Papers in theoretical linguistics (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

PAPERS IN THEORETICAL LINGUISTICS

AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE General Editor E.F. KONRAD KOERNER (University of Ottawa)

Series IV - CURRENT ISSUES IN LINGUISTIC THEORY

Advisory Editorial Board Henning Andersen (Los Angeles); Raimo Anttila (Los Angeles) Thomas V. Gamkrelidze (Tbilisi); John E. Joseph (College Park, Md.) Hans-Heinrich Lieb (Berlin); Ernst Pulgram (Ann Arbor, Mich.) E. Wyn Roberts (Vancouver, B.C.); Danny Steinberg (Tokyo)

Volume 23

Niels Danielsen Papers in Theoretical Linguistics

NIELS DANIELSEN

PAPERS IN THEORETICAL LINGUISTICS Edited by

PER BÆRENTZEN University of Aarhus

JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA 1992

Library of Congress Cataloging-in-Publication Data Danielsen, Niels. Papers in theoretical linguistics / Niels Danielsen ; edited by Per Baerentzen. p. cm. -- (Amsterdam studies in the theory and history of linguistic science. Series IV, Current issues in linguistic theory, ISSN 0304-0763; v. 23) Includes bibliographical references and index. 1. Linguistics. I. Baerentzen, Per. II. Series. P125.D37 1992 410--dc20 91-39469 ISBN 90 272 3509 0 (alk. paper) CIP © Copyright 1992 - 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.

CONTENTS Foreword

vii

Niels Danielsen: List of Publications

xi

Niels Danielsen in memoriam (Laurits Rendboe)

xv

Papers by Niels Danielsen: On the Universality of Language

1

Die Relativa im Neuhochdeutschen — und anderen Sprachen. Eine Vorstudie zu einer konstitutionellen Sprachtheorie

37

On the Nucleus Constructions in Human Language. Preamble to a Semasiosyntax

61

A First Constitutional Step Towards a Universal Syntax

97

The Cases Unravelled (Or: Back to Linguistics!)

105

A Final Wrest: On Existence, Existential Disseminations and their Delegations

141

Plädoyer gegen die generativen Tiefenoperationen. Kritik einer Scheinlehre

171

Kinetic Analysis of Sentences

193

A Concise List of the most Important Symbols and Abbreviations used in a Nomic Analysis

217

Index of Languages Index of Authors

219 223

FOREWORD This book has taken a long time to produce. More than ten years ago the late Niels Danielsen made plans to publish a selection from his linguistic writings in a volume tentatively entitled Papers in Linguistic Theory. But because of the progression of his illness he was forced to abandon its realization. Following the death of Niels Danielsen in Oc­ tober 1987 I was asked by the series editor to take charge of the edi­ ting of this volume. I gladly accepted the task. It is my hope that this book will help to communicate the linguistic theories of Niels Danielsen to a wider scholarly audience. The articles reprinted here have been chosen with a view to throw a light on the essential components of Niels Danielsen's linguistic theo­ ry whose key word is 'linguistic universality'. With one exception the articles are reprinted in their original form without any alterations or corrections. Niels Danielsen frequently made reference to his own wri­ tings but rarely was very accurate in this, and so the reader is referred to the complete list of his linguistic publications here appended for exact information. The only paper not reprinted in its original form is the one entitled "On the Universality of Language". It has been thought preferable to provide an English translation of Danielsen's "Zur Universalität der Sprache" (1976a) because it serves as a fine introduction to his whole work. It presents Danielsen's entire theory in nuce by mentioning the most important universals of his, such as Polarity, the Constitutional Axis, the Nuclei, the Nomic Structure of the Sentence, Transversal Re­ lations, Interference (Case) Theory, the Critical Field of Distribution, and the Formal Garment of Sentences. The paper "Die Relativa im Neuhochdeutschen — und anderen Spra­ chen: Eine Vorstudie zu einer konstitutionellen Sprachtheorie" (Daniel­ sen 1975) deals with the Constitutional Axis, i.e., the fundamental se­ mantic entities of human language. The constitutional axis of a given language is made up by the entire set of specific pronominal, verbal and adverbial interrogatives and exclamatives of that language. Niels Danielsen in great detail shows us the constitutional axis of modern German and then in an addendum offers a survey of the constitutional axis of French and Latin, with reference being made to the constitutio­ nal axis of other languages.

Viii

FOREWORD

The Universal of the Verbal Nuclei grows out of the constitutional axis. What Niels Danielsen sees as verbal nuclei are in fact the wellknown auxiliary verbs and some further, very abstract verbs. The chap­ ter "On the Nucleus Constructions in Human Language: Preamble to a semasiosyntax" has been taken from his book Linguistic Studies (Da­ nielsen 1980:24-58). In his paper "A First Constitutional Step towards a Universal Syn­ tax" (1977b) Niels Danielsen deals with the universal which he calls the Critical Field of Distribution. It combines syntactical valency and topo­ logy and, to me at least, makes up the most complex and difficult part of his theory. The two papers "The Cases Unravelled (or: Back to linguistics!)" (1983a) and "A Final Wrest: On existence, existential disseminations and their delegations" (1985a) are among the last of Danielsen's publi­ cations and should be seen complementary as both of them discuss Ca­ se Theory, Transversal Relations in syntax and Verbal Nuclei. For many years Niels Danielsen was campaigning against generative transformational grammar. The paper "Plädoyer gegen die generativen Tiefenoperationen" (1973a) is a clear example of this. Here Danielsen demonstrates how the generative transformational procedure, when ap­ plied to languages somewhat different from English and German, will inevitably lead to absurdities. Transformations like deletion and permu­ tation are seen as entirely arbitrary operations that make it only too easy for the linguist to transform any unacceptable sentence generated by the grammar into a streamlined, i.e., a so-called 'well-formed' sen­ tence. Niels Danielsen's own theory, however, is not free from comparable weaknesses. This can be seen in the chapter "Kinetic Analysis of Sen­ tences" taken from his monograph An Essay on Nomos and Human Language (1976c). The various syntactic and semantic levels and the way of transition between levels show clear correspondences with the rules of generative transformational grammar so eagerly fought by him. In spite of the technical and graphic formalism demonstrated by Niels Danielsen in his writings, his linguistic theory first and foremost is a semantic theory. Thanks to his acquaintance with so many langua­ ges Danielsen is able to make comparisons between them very freely but, in my opinion, he at times overemphasizes mere semantic paral­ lelism between languages and neglects the formal grammatical aspects. It is my sincere hope, however, that this insistance on semantic charac­ teristics of language represented by the present selection of papers will

