operate. In this volume there are several case studies in different domains that reflect the human cognitive problem of constructing epistemically high-valued representation – i.e. true representations from the point of view of some cognisor. In fact this concern – one might think of it as a fundamental search for truth arising in evolution from the survival value of accurate information about one’s environment – is built into the fundamental architecture of discourse.
TSL[v.20020404] Prn:12/12/2007; 12:23
F: PB17010.tex / p.5 (255)
Chapter 10. Reflections on blends and discourse
This “search for truth” is reflected in several papers and indicates the importance of understanding the base or reality space in relation to spaces embedded within it. This in turn points to the importance of understanding how space builders, and more especially the absence of linguistic space builders, operate as a function of context. It appears to me from several of the papers that the representation of the speech of other participants is especially crucial. This is demonstrated by the perceptive analysis of Dancygier for literary narratives, but this particular kind of blend is arguably even more important in the non-literary contexts, for example that of the courtroom, an institution surely crucial to modern social organisation. In the courtroom and many other social contexts, it is of great interest to know the exact nature of the relationship between the speaker’s reality space and any embedded space attributed to another speaker. What these papers are looking at is, in terms of the western rhetorical tradition, prosopopoeia, classed appropriately enough by Cicero, Quintilian and other rhetors as a “figure of thought”. CIT of course tells us much more about how this figure works cognitively but still I think leaves scope for more research in relation to discourse. In particular, not only are we confronted with the problem of paraphrase relations but also with speakers’ attitudes towards embedded speech. This relation is referred to in the papers as “viewpoint”. However, it seems to me that the detailed analyses found here require us to differentiate between viewpoint and speaker interpretation. Viewpoint is essentially a deictic spatial notion but extended to time and to epistemic evaluation by the self or speaker and is fundamental to all discourse spaces, including embedded discourse spaces (cf. Chilton 2005). This understanding of viewpoint, in which a speaker “distances” him or herself spatially, temporally or epistemically, might need to be separated from “viewpoint” in the sense of interpretation, e.g. of the facts in a court case, of a narration in a police enquiry (whether fictive as in the Blind Assassin studied by Dancygier or in social world). To characterize a narrative of events as “telling” someone’s guilt or innocence, for example, is an interpretation of those events and only metaphorically a “point of view”. In general it is important to differentiate between mental space approaches that model communication involving the separation of spaces, a separation needed for good cognitive reasons, and blending, a creative cognitive operation that occurs under very specific conditions. The contributors to this volume have set us well on the way toward a more precise exploration of what these conditions might be.
TSL[v.20020404] Prn:12/12/2007; 12:23
F: PB17010.tex / p.6 (256)
Paul Chilton
References Baron-Cohen, S. 2001. Mindblindness, Cambridge MA, MIT Press. Brandt, L. and Brandt, P.A. 2005. “Making Sense of a Blend.” In Annual Review of Cognitive Linguistics 5. R. Mendoza Ibanez & F. Jose (eds.), 216–249. Amstredam: John Benjamins Press. Chilton, P. 2005. “Vectors, Viewpoint and Viewpoint Shift.” In Annual Review of Cognitive Linguistics 3: 78–116. Amsterdam: John Benjamins Press. Cosmides, L. and John T. 2000. “Consider the Source: the Evolution of Adaptations for Decoupling and Metarepresentations.” In, Metareprestentation: A Multidisciplinary Perspective, D. Sperber (ed.), pp. 53–117. Oxford: Oxford University Press. Duranti, A. and Goodwin, C. 1993. Rethinking Context. Cambridge: Cambridge University Press. Langacker, R. 1987. Foundations of Cognitive Grammar, Volume 1: Theoretical Perspectives, Stanford: Stanford University Press. Swales, J.M. 1990. Genre Analysis. Cambridge University Press, Cambridge.