FOREWORD

ix

do Niels Danielsen's linguistic theory full justice and at the same time confer inspiration on others. The papers here united are preceded by an obituary written by Laurits Rendboe for the volume 13 of NOWELE (1989) dedicated to the me­ mory of Niels Danielsen. I wish to thank Laurits Rendboe and the edi­ tor for their permission to include this fine appraisal of Niels Danielsen as a man and a scholar. The editors and publishers responsible for the original publication of the articles included in the present book have shown the greatest kind­ ness by giving their consent to the reprinting of the texts. I want to express my thanks to the following persons and companies: Professor Dr. Rudolf Schützeichel, Münster, and the Carl Winter Universitätsver­ lag, Heidelberg, for "On the Universality of Language" (1976a) and "On the Nucleus Constructions in Human Language. Preamble to a semasiosyntax" (1980b); Professor Dr. Werner Neumann, Berlin, and the Akademie-Verlag Berlin, for "Die Relativa im Neuhochdeutschen — und anderen Sprachen: Eine Vorstudie zu einer konstitutionellen Sprach­ theorie" (1975) and "A First Constitutional Step towards a Universal Syntax" (1977b); Professor Hans F. Nielsen, Odense, and Odense Uni­ versity Press, for "The Cases Unravelled (or: Back to linguistics!)" (1983a) and "A Final Wrest: On existence, existential disseminations and their delegations" (1985a); Professor Dr. Herbert Kolb, München, and the Erich Schmidt Verlag, Berlin, for "Plädoyer gegen die generati­ ven Tiefenoperationen: Kritik einer Scheinlehre" (1973a), and The Royal Danish Academy of Sciences and Letters, Copenhagen, for "Kinetic Ana­ lysis of Sentences" (from Danielsen 1976c). Aarhus, Denmark July 1991

Per Bærentzen

NIELS DANIELSEN: LIST OF PUBLICATIONS 1963. "Versuch einer neuen Deutung des westgermanischen Wortes für 'Welt'". Språkliga Bidrag 4, Nr. 18.101-106. Lund. 1967a. "Ahd. skepfen, heffen und swerien". Studia Neophilologica 39.281283. Uppsala. 1967b. "Zur definierten Aussage". Oriens 20.169-196. Leiden. 1968a. Zum Wesen des Konditionalsatzes, nicht zuletzt im Indoeuro­ päischen. Odense: Odense University Press, 75 pp. 1968b. Status und Polarität im Gotischen, im Lichte des Kymrischen dargestellt. Odense: Odense University Press, 154 pp. 1968c. "Die negativen unbestimmten Pronominaladjektiva im Alt- und Mittelhochdeutschen". Zeitschrift für deutsche Sprache 24.92-117. Berlin. 1971a. "Das generative Abenteuer". Språkliga Bidrag 6, Nr. 26.105146. Lund. 1971b. "Die Relativa im Neuhochdeutschen — und anderswo: Eine Vor­ studie zu einer konstitutionellen Sprachtheorie". Spràkliga Bidrag 6, Nr. 27.83-116. Lund. 1972a. Die Frage: Eine sprachwissenschaftliche Untersuchung. Copen­ hagen: Munksgaard, 68 pp. 1972b. "Das generative Abenteuer". Zeitschrift für Phonetik, Sprach­ 25.255-280. Berlin. (Rev. wissenschaft und Kommunikationsforschung version of 1971a.) 1972c. Review of Ernesto Zierer, The Theory of Graphs in Linguistics (The Hague: Mouton, 1970). Ural-Altaische Jahrbücher 44.314-315. Wiesbaden. 1972d. Review of Ernst Pulgram, Syllable, Word, Nexus, Cursus (The Hague: Mouton, 1970). Ural-Altaische Jahrbücher 44.315-317. Wies­ baden. 1973a. "Plädoyer gegen die generativen Tiefenoperationen: Kritik einer Scheinlehre". Archiv für das Studium der neueren Sprachen und Lite­ raturen 210.241-262. Braunschweig. (Repr. in the present volume.) 1973b. "Apologie der Oberflächenstruktur". Zeitschrift für Phonetik, 26.11-21. Berlin. Sprachwissenschaft und Kommunikationsforschung 1974a. "Semasiosyntaktische Universalien im Finnischen". UralAltaische Jahrbücher 46.54-83. Wiesbaden.

Xii

LIST OF PUBLICATIONS

1974b. "Fokus pâ syntaksen [Focus on Syntax]". Meddelelser fra Gymnasieskolernes Tysklærerforening 55.24-85. Copenhagen. 1975. "Die Relativa im Neuhochdeutschen — und anderen Sprachen: Eine Vorstudie zu einer konstitutionellen Sprachtheorie". Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 28.5073. Berlin. (Rev. version of 1971b; repr. in the present volume.) 1976a. "Zur Universalität der Sprache". Sprachwissenschaft 1.1-45. Heidelberg. (English translation in the present volume.) 1976b. "Das Satzverbal () und die Kasus". Sprachwissenschaft 1.262291. Heidelberg. 1976c. An Essay on Nomos and Human Language. Copenhagen: Munksgaard, 160 pp. (The third chapter "Kinetic Analysis of Sentences" reprinted in the present volume.) 1976d. "Bezdna bez glubiny bez dna [A bottomless pit without depth]". Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 29.388-395. Berlin. 1976e. "Tod vor dem Sterben: Gedanken über sprachpolitische Durch­ schnittlichkeit und Minoritätentod". Sprachen und Staaten: Festschrift Heinz Kloss, Teil I: Der politische und soziale Status der Sprachen in den Staaten der Europäischen Gemeinschaft ed. by Harald Haar­ mann & Anna-Liisa Värri Haarmann, 315-322. Hamburg: Stiftung Europa-Kolleg. 1977a. "Digression über Delimitationen". Sprachwissenschaft 2.368-410. Heidelberg. 1977b. "A First Constitutional Step towards a Universal Syntax". Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 30.354-361. Berlin. (Repr. in the present volume.) 1977c. "Zu den Nucleuskonstruktionen in der menschlichen Sprache: Vorbemerkungen zu einer Semasiosyntax". Zeitschrift für Phonetik, 30.589-608. Berlin. Sprachwissenschaft und Kommunikationsforschung 1977d. "Richieste di chiarimenti epistemologici [The need for epistomological clarification]". Problemi della ricostruzione in linguistica: Atti del convegno internazionale di studi, Pauia, 1-2 ottobre 1975 ed. by Raffaele Simone & Ugo Vignuzzi, 133-142. Rome: Bulzoni. 1978. "Zum Satzelektron". Sprachwissenschaft 3.184-224. Heidelberg. 1979a. "The Problem of the Syllable". Sprachwissenschaft 4.13-23. Heidelberg. 1979b. "Two Clear-Cut Instances of Immanent Causativity". Sprachwis­ senschaft 4.355-358. Heidelberg. 1979c. "A Short Note on the Nonsensicality of Localistic Case Theories". Sprachwissenschaft 4.478-479. Heidelberg.

LIST OF PUBLICATIONS

Xiii

1979d. "Georgisches Gequengel in dem CF-Internat für Kasus". GIP (= Germansk Instituts Publikationer, Odense Universitet) 14.1-13. Odense. 1980. Linguistic Studies. Heidelberg: Carl Winter, 153 pp. This book contains the following papers: 1980a. "The Combined TRs and Their Conglomerations" (11-23); 1980b. "On the Nucleus Constructions in Human Language: Preamble to a semasiosyntax" (24-58). (English translation of 1977c; repr. in the present volume); 1980c. "The Basic Role of Welsh in a Universal Syntax" (59-68); 1980d. "A Constitutional Approach to the Lithuanian Case System" (69-83); 1980e. "The 'Cases', Hjelmslev, and the Cases" (84-113); 1980f. "The Total Analysis of a Sentence" (114-127); 1980g. "Evidence for the Dativity of Datives: Japanese and German" (128-138), and 1980h. "Death before Dying. Some reflections on linguistic policy in connexion with averageness and the death of minorities" (139-148). (English translation of 1976e.) 1981. "Qu'est-ce qu'un nom?". Logos semantikos: Studia linguistica in honorem Eugenio Coseriu 1921-1981, Vol. II: Sprachtheorie und Sprachphilosophie ed. by Harald Weydt, 317-329. Berlin: Walter de Gruyter; Madrid: Gredos. 1982. "A Semiotical Analysis of Complex Mesonomic Verb Structures in German and North Frisian (Sölring): A constitutional contrastive study". Friserstudier 2.9-74. Odense. 1983a. "The Cases Unravelled (or: Back to linguistics!)". NOWELE 1.77112. Odense. (Repr. in the present volume.) 1983b. "A Disseminational Analysis of Human Language Sentences: Prolegomena to a constitutional theory of language". Proceedings of the XIIIth International Congress of Linguists, August 29 — Septem­ ber 4, 1982, Tokyo ed. by Shirô Hattori & Kazuko Inoue, 523-527. Tokyo: Sanseido. 1985a. "A Final Wrest: On existence, existential disseminations and their delegations". NOWELE 5.77-105. Odense. (Repr. in the present volume.) 1985b. "Zum Gliederreißen der linguistischen Satzanalysen". Studi Linguistici e Filologici per Carlo Alberto Mastrelli, 123-142. Pisa: Pacini.

NIELS DANIELSEN IN MEMORIAM LAURITS RENDBOE Odense

Universitet

Niels Skovgaard Danielsen, by some held to be one of the most gifted Danish linguists since Rasmus Rask, was born on 20 May, 1933, some 100 years after his pioneering predecessor. Like Rask, he died prema­ turely, before he had realized his full potential, and, in a number of ways, his life and career resembled that of Rask: both were country boys who manifested their linguistic excellence at an early age, and, luckily, both were given a proper education, having competent teachers. Also, both set their course towards philological studies right away, developing their extreme talents for independent research, shaking free of the bonds of tradition to pioneer new methods and theories in their chosen fields. And both had to contend with disparagement and lack of appreciation, and, tragically, both had their brilliant careers brutally curtailed by malignant illnesses. Nevertheless, both left legacies of lasting value, and their achievements will long stand as examples to others. To chronicle the life of Niels Danielsen is in itself a rather straightforward affair, because he knew only one way of movement, namely forward. From the outset he set his course straight towards his goals, never looking back, but forging ahead as long as his health permitted. Thus, this will mostly be a straight narrative, moving forward with Danielsen through his fast-moving career, with only a few comments here and there, to elucidate the development. Although Danielsen was of old farming stock, he happened to be born in Hellerup, a northern suburb of Copenhagen (his mother was visiting with her mother there), but he grew up on his father's farm 'Hillestedgaard', which was situated in the flat but fertile sugar-beet farmland of Lolland, and it was here that he showed his foreign language acu­ men at an early date. Having been taught the alphabet at the age of three, 'Little Niels', as he was called (he was the youngest of three), was soon an avid reader, and before long he was absorbing large amounts of spoken Polish from the seasonal workers imported from across the Baltic. Owing to their illiteracy, however, they could not teach him the written language, a fact which he deplored — but a good