TSL[v.20020404] Prn:23/01/2008; 10:47
F: PB170AI.tex / p.1 (257)
Author index
A Abelson, Robert Atwood, Margaret , Austin, J. L. B Baddeley, Aaron D. Bakhtin, M. M. Bal, Mieke Baron-Cohen, Simon Barsalou, Lawrence , Bartlett, Fredric Barwise, Jon Bates, Elizabeth Bawarshi, Anis Benveniste, Emile , Bitzer, Lloyd Booth, Wayne Brandt, Line , , , , , , , , , , , , , –, , , , , , , , , , Brandt, Per Aage , , , , –, , , , , , , Butler, Judith C Chafe, Wallace , , , –, , , Chatman, Seymour , , Chilton, Paul , , , Chomsky, Noam , Cicourel, Aaron , Cienki, Alan , , , , , , – Clark, Herbert , Clarke, Richard , , , , – Clinton, Bill , Cosmides, Leda Cotterill, Janet
Coulson, Seana , , , , , –, , , , , –, , , , , , , , , , , Coulter, John Cutrer, Michelle ,
D Damasio, Antonio Dancygier, Barbara , , , –, , , , , , , , , , , , , Davies, Dave , , –, , – Descartes, Rene , Devitt, Amy Dick, Frederic Dijk, Teun van Driscoll, Marcy P. Du Bois, Jack Duncan, Susan Duranti, Alessandro
E Edwards, Derek , Emmott, Catherine , ,
F Fauconnier, Gilles –, –, , , –, , , , , , , –, , , –, , , , , , , , , , , , , , , , , , , , , , , Fillmore, Charles , , Fitzgerald, F. Scott Fodor, Jerold Fornel, Michel de
G Gallagher, Shawn Gallese, Vittorio Garfinkel, Harold , Garrod, Simon Genette, Gerard , Gibbs, Raymond W. , Glenberg, Arthur Goffman, Erving , Goodwin, Charles , , , , , , , , Goodwin, Marjorie Harness Gross, Terry H Harder, Peter Harré, Romano Herman, David , , , Hopper, Robert Hougaard, Anders –, , , –, , , –, , , , , –, , Hougaard, Gitte R. , , , , , , – Hutchins, Edwin , , , , J Jefferson, Gail , Jensen de Lopez, Kristine Johnson, Mark , , , , , , , , Johnson-Laird, Philip K Kaufer, David S. , , , , –, , , Kintsch, Walter , , L Labov,William ,
TSL[v.20020404] Prn:23/01/2008; 10:47
F: PB170AI.tex / p.2 (258)
Author index Lakoff, George , , , , , , , , , Langacker, Ronald W. , , , –, , Laporte, Leo , , –, M Müller, Cornelia , Mackenzie, Donald Mandler, Jean Matsumoto, Yo Maynard, Douglas McEwen, Ian McNeill, David Merleau-Ponty, Maurice Meyer, Charles F. Molder, Hedwig te , N Narayanan, Srini Neisser, Ulric Nelson, Katherine O Oakley, Todd –, , , , , , , , –, , , , , , , , , , –, , , , , , , , , P Panther, Klaus Parrill, Fey Pascual, Esther , , , , , , , , –, , , ,
, , , , , , –, –, , , , , , , Perlman, Marcus Pickering, Andrew , Pickering, Martin , Polkinghorne, Donald Posner, Richard , –, , , Q Quintilian , R Rimmon-Kenan, Shlomith Robertson, David Rohrer, Tim , , , S Sacks, Harvey , , Safire, William Sapir, Edward Schank, Roger Schegloff, Emanual Schnepf, Jennifer , , Schutz, Alfred , Sedgwick, Eve Sinha, Chris , Skinner, B. F. Sperber, Dan Spinoza, Baruch Stygall, Gail Swales, John , Sweetser, Eve , , , , , , , , , , , , , , ,
T Talmy, Leonard , , , Tannen, Deborah Thornburg, Linda Tobin, Vera Tomasello, Michael Tooby, John Turner, Mark –, –, , , , , , , , , , , , , , , –, , , , , , , , , , , , , , , , , , , , ,
U Uttal, William
V Vandelanotte, Lieven Verhagen, Arie
W Waletsky, Joshua Walter, Bettyruth , Wiemar-Hastings, Katja Williams, Robert F. , , , , –, , , , , , , , , –, Wilson, Deirdre ,
Z Zlatev, Jordan
TSL[v.20020404] Prn:25/01/2008; 8:11
F: PB170SI.tex / p.1 (259)
Subject index
A access principle addressee(s) , , , , , , , , – American English American judiciary analogy (see blends) anaphora , anchor , , , , , , , –, , , argumentational relevance –, –, artifact layer –, , assimilates v. accommodates attorneys , , , –, , –, , autistic spectrum B backstage cognition base space (see mental space(s)) behavioral cognitive account behaviorism , behaviorist , , blend(s) , , –, –, –, –, –, –, –, –, , , –, –, –, , –, –, – analogy blends , – anchored blends blended space –, , , , , , , , , , , , –, –, , , , , , –, Clock Quarters blend , –, , , , Clock Quarters Motion blend , –