XVi

NIELS DANIELSEN IN MEMORIAM

start had been made, and soon another foreign language intruded itself into the hitherto peaceful world of young Danielsen. In the spring of 1940 (Danielsen was then seven) his homeland was occupied by the Germans, and some German troops were garrisoned in Maribo, the little town where Danielsen had entered school that same year. The bright boy soon began to pick up the German idiom, first from the marching songs echoing through the streets, and then from the daily language of the soldiers. Thus aroused linguistically, he soon set himself the task of learning more foreign languages, beginning to study, on his own, first Russian (he had discovered a Russian grammar in the local library), Italian (his father had given him a second-hand Italian grammar book for Christmas), French and Spanish (for which he bought textbooks himself), and in 1946, right after the end of the war, he was able to supplement his studies with some radio courses then introduced. All this naturally gave him a tremendous head start before entering the gymnasium (high school), from which he, after ardent studies, graduated in 1951, a full year ahead of his contemporaries. His final score consisted of straight A's, except for one A-minus in written German — the reason for which was symptomatic of Danielsen's pen­ chant for the unorthodox: at the final exam he decided to write his paper, not in the standard script, but in the old German hand, but in all the excitement he clean forgot that in this style a double -s has two forms, one for medial and one for final position, and he used just one form throughout! Following this unique experience, Danielsen relaxed during his sum­ mer holidays by hitch-hiking to Morocco, taking his first look at Arabic at the same time, and then, at the start of the autumn term, he enrol­ led in Copenhagen University, to study German and French. Happily, among his teachers were found some of the most illustrious academic personalities of the university, such as L.L. Hammerich, professor of German, a giant among philologists, and the idol of the students, Da­ nielsen not excepted. Another professor of German was Peter J0rgensen, who was an expert on Low German and Frisian, while Kaj Barr, pro­ fessor of Latin and Greek, was a great authority on Turkish and Irani­ an philology. Close to this linguistic corner was Kaare Gr0nbech, profes­ sor of Central Asiatic languages, and last but not least must be men­ tioned the internationally acclaimed phonologist and expert on all things Baltic, Louis Hjelmslev, professor of linguistics and author of the theory of glossematics. No wonder, then, that young Danielsen's academic studies soon pro­ gressed at a fast pace, also his extra-curricular research, which at ti-

NIELS DANIELSEN IN MEMORIAM

xvii

mes led him to far-away places to find the needed guidance. Thus, most of 1956 was spent studying in Berlin, and in 1958 he stayed a full term in Perugia. In 1959 he studied for a term in Beograd, going on from there to pursue his Greek studies in Athens, and in 1960 he went to Munich. He was not idle at home, either; in 1957 he earned his first academical award as runner-up in a prize contest, gaining a proxime accessit for a paper on Numeral Systems in Various Languages. This, indeed, is a most amazing paper from a young student of 23, if for nothing else then for the languages treated: Indo-European, Gothic, Old Norse, Old High German, Old Saxon, Old Frisian, Latin, Classical Greek, Old Bulgarian, Lithuanian, Old Irish, Albanian, Armenian, Avestan, Sanskrit, Finnish and Chinese! Moreover, the budding young linguist has the temerity to deplore that he cannot include Old Turkish, Eskimo and some Semitic and Bantu languages within the framework! The 18 languages included show Danielsen's predilection for study in depth, for going back to the old forms, the very roots, so to speak, of human language. To him all languages were of equal value and worthy of diligent study. As for the paper itself, it is quite detailed and well structured, and if the presentation appears slightly stiff at times, it is no doubt because the youthful author stuck to his sources with meticu­ lous precision. Danielsen, however, never content with second place, soon won his academic spurs: in 1959 he was awarded the gold medal of Copenhagen University for a paper on Negations and their Use in Middle High German, an effort which marks his real breakthrough in linguistic research. The reason is not so much the number of languages studied (about 30 this time), but because here Danielsen sets the pattern for his future research: he is no longer content to stay within the bounda­ ries set by others, but from now on he marks off his own limits and course, as he sees fit. He also aims high, trying for the universale of language as shown in this passus about one of his goals, which is to arrive at an absolutely unambiguous method to be used in a concise compa­ rative description of the structural and functional characteristics of the nega­ ting elements. The most profitable result of this study would be if this method could be used, essentially, not merely on the language treated in a given situa­ tion, but on any given language.

From then on Danielsen would pursue this course, and as in this case he would also coin new terms for the linguistic phenomena discovered, when such were not already available. In 1961 Danielsen studied for nine months at the Sorbonne, and in 1962 he graduated from Copenhagen University as candidatus magiste-

XViii

NIELS DANIELSEN IN MEMORIAM

rii, after which a UNESCO-scholarship enabled him to go to Krakow University for the first of four study periods spent there during the years 1962-64. The same year he also studied in Moscow and Warsaw, before returning to his old university, to teach German in the autumn term of 1962. At the same time he pursued his own special studies, i.a. taking lessons in Chinese, while concentrating on the completion of his formal education by taking his High School Teacher's Certificate in Ger­ man, French and Russian. In 1963 his studies took him to Rumania and in 1964 to Hungary, and because of his achievements as a teacher of Turkish in the University Extension ('Folkeuniversitetet'), the Turk­ ish government invited him to visit the country as its guest in 1965. The sixties were a busy time for Danielsen, now an amanuensis (lecturer) in the German Department of Copenhagen University, as he also taught in the University Extension and worked as a free-lance producer for the Danish State Broadcasting Service. Life, however, is not all work and study, and finally its realities caught up with our swift-moving linguist: in 1964 he got engaged to a girl from his own island, and in 1966 they were married. Theirs was a blessed union, as a boy was born in 1967, and a girl in 1968; indeed, Mrs Dorrit Danielsen proved herself a devoted wife and mother, supporting and assisting her husband in all his activities over the years. About this time Danielsen also changed his scene of work: a new university was being built up in Odense, the old city of Rasmus Rask and Hans Christian Andersen, and Danielsen now accepted the post as senior lecturer in the German Department there, besides assisting in teaching the classics. In 1967 he went on a study tour of Greece, and in 1968 he published his first major work, Zum Wesen des Kondi­ tionalsatzes, as No. 1 in the series Odense University Linguistic Studies. Having found this new base, Danielsen really got into his stride, and in the same year he defended his doctoral thesis Status und Polarität im Gotischen, im Lichte des Kymrischen dargestellt, at Aarhus Universi­ ty. Besides treating the languages given in the title, this was in reality Danielsen's first major presentation of his new semasio-semantic theory, of which a mere inkling had been noticed in his gold medal paper al­ most a decade before. It was very well received, by some even acclaim­ ed as a masterpiece. After this Danielsen was made the first full professor of Germanic languages and Medieval philology at Odense University, but now he al­ so faced an entirely different problem. In 1965 he had been taken ill with some mysterious ailment, which the doctors found it hard to diagnose, but in 1968 he had another attack, and now the true nature of the illness was determined: Danielsen was suffering from dissemina-