contrast blends , –, contrastive co-temporality blends – Debate with Kant , , , , , , , , , , , , , , Dream of the Rood , fictive interaction blends , , , , , , semiotic blend virtual being blends , – Voting as Speaking , , , , blended space (see blend(s)) C challenge strategy , , chunkings Clinical Experience –, , , , , clock-reading (see time-telling) , Cognitive Linguistics , , , , , , , , , cognitive phenomena cognitive process cognitive science , , –, , , , , , , , common ground –, , , common sense , , , commonality –, compression , –, , –, , –, , –, , , , , , conceptual blend (see also conceptual blending) , , , , , , , , conceptual blending (see also blends and conceptual blend)
, , , , , , , , , , , , , conceptual integration (see conceptual blending) , , –, , , , –, , , , , , , –, conceptual integration networks , , , , , , Conceptual Integration Theory (CIT) , , , , , , , , – conceptual mapping –, conditionals , , Connect the Dots , , , , –, , , – Connecting the Dots consultation strategy , , , container schema , context , , –, –, , –, , , –, , , , , , –, context renewing context-shaped Conversational Analysis (CA) –, –, , –, –, –, counterargument , , cross-input projection , cross-space mappings , , , , cultural knowledge , , , current hour
D data mining , Democrats , disanalogy , , , ,
TSL[v.20020404] Prn:25/01/2008; 8:11
F: PB170SI.tex / p.2 (260)
Subject index discourse –, –, –, –, , , , –, –, –, –, , –, –, –, –, –, –, , , , , , , –, – discourse participants , , , , , , , , , Discursive Psychology (DP) , , – Docuscope , , , , , Dramatization , , , , , , E embedded enunciation , , , , embedded metonymic enunciation , , , , embodiment , , , emergent story , , , –, , , , emic , empirical research enunciation –, –, –, –, –, , –, enunciational shift episodes , , epistemic stance , ethnomethodological techniques ethnomethodology , , , , , event related potentials (ERP) , evidence , , , –, –, , , , , , –, , , –, , , experientialism eye gaze direction F factive , , Factor Analysis (FA) , , , , fictive interaction , –, –, –, –, –, –, –, –, – fictive interaction specifiers
fictive motion , , fictive trialogues First Generation Cognitive Science First Gulf War formulations –, , , Fresh Air , functional Magnetic Resonance Imaging (fMRI) , G Gat System, the generative linguists – generic enunciation , –, , , genre , , , , , –, , –, , genre layer , genre studies – George H. W. Bush , – George W. Bush , – gestalt gesture , , , , –, , , , , , , – gestures , , , –, , , , , , , –, Giant Sixty Eight KNBR, the goal (see source-path-goal schema) , , , , , , , , , , , –, grammar layer –, , , , , , guided conceptualization , , H hearer(s) , , , , , , , , , high-valued representation Hospital Practice , , human cognitive architecture hypothesis of blending , I ideal speaker-hearer , illocutional relevance , image schema
image schemas imagistic anchors individual mind , , innovation , –, , instruction , , , instructional discourse , –, , interactional memory chunk , interactional pattern , interactional working memory interlocutors , , , , , –, –, –, , –, interpretation , , , , , , , , , , , –, –, , –, , , , , – intersubjective meaning intonation unit , , , , fragmentary intonation unit regulatory intonation unit substantive intonation unit IT-constructions , J joint attention , jurors , , , –, – K KNOWING IS SEEING metaphor L language action type (LAT) , , legal monologues , , linguistic knowledge , , , , , , Long Term Working Memory model (LTWM) , M Managed Care manipulation of objects
TSL[v.20020404] Prn:25/01/2008; 8:11
F: PB170SI.tex / p.3 (261)
Subject index mappings , – , , , , , –, , , , , –, –, material anchors , , , , meaning –, , –, –, –, , , , , , , , , , , , , –, , , –, , , , –, , , , , , , , , , , , –, , – meaning construction , , –, , –, –, , , –, , , , mental simulation , –, mental space(s) , –, , , –, –, , , , , , , –, –, , , –, –, , –, , –, –, –, –, –, –, –, – base space , , –, , , , , , , , , , counterfactual space displaced grounding focus spaces , future space , grounding space HERE-AND-NOW Space hypothetical space –, law space meaning space , normative space , , presentation space , , , , , , , , , , –, , , –, , , reality space , , , , –, , , reference space , , , , –, , , , , , , , –, , , –, , semiotic base space , , , Under-Oath space viewpoint spaces , virtual space , , –, , –,
mental space embedding mental spaces and conceptual integration (MSCI) –, –, –, –, – mental spaces theory , , , , , , –, , metaphor –, –, , –, , , , , , , metaphoric language metarepresentation metonymy , , , EFFECT FOR CAUSE Michigan Corpus of Spoken Academic English (MICASE) , motion-based conceptualizations N narrative anchors , –, –, narrative discourse , , , narrative viewpoint , –, – naturally occurring behavior Neural Theory of Language (NTL) neuro-imaging techniques novelty – O ontological differentiation of mental spaces ontological grounding optimality principles P packing up turns , , , path of motion , , , , , pathological brain states personal enunciation , , , phenomenology , , pheno-world phonology pointing gestures pragmatic implication , , , –,
preclosings predication , , , presentation strategy , , , –, producer , profiling , , , , , , , , , , , prosopopoeia Q queuing – R reference hour Relevance Theory representation , , , , , –, , –, –, – researcher , , –, responding to turns rhetorical situation , , rhetorical strategies , , , role-value mappings Russian Formalism S schemas , , , , schizophrenia screaming dots , , – scripts , selective projection , , semantics , –, , –, , –, , –, , –, –, semiosis , , , , semiotic integration sense , , , , separation of spaces sequence of events , sequence of expressions setting , , , , , , –, –, , , , shared processes signification , Situation Semantics situational knowledge –, , , , situational relevance , , ,
TSL[v.20020404] Prn:25/01/2008; 8:11
F: PB170SI.tex / p.4 (262)
Subject index
source (see source-path-goal schema) –, , , , –, –, , , , , –, –, , source-path-goal schema , , Spanish (judiciary) system , speaker , –, , , , , , –, –, –, –, , , , , –, –, –, –, –, , – speaker interpretation spontaneous manual gesture and speech standing in line (see queuing) , static conceptualization story , , , , –, –, –, , –, , story construction –, , subjective change syntax , ,
T talk-in-interaction , , talkbank
teaching , , , –, –, , , text , , , –, –, –, –, , , , –, –, , , , , theatricality theory of mind THERE-constructions , thinking as speaking time-telling , , , , , time-telling instruction tracing gestures trajectory (see source-path-goal schema) , truth –, , , , , , , , turn , , , , , , , , , , , –, , , , –, , , , –, –, , –, –, , , –, – turn packing utterance (see packing up turns) , turn-constructional unit (TCU) –, , turn-transition ,
U understanding , –, –, , –, , , –, , , –, –, , , , , –, , usage based approach utterance(s) , , , , , , , , , , , , –, –, –, , , , , , , , , , –, , –, , , , , , –, , , , , –, V video-recorded conversations viewpoint , , , –, –, , , , , , , , , , , , , virtual reality , vital relations , , , , W Wallace and Associates –, , white coat effect , white coat hypertension ,