NIELS DANIELSEN IN MEMORIAM

xix

ted sclerosis. This was a heavy blow to a man who had always enjoyed perfect health, being able to do what he wanted, in work and in play, as well as in travelling — but who was now destined to become more and more dependent on others. Here his wife stood him in good stead, and, assisted by her and his friends, Danielsen decided to go on with his work as long as possible, teaching and researching, even travelling abroad, almost as before. Indeed, in 1972 he received an honorary grant from UNESCO, making it possible for him to attend the XIth Congress of Linguists in Bologna, and the following year saw him in America where he, inter alia, studied the language of the Menomini Indians, who requited the visit by making him an honorary member of their tribe! At long last he was able to visit Iceland in 1974, to study the favourite language of Rasmus Rask on its own home ground. Besides adding this 'world language' to his collection, he also became the lifelong friend of Kristján Eldjárn, the late president of Iceland, who invi­ ted him to write the syntax of Modern Icelandic according to scientific principles, a task he did not live to complete. As the bibliography of this volume shows, Danielsen did not let his gradual incapacitation deter his production, and during the seventies he wrote a number of scholarly articles and books, including two of his major theoretical works. First came the only comprehensive presentation of his new language theory ever to appear in his own native Danish, Fokus pá syntaksen ('Focusing on Syntax'), published in 1974, and then, in 1976 An Essay on Nomos and Human Language, published by the Royal Academy of Sciences and Letters. Years earlier Professor Hamme­ rich had stated that he had found it 'heavy going' to read Danielsen's thesis (he was one of the official opponents), but this Essay is very much heavier! Not because of the language, though, for it is written in Danielsen's usual lucid style, but because here he goes all out to show what his semasio-syntactic theory contains, making use of his entire ar­ ray of new grammatical terms, and of almost his whole arsenal of exo­ tic languages, to illustrate the linguistic phenomena described in the book. His close friend, the noted Swedish sinologist Olov Bertil Ander­ son, says: His greatest work, An Essay on Nomos and Human Language, is a unique ef­ fort with examples from more than 40 languages. This means that no little energy is required just to read through it. But the strange thing about it is that apparently all the examples are correct and appropriate in all these fields. (I can handle only something like twenty-odd languages, but I feel confident that all the other examples are equally correct down to the last comma).

XX

NIELS DANIELSEN IN MEMORIAM

No doubt this assessment is correct, but the count of languages is not: Danielsen uses more than 80 languages to exemplify his linguistic theo­ ry! During this period Danielsen received Handelsbankens Forskningspris ('Research Prize of the Bank of Commerce', 1977), and various grants for his studies and travels. In 1977 he also served a full term as visi­ ting professor in the University of Firenze, and in 1978 he lectured on Baltic (esp. Lithuanian) syntax in the University of Toronto. In 1979 he lectured on linguistics in Galway, and the same year he initiated Frisian studies in his own university, arranging for a regular Friserdag ('Day of Frisian Studies') to be held, alternating with visits by scholars and students from Odense to the North Frisian region. From this ema­ nated a series of Friserstudier ('Frisian Studies'), being collections of the scholarly material presented at these conferences, including Danielsen's own original contribution (cf. 1982 in the bibliography). Then, in 1980, he undertook a most ambitious project, as he accepted an invitation from the Austrian universities to lecture in Vienna, Graz, Klagenfurt, Innsbruck and Salzburg. For the planning, funding and execution of this tour he had the assistance of a number of friends and institutions, being sent off in grand style in a specially equipped Merce­ des car, driven by a trained ambulance driver. This eleven-day trip was a great success, as Danielsen was acclaimed everywhere, not so much for his remarkable ability to overcome his severe handicap, but all the more for his lucid exposition of modern linguistic theory. When later on the Danish Sclerosis Society produced a film about this illness and its victims, Danielsen had the leading part, and not a few pictures from the grand tour of Austria were included. In 1982 Danielsen made his longest journey ever in the service of modern philology, as he attended the XIIIth Congress of Linguists in Tokyo, delivering a lecture later published by his hosts (cf. 1983b in the bibliography). This time he was escorted by his wife, who better than anyone else was able to minister to his special needs. The next year he assisted in starting a new linguistic journal at Odense Univer­ sity, namely NOWELE, which honours him with this volume, and he was on its Editorial Board from the start to his death. Here he publish­ ed his final major effort, fittingly called 'A Final Wrest' (cf. 1985a in the bibliography). By this time, however, Danielsen's health had deteri­ orated severely, and in 1986 he was hospitalized, necessitating extended sick-leave. It soon became clear, that he would no longer be able to handle the strenuous work at the university, and so he was forced to retire from his post. While he continued to show great fortitude in his increased sufferings, it soon became evident that this was a fight not

ON THE UNIVERSALITY OF LANGUAGE* And God said, Let there be light: and there was light. These are the most surprising words ever committed to paper by human hand. How could they possibly be formulated. In a split second a maximum of ima­ ginative thinking, couched in a minimum of linguistic garment, presents itself to the wondering eyes and ears of the human race. This extraordinary passage from Genesis expresses a metaphysical theory with the remarkable feature of not being contradictory. Divinity, by nature, is never contradictory. Instead it is miraculous: It conveys to us how Something may arise from Nothing. Non-light turns into light. How it happens that Something arises from Nothing, only the Lord and, so it seems, Noam Chomsky can tell. The latter, however, sets out from the wrong starting point: Ab initio erat nox. And the night retrea­ ted before the imperative of transformations and turned into non-night. But deep darkness remained. All this happened, because the generativists thought themselves wiser than the Lord. But nowhere Genesis has it this way: And God said, Let there be language. The ancient Jews were clever enough not to make such a contradictory statement. For what already existed could not come into existence. In the beginning was the Word — as is well known. And the Word was with God, and the Word was God — as is also well known. This declaration, however, does not come from a lin­ guist, but from St. John, the Evangelist, whose formulation remains un­ surpassed to this day. The generativists grossly offended against theory of knowledge when they, taking darkness as their point of departure, tried to persuade people of the existence of something which we would not know before­ hand but would have to discover.1 This is the origin of the most severe tautology in the history of linguistics.