Pragmatics & Beyond New Series A complete list of titles in this series can be found on the publishers’ website, www.benjamins.com 176 Jucker, Andreas H. and Irma Taavitsainen (eds.): Speech Acts in the History of English. xxxi, 286 pp. + index. Expected April 2008 175 Gómez González, María de los Ángeles, J. Lachlan Mackenzie and Elsa M. GonzálezÁlvarez (eds.): Languages and Cultures in Contrast and Comparison. xxii, 354 pp. + index. Expected April 2008 174 Heyd, Theresa: Email Hoaxes. Form, function, genre ecology. vii, 245 pp. + index. Expected April 2008 173 Zanotto, Mara Sofia, Lynne Cameron and Marilda C. Cavalcanti (eds.): Confronting Metaphor in Use. An applied linguistic approach. vii, 310 pp. + index. Expected March 2008 172 Benz, Anton and Peter Kühnlein (eds.): Constraints in Discourse. xi, 278 pp. + index. Expected March 2008 171 Félix-Brasdefer, J. César: Politeness in Mexico and the United States. A contrastive study of the realization and perception of refusals. 2008. xiv, 195 pp. 170 Oakley, Todd and Anders Hougaard (eds.): Mental Spaces in Discourse and Interaction. 2008. vi, 262 pp. 169 Connor, Ulla, Ed Nagelhout and William Rozycki (eds.): Contrastive Rhetoric. Reaching to intercultural rhetoric. 2008. viii, 324 pp. 168 Proost, Kristel: Conceptual Structure in Lexical Items. The lexicalisation of communication concepts in English, German and Dutch. 2007. xii, 304 pp. 167 Bousfield, Derek: Impoliteness in Interaction. 2008. xiii, 281 pp. 166 Nakane, Ikuko: Silence in Intercultural Communication. Perceptions and performance. 2007. xii, 240 pp. 165 Bublitz, Wolfram and Axel Hübler (eds.): Metapragmatics in Use. 2007. viii, 301 pp. 164 Englebretson, Robert (ed.): Stancetaking in Discourse. Subjectivity, evaluation, interaction. 2007. viii, 323 pp. 163 Lytra, Vally: Play Frames and Social Identities. Contact encounters in a Greek primary school. 2007. xii, 300 pp. 162 Fetzer, Anita (ed.): Context and Appropriateness. Micro meets macro. 2007. vi, 265 pp. 161 Celle, Agnès and Ruth Huart (eds.): Connectives as Discourse Landmarks. 2007. viii, 212 pp. 160 Fetzer, Anita and Gerda Eva Lauerbach (eds.): Political Discourse in the Media. Cross-cultural perspectives. 2007. viii, 379 pp. 159 Maynard, Senko K.: Linguistic Creativity in Japanese Discourse. Exploring the multiplicity of self, perspective, and voice. 2007. xvi, 356 pp. 158 Walker, Terry: Thou and You in Early Modern English Dialogues. Trials, Depositions, and Drama Comedy. 2007. xx, 339 pp. 157 Crawford Camiciottoli, Belinda: The Language of Business Studies Lectures. A corpus-assisted analysis. 2007. xvi, 236 pp. 156 Vega Moreno, Rosa E.: Creativity and Convention. The pragmatics of everyday figurative speech. 2007. xii, 249 pp. 155 Hedberg, Nancy and Ron Zacharski (eds.): The Grammar–Pragmatics Interface. Essays in honor of Jeanette K. Gundel. 2007. viii, 345 pp. 154 Hübler, Axel: The Nonverbal Shift in Early Modern English Conversation. 2007. x, 281 pp. 153 Arnovick, Leslie K.: Written Reliquaries. The resonance of orality in medieval English texts. 2006. xii, 292 pp. 152 Warren, Martin: Features of Naturalness in Conversation. 2006. x, 272 pp. 151 Suzuki, Satoko (ed.): Emotive Communication in Japanese. 2006. x, 234 pp. 150 Busse, Beatrix: Vocative Constructions in the Language of Shakespeare. 2006. xviii, 525 pp. 149 Locher, Miriam A.: Advice Online. Advice-giving in an American Internet health column. 2006. xvi, 277 pp. 148 Fløttum, Kjersti, Trine Dahl and Torodd Kinn: Academic Voices. Across languages and disciplines. 2006. x, 309 pp. 147 Hinrichs, Lars: Codeswitching on the Web. English and Jamaican Creole in e-mail communication. 2006. x, 302 pp.
146 Tanskanen, Sanna-Kaisa: Collaborating towards Coherence. Lexical cohesion in English discourse. 2006. ix, 192 pp. 145 Kurhila, Salla: Second Language Interaction. 2006. vii, 257 pp. 144 Bührig, Kristin and Jan D. ten Thije (eds.): Beyond Misunderstanding. Linguistic analyses of intercultural communication. 2006. vi, 339 pp. 143 Baker, Carolyn, Michael Emmison and Alan Firth (eds.): Calling for Help. Language and social interaction in telephone helplines. 2005. xviii, 352 pp. 142 Sidnell, Jack: Talk and Practical Epistemology. The social life of knowledge in a Caribbean community. 2005. xvi, 255 pp. 141 Zhu, Yunxia: Written Communication across Cultures. A sociocognitive perspective on business genres. 2005. xviii, 216 pp. 140 Butler, Christopher S., María de los Ángeles Gómez González and Susana M. Doval-Suárez (eds.): The Dynamics of Language Use. Functional and contrastive perspectives. 2005. xvi, 413 pp. 139 Lakoff, Robin T. and Sachiko Ide (eds.): Broadening the Horizon of Linguistic Politeness. 2005. xii, 342 pp. 138 Müller, Simone: Discourse Markers in Native and Non-native English Discourse. 2005. xviii, 290 pp. 137 Morita, Emi: Negotiation of Contingent Talk. The Japanese interactional particles ne and sa. 2005. xvi, 240 pp. 136 Sassen, Claudia: Linguistic Dimensions of Crisis Talk. Formalising structures in a controlled language. 2005. ix, 230 pp. 135 Archer, Dawn: Questions and Answers in the English Courtroom (1640–1760). A sociopragmatic analysis. 2005. xiv, 374 pp. 134 Skaffari, Janne, Matti Peikola, Ruth Carroll, Risto Hiltunen and Brita Wårvik (eds.): Opening Windows on Texts and Discourses of the Past. 2005. x, 418 pp. 133 Marnette, Sophie: Speech and Thought Presentation in French. Concepts and strategies. 2005. xiv, 379 pp. 132 Onodera, Noriko O.: Japanese Discourse Markers. Synchronic and diachronic discourse analysis. 2004. xiv, 253 pp. 131 Janoschka, Anja: Web Advertising. New forms of communication on the Internet. 2004. xiv, 230 pp. 130 Halmari, Helena and Tuija Virtanen (eds.): Persuasion Across Genres. A linguistic approach. 2005. x, 257 pp. 129 Taboada, María Teresa: Building Coherence and Cohesion. Task-oriented dialogue in English and Spanish. 2004. xvii, 264 pp. 128 Cordella, Marisa: The Dynamic Consultation. A discourse analytical study of doctor–patient communication. 2004. xvi, 254 pp. 127 Brisard, Frank, Michael Meeuwis and Bart Vandenabeele (eds.): Seduction, Community, Speech. A Festschrift for Herman Parret. 2004. vi, 202 pp. 126 Wu, Yi’an: Spatial Demonstratives in English and Chinese. Text and Cognition. 2004. xviii, 236 pp. 125 Lerner, Gene H. (ed.): Conversation Analysis. Studies from the first generation. 2004. x, 302 pp. 124 Vine, Bernadette: Getting Things Done at Work. The discourse of power in workplace interaction. 2004. x, 278 pp. 123 Márquez Reiter, Rosina and María Elena Placencia (eds.): Current Trends in the Pragmatics of Spanish. 2004. xvi, 383 pp. 122 González, Montserrat: Pragmatic Markers in Oral Narrative. The case of English and Catalan. 2004. xvi, 410 pp. 121 Fetzer, Anita: Recontextualizing Context. Grammaticality meets appropriateness. 2004. x, 272 pp. 120 Aijmer, Karin and Anna-Brita Stenström (eds.): Discourse Patterns in Spoken and Written Corpora. 2004. viii, 279 pp. 119 Hiltunen, Risto and Janne Skaffari (eds.): Discourse Perspectives on English. Medieval to modern. 2003. viii, 243 pp. 118 Cheng, Winnie: Intercultural Conversation. 2003. xii, 279 pp. 117 Wu, Ruey-Jiuan Regina: Stance in Talk. A conversation analysis of Mandarin final particles. 2004. xvi, 260 pp. 116 Grant, Colin B. (ed.): Rethinking Communicative Interaction. New interdisciplinary horizons. 2003. viii, 330 pp.