Translated from the German original "Zur Universalität der Sprache", Sprachwissenschaft 1.1-45 (1976a), by Per Bærentzen. 1 Maria Vittoria Giuliani in her excellent paper (1969:120) says: "Una teoría linguistica per Chomsky deve essere in grado infatti di definire non solo la gram­ matica di una lingua particolare, ma anche di tutte le lingue naturali; non solo di descrivere, ma anche di spiegare le lingue."

2

PAPERS IN THEORETICAL LINGUISTICS

Tautologies could become extremely tiresome because if they are not banned, they tend to grow infinitely. In the end their growth will cause their death, since nothing new will ever emanate from them. The Old Testament theory of the birth of light satisfies a principle that is crucial to any theory of interest: it is simple. Also the theories of Copernicus, Galileo and Einstein are simple. Otherwise they would probably be of no interest. However, the biblical theory has no bearing on scientific work. It has to do with its supernatural character. We shall claim that a theory must be a theory of something, which may sound like a truism but regrettably is none. The linguisticism of late which has so often been warmly welcomed demonstrates this only too clearly. As a matter of course the science of language will always be obliged to accept the fact that its object must be identified with the science itself and its tools. For that very reason linguistics — unjustly — is taken in to dinner by scholars working in many other fields and displaying an affability of the highest order. Especially the epistemological somer­ saults of the last two decades will make the coming generations look with great amazement upon linguistic pseudoproblems derived from old and faded discussions. During the interregnum of transformationalism terms like 'competence', 'performance' and 'generative' became extremely injuring to the re­ putation of our science. In the mid-fifties about four billions of people said good night to each other without imagining that they were going to wake up a few hours later with the honorary term of 'competence' stamped on them. That very night 'competence' was deprived of its meaning. A 'performer' automatically became a 'competent' if he chose the sensible alternative to absolute silence. People neither would nor could fully agree about the real meaning of 'generative' and 'generate'. Some of them thought that these terms referred to the deep and genuine meaning of words (see Chomsky 1970), while others realised the nonsense of this and with some modification interpreted them as 'describing' and 'describe' respectively (see Hovdhaugen 1971:9; Fowler 1971:6). That scholars describe, is nothing new, I should think. Already Pānini describes, and so did the grammarians before him. Linguistics to a great extent turned into a leisure-time activity of people dealing with English syntax. Noam Chomsky makes the follow­ ing comment on this: These are, then, two distinct ways of approaching what is clearly the most fundamental question of linguistic science, namely, the question of linguistic

ON THE UNIVERSALITY OF LANGUAGE

3

universals. One way is by an investigation of a wide range of languages. Any hypothesis as to the nature of linguistic universals must meet the empirical condition that it is not falsified by any natural language, any language acqui­ red and used by humans in the normal way. But there is also another and, for the time being, somewhat more promising way of studying the problem of universals. This is by deep investigation of a particular language, investiga­ tion directed towards establishing underlying principles of organization of great abstractness in this language. (Chomsky 1970:57-58)

This well-known argumentation of great trenchancy seemed to provide a whole generation of linguists trying to follow fashion with comfortable solutions to all their problems. I am writing this paper just to show that no Chomskian highway leads to universal linguistics. It is not my ambition to prove that I am right. I just want to show that even in 1970 Chomsky lacked the broad view of things and therefore was unable to say anything true about human language. The revised edition of Language and Mind (Chomsky 1972) does not make me change my mind. Like Noam Chomsky I believe in linguistic universals. As is well known, many people take only little interest in the study of linguistic universals and would rather deny the existence of such things. In a certain way I find it a pleasant experience to be in full agreement with Noam Chomsky. But I am afraid that this will not last for long. His universals seem to have nothing in common with mine. When I speak about universals, I am referring to phenomena which mirror a universe. The Chomskian universals, however, only let the generative theory ap­ pear as the reflection of a fake universe. This is what happens so often: in the hands of a prophet the meaning of words change without diffi­ culty and normally with nobody paying attention to it. To Chomsky 'universal' means "that, which is going to be universal" instead of "that which is universal". The latter is the normal meaning of the word, and from this it would not be possible to derive universals from a random state of language like Chomsky suggests in the above quotation. On which specific state of language he is actually thinking, I will not tell. Those who want to know should consult Thomas Gardner, who in the der modernen Linguistik foreword to his so-called Hauptströmungen (1973) gratefully acknowledges the inspiration and the technical help which he has received from the members of his team of Anglistic lin­ guists. In that book religion is substituted for linguistics, common sense is drowned by the gospel of monoglottiness, the study of language is dead, and only dry demagogy remains, but they do call it linguistics. Where you would usually find three or four linguists, you now find an

4

PAPERS IN THEORETICAL LINGUISTICS

unlimited number of them. Economists have a name for this phenome­ non which is very much to the point here. They call it devaluation. And sometimes a devaluation may lead to what is called bankruptcy. It is our task to prevent total bankruptcy. In most countries the pro­ blem seems to have been solved already. During the last years many people in despair have proclaimed the death of linguistic science. This truly would have been something of a stroke to many of us, who now see numerous important problems lying before us, the solution of which demands our full involvement. We have to be very sensible when for­ mulating the objectives of our linguistic work, because the language of homo sapiens at last is seen to be fundamental and prerequisite for any other scientific study. Therefore we must try to consolidate the va­ lue of our work, to abstain from sectarianism and concentrate all our energies on honest work in the field of linguistics. Honest scientific work distinguishes itself first and foremost by being open to falsification and by its readiness, when falsified, to confess its failings. Again and again the scholar should willingly put his theory to a test; otherwise he does not deserve to be counted among scholars. Such behaviour, however, is possible only if firm principles have been established beforehand in this particular branch of study. A theory of nothing will not do. A theory has always to be a theory of something. This something is independent of the theory itself. So, for instance, the stars and the laws that they obey are independent of the theory that we have formulated about them. Already before Giordano Bruno was burned to death, his ideas about the universe were more true than the theories of those who sent him to the stake. The phenomena of which we formulate theories exist in total independence of our assumptions about them. A scientist should seek first and foremost to bring this ideas about things into full concord with the things themselves. And this is done by way of advancing theories which are open to falsifica­ tion. The possibility of falsifying a theory implies its universality. 2 When, 2 Cf. Karl R. Popper (1959, p.121): "Thus I regard the comparison of the empiri­ cal content of two statements as equivalent to the comparison of their degrees of falsification. This makes our methodological rule that those theories should be given preference which can be most severely tested equivalent to a rule favouring theories with the highest possible empirical content", and further (pp.121-122): "There are other methodological demands which may be reduced to the demand for the highest possible empirical content. Two of these are outstanding: the demand for the highest attainable level (or degree) of universality, and the demand for the highest attainable degree of precision". And finally (p.280): "Those

ON THE UNIVERSALITY OF LANGUAGE

5

for instance, we formulate a theory of the system of our planets, we may easily choose to leave out or to delete Saturn, if it does not fit into the system that we are striving for. But such a theory is subject to immediate falsification thanks to competent astronomers, who will in­ disputably prove the existence of that planet. Until the discovery of the planet Pluto we were convinced, that we possessed a total picture of the planetary system. The discovery of Pluto meant the falsification of that picture. The registration of the planet that we called Pluto was not the result of transformations or insertions, but a reward for accurate measuring of pertubations and for drawing correct logical conclusions from observations. I have been very careful in choosing as an illustration of my point this microcosm, harmoniously at rest in itself, which suddenly one day grows more univeral. It reveals a very far-reaching problem in a way, which is well-known at the moment and can be easily understood be­ cause we here actually are dealing with the universe so near to us and which has for any number of centuries been discussed eagerly in the learned societies and the smallest hut. Most people start listening, when the universe is mentioned. Indeed, a special interest for every­ thing universal seems to inhabit every human being. Ultimately this is the primitive force which sets quite a few hearts afire, and on lucky days it may lead to honest science. The demand for universality can be brutal, but it seems to be the highest purpose of any branch of science, or rather: it is the only raison d'être for science. A science that has not got the touch of universality or does not contribute to some sort of universal insight is not science. This holds true also of a so-called science which makes use of all sorts of tricks in order to create its own (false) universe. It is characteristic of universes that they exist whether we want them to or not. We just have to find them. Universes may be constructed by everybody. Linguistics often calls itself 'general' and thereby suggests that it has a universal character. But to a great extent the important and noble purpose gets out of focus and withdraws to a faraway place where it will not be subject to misconception and total rejection. Only few lin­ guists are fully aware that it is just this universality of linguistics (and of any theory) which guarantees its epistemological value, provided that it does not lack the empirical, the indispensable empirical founda­ tion. among us who are unwilling to expose their ideas to the hazard of refutation do not take part in the scientific game."

6

PAPERS IN THEORETICAL LINGUISTICS

I know the costs of maintaining this strict attitude and that it can be very testing to pay the bill. But I also know that there exists such a thing as a scientific conscience, and this master recompenses his ser­ vant immensely compared to the commanders of today's fashion. Through the years linguistic study has taught me that human lan­ guage contains a great number of universale. The simple fact that every human being of average intelligence is capable of learning a foreign language which may even be only distantly related to his mother ton­ gue, seems to show that some fundamental features of the language mechanism are shared by all human languages. This may sound all too primitive, like a commonplace statement. On the following pages, how­ ever, I will go into details about this enormous subject, at least as far as the normal limits of a research article permit. The linguistic universe is studded with stars that do not twinkle, and therefore some important universals may remain undiscovered for a long time. A very striking example of this we find in the universal of Polarity. Because of this universal being unknown through centuries the study of 'negation' in language has been totally bungled. Neither logicians nor linguists have paid attention to this astonishing feature, which is of the greatest importance for any linguistic study within semasiosyntax. Let us consider two German sentences like: Du redest : Du redest nicht Until 1968 the interest of linguists concentrated exclusively on senten­ ces containing the word nicht like the one to the right. Characteriza­ tions like 'negation', 'negative' sentence or 'negative' phrase are in com­ mon use. This, however, permanently led to serious contradictions since the word 'negative' has at least two entirely different meanings, some­ thing which was not realised at once. The universal of polarity, which is a condition of any sentence analysis, was revealed by mere accident. Some years ago an Indian student persuaded me to take a glance at some Dravidian languages. I was at that time writing my doctoral the­ sis and was eagerly engaged in the study of Celtic. But this juxtaposi­ tion had the most wonderful consequences which, even in my most fan­ tastic moments, I would never have dreamt of. Afterwards, when every­ thing looks that simple, you always ask yourself: Why did you not grasp this solution of the utmost simplicity at a much earlier time. Why has nobody else realised that when you have languages showing a pair of sentences like

ON THE UNIVERSALITY OF LANGUAGE

7

Du redest : Du redest nicht there will very probably be languages with the opposite distribution of linguistic signs of polarity, i.e., languages which will express explicitly the positive position whereas they keep the negative alternative un­ marked. There is no need of having more than one half of the polarity clearly marked against the other. In Kannada, a Dravidian language, I found the first signs of such a system which has been turned upside down (at least in comparison with German or Danish, for which of the systems is, in final analysis, upside down and which one shows the prime orientation?). In Telugu, the second most widely spoken language of India, I found this system fully developed: Chepputaavu