115 Kärkkäinen, Elise: Epistemic Stance in English Conversation. A description of its interactional functions, with a focus on I think. 2003. xii, 213 pp. 114 Kühnlein, Peter, Hannes Rieser and Henk Zeevat (eds.): Perspectives on Dialogue in the New Millennium. 2003. xii, 400 pp. 113 Panther, Klaus-Uwe and Linda L. Thornburg (eds.): Metonymy and Pragmatic Inferencing. 2003. xii, 285 pp. 112 Lenz, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. 2003. xiv, 279 pp. 111 Ensink, Titus and Christoph Sauer (eds.): Framing and Perspectivising in Discourse. 2003. viii, 227 pp. 110 Androutsopoulos, Jannis K. and Alexandra Georgakopoulou (eds.): Discourse Constructions of Youth Identities. 2003. viii, 343 pp. 109 Mayes, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and America. 2003. xiv, 228 pp. 108 Barron, Anne: Acquisition in Interlanguage Pragmatics. Learning how to do things with words in a study abroad context. 2003. xviii, 403 pp. 107 Taavitsainen, Irma and Andreas H. Jucker (eds.): Diachronic Perspectives on Address Term Systems. 2003. viii, 446 pp. 106 Busse, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person pronouns. 2002. xiv, 344 pp. 105 Blackwell, Sarah: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. xvi, 303 pp. 104 Beeching, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. x, 251 pp. 103 Fetzer, Anita and Christiane Meierkord (eds.): Rethinking Sequentiality. Linguistics meets conversational interaction. 2002. vi, 300 pp. 102 Leafgren, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and objects. 2002. xii, 252 pp. 101 Luke, K. K. and Theodossia-Soula Pavlidou (eds.): Telephone Calls. Unity and diversity in conversational structure across languages and cultures. 2002. x, 295 pp. 100 Jaszczolt, Katarzyna M. and Ken Turner (eds.): Meaning Through Language Contrast. Volume 2. 2003. viii, 496 pp. 99 Jaszczolt, Katarzyna M. and Ken Turner (eds.): Meaning Through Language Contrast. Volume 1. 2003. xii, 388 pp. 98 Duszak, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002. viii, 522 pp. 97 Maynard, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. 2002. xiv, 481 pp. 96 Haverkate, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002. vi, 241 pp. 95 Fitzmaurice, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002. viii, 263 pp. 94 McIlvenny, Paul (ed.): Talking Gender and Sexuality. 2002. x, 332 pp. 93 Baron, Bettina and Helga Kotthoff (eds.): Gender in Interaction. Perspectives on femininity and masculinity in ethnography and discourse. 2002. xxiv, 357 pp. 92 Gardner, Rod: When Listeners Talk. Response tokens and listener stance. 2001. xxii, 281 pp. 91 Gross, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. xxviii, 341 pp. 90 Kenesei, István and Robert M. Harnish (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. xxii, 352 pp. 89 Itakura, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. 2001. xviii, 231 pp. 88 Bayraktaroğlu, Arın and Maria Sifianou (eds.): Linguistic Politeness Across Boundaries. The case of Greek and Turkish. 2001. xiv, 439 pp. 87 Mushin, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. 2001. xviii, 244 pp. 86 Ifantidou, Elly: Evidentials and Relevance. 2001. xii, 225 pp. 85 Collins, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. xx, 384 pp. 84 Andersen, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. ix, 352 pp.