: Cheppavu

Welsh represents the third type with both polarizations having a spe­ cific sign: Yr wyt ti'n siarad : Nid wyt ti'n siarad This seems peculiar to most Aristotelians. 3 The reason why they wonder is quite simple. It never occured to them that there are languages which in comparison with German, English, French, etc. have their polarity turned upside down. You have to comprehend this fully in order to see how the Welsh system is almost ideal. Redundant, but charming. Here we face a principle which to a large extent seems to govern human language, i.e., the principle of Equivalence b e t w e e n Diame­ trical Opposites. These opposites at the same time through combina­ tion or disjunction imply a third position. In other words: the opposition X and Z implies the existence of Y (Y being "either X or Z" or "neither X nor Z" or both at a time). This sounds like a contradiction without necessarily being one. The Welsh polarization system is identical to neither the German nor the Telugu system but represents both at the same time. Whether you prefer to stress the "either-or" or the "neithernor" depends on the intention of your description. The following dia-

3 Thus Bowen & Rhys Jones (1960:15): "Note that y and yr are here meaning­ less particles" (!).

8

PAPERS IN THEORETICAL LINGUISTICS

gram shows our principle of identifying language universale: opposition

implication

The principle of equivalence between diametrical opposites has been totally neglected by generativists. Their transformations are thus a vio­ lation of the principle, because transformations are normative and antiuniversal, i.e., they are by nature incompatible with universality. But transformations are immensely destructive as they try to conquer a universe from which they have from the very beginning and for all ti­ mes excluded themselves. I will return to this point later, but for the moment I go on with the study of linguistic polarity. After the discovery of the simple polarity universal one thing is clear: the word nicht must be renamed. And in accordance with this the word 'nicht nicht' has to be named. J u s t in order to know what we are talking about. 'Negation' never was a happy term for the polarity sign nicht. Altogether, what is the negation? Logicians have never been able to explain that. Is the sentence Sie ist nicht

glücklich

more negative than the sentence Sie ist

unglücklich

and if it is, what do we mean by saying so? Semantically or asemantically? I only ask this question because so many stupid ideas have been put to paper about semantics. The English sentence Nobody ain't never doing nothing here (Cockney) would probably be much more 'negative' than its German equivalent Niemand

macht jemals irgend etwas hier

ON THE UNIVERSALITY OF LANGUAGE

9

and the sentence Wir sind alle keine Engel seems to be equally positive and negative. The subordinate clause in Ich tue es nicht, bevor er nicht wieder hier ist clearly contains a 'negation' without being 'negative'. This feature is found in many languages and seems to reveal an extremely interesting universal: Cekajmo dok ne prestane vetar (Serbo-Croatian "let us wait until the wind stops") Non lo dico finché non riviene (Italian "I do not say it until he doesn't come back") Yävat tava vinäso na bhavati, tävad gaccha (Sanskrit "before your ruin occurs, (then) go!") Hatutasoma kabla hajafika (Swahili "we will not read until he doesn't come back") The word nicht and logical negation (whatever that means) are in no way identical. The only thing we can do seems to be something entirely linguistic: We can speak about two types of sentences, the first one being sentences containing the word nicht (or an equivalent conjugation of the verb) and the second one being sentences without the word nicht (or an equivalent conjugation of the verb). The first type we will call recusative sentences and the second type we will call propositive sentences (cf. Danielsen 1968b). We now have at our disposal universal criteria which enable us, with respect to polarity, to make an optimal division of human sentences into four main groups: +/p +/r -/p -/r

(logically (logically (logically (logically

positive, linguistically propositive) positive, linguistically recusative) negative, linguistically propositive) negative, linguistically recusative)

Each of the four categories will be illustrated through the following examples:

10

PAPERS IN THEORETICAL LINGUISTICS

+/p:

+/r:

Er kommt morgen Sie ist unglaublich

häßlich

Was du nicht alles sagst! Frage ihn, ob er nicht morgen kommen

möchte

-/p:

Ihrer Ansicht nach schuldete keiner keinem etwas Niemand hat ihm jemals irgend etwas Böses zugefügt

-/r:

Er kommt nicht morgen Er ist nicht ohne Humor

Surely the universal of polarity is the most categorical universal of hu­ man language. A sentence has to be either propositive or recusative. Only the newly invented German combination jein, a contraction of ja and nein, does not obey that general law. The following diagram illu­ strates our universal: Universal of Linguistic Polarity: propositive recusative jein German: Du redest Welsh: Yr wyt ti'n siarad Telugu: Chepputaavu

Du redest nicht Nid wyt tVn siarad Cheppavu

Many languages express the two polarizations through a specific propo­ sitive or a specific recusative conjugation of the verb. This holds good of Eskimo, Swahili and Japanese: propositive

recusative

Eskimo:

Naluvara "I am unconscious of it"

Nalúngilara

Swahili:

Ninasoma

Sisomi "I do not read"

Japanese:

Chichi wa zasshi wo yomimasu "my father reads illustrated papers"

"I read"

"I know it"

Chichi wa zasshi wo yomimasen "my father does not read illustrated papers"

ON THE UNIVERSALITY OF LANGUAGE

11

In Finnish the recusative sign is an auxiliary verb, i.e., Finnish uses a polar nucleus for expressing the recusative polarization (see Danielsen 1974a; Anderson 1975). If one had to translate literally the present ten­ se paradigm of the Finnish verb into German it would look like the fol­ lowing: Ich nickte komm Du nicktest komm Er nicktet komm Wir nickten komm Ihr nicktet komm Sie nickten komm This discussion of the polarity universal of language already tends to exceed the limits of a normal article, and we therefore must force our­ selves to leave this topic and turn to other parts of universal semasiosyntax which are just as fascinating. Besides the polarity universal we find a series of other important semasiosyntactic universals. We will be dealing with the universal of Nuclei, the universal of Transversal Relations, the universal of Nomic Sentence Structure, the universal of the Critical Field of Di­ stribution and the universal of the Formal Garment of Sentences. On the following pages we will discuss these universals but first of all we will have to cut a knot just as tangled as the Gordian one. I am speaking of the universal of Sememes, as we cannot do without seman­ tics. The intensive work of isolating the basic semantic elements requires great effort. Many people choose not to lift a hand at all, and if this is the case, we all know too well what the mouth will speak. For a long time semantics was by many linguists regarded merely as a part of lin­ guistics which was seriously studied only by madmen