83 Márquez Reiter, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. xviii, 225 pp. 82 Khalil, Esam N.: Grounding in English and Arabic News Discourse. 2000. x, 274 pp. 81 Di Luzio, Aldo, Susanne Günthner and Franca Orletti (eds.): Culture in Communication. Analyses of intercultural situations. 2001. xvi, 341 pp. 80 Ungerer, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. xiv, 286 pp. 79 Andersen, Gisle and Thorstein Fretheim (eds.): Pragmatic Markers and Propositional Attitude. 2000. viii, 273 pp. 78 Sell, Roger D.: Literature as Communication. The foundations of mediating criticism. 2000. xiv, 348 pp. 77 Vanderveken, Daniel and Susumu Kubo (eds.): Essays in Speech Act Theory. 2002. vi, 328 pp. 76 Matsui, Tomoko: Bridging and Relevance. 2000. xii, 251 pp. 75 Pilkington, Adrian: Poetic Effects. A relevance theory perspective. 2000. xiv, 214 pp. 74 Trosborg, Anna (ed.): Analysing Professional Genres. 2000. xvi, 256 pp. 73 Hester, Stephen K. and David Francis (eds.): Local Educational Order. Ethnomethodological studies of knowledge in action. 2000. viii, 326 pp. 72 Marmaridou, Sophia S.A.: Pragmatic Meaning and Cognition. 2000. xii, 322 pp. 71 Gómez González, María de los Ángeles: The Theme–Topic Interface. Evidence from English. 2001. xxiv, 438 pp. 70 Sorjonen, Marja-Leena: Responding in Conversation. A study of response particles in Finnish. 2001. x, 330 pp. 69 Noh, Eun-Ju: Metarepresentation. A relevance-theory approach. 2000. xii, 242 pp. 68 Arnovick, Leslie K.: Diachronic Pragmatics. Seven case studies in English illocutionary development. 2000. xii, 196 pp. 67 Taavitsainen, Irma, Gunnel Melchers and Päivi Pahta (eds.): Writing in Nonstandard English. 2000. viii, 404 pp. 66 Jucker, Andreas H., Gerd Fritz and Franz Lebsanft (eds.): Historical Dialogue Analysis. 1999. viii, 478 pp. 65 Cooren, François: The Organizing Property of Communication. 2000. xvi, 272 pp. 64 Svennevig, Jan: Getting Acquainted in Conversation. A study of initial interactions. 2000. x, 384 pp. 63 Bublitz, Wolfram, Uta Lenk and Eija Ventola (eds.): Coherence in Spoken and Written Discourse. How to create it and how to describe it. Selected papers from the International Workshop on Coherence, Augsburg, 24-27 April 1997. 1999. xiv, 300 pp. 62 Tzanne, Angeliki: Talking at Cross-Purposes. The dynamics of miscommunication. 2000. xiv, 263 pp. 61 Mills, Margaret H. (ed.): Slavic Gender Linguistics. 1999. xviii, 251 pp. 60 Jacobs, Geert: Preformulating the News. An analysis of the metapragmatics of press releases. 1999. xviii, 428 pp. 59 Kamio, Akio and Ken-ichi Takami (eds.): Function and Structure. In honor of Susumu Kuno. 1999. x, 398 pp. 58 Rouchota, Villy and Andreas H. Jucker (eds.): Current Issues in Relevance Theory. 1998. xii, 368 pp. 57 Jucker, Andreas H. and Yael Ziv (eds.): Discourse Markers. Descriptions and theory. 1998. x, 363 pp. 56 Tanaka, Hiroko: Turn-Taking in Japanese Conversation. A Study in Grammar and Interaction. 2000. xiv, 242 pp. 55 Allwood, Jens and Peter Gärdenfors (eds.): Cognitive Semantics. Meaning and cognition. 1999. x, 201 pp. 54 Hyland, Ken: Hedging in Scientific Research Articles. 1998. x, 308 pp. 53 Mosegaard Hansen, Maj-Britt: The Function of Discourse Particles. A study with special reference to spoken standard French. 1998. xii, 418 pp. 52 Gillis, Steven and Annick De Houwer (eds.): The Acquisition of Dutch. With a Preface by Catherine E. Snow. 1998. xvi, 444 pp. 51 Boulima, Jamila: Negotiated Interaction in Target Language Classroom Discourse. 1999. xiv, 338 pp. 50 Grenoble, Lenore A.: Deixis and Information Packaging in Russian Discourse. 1998. xviii, 338 pp. 49 Kurzon, Dennis: Discourse of Silence. 1998. vi, 162 pp. 48 Kamio, Akio: Territory of Information. 1997. xiv, 227 pp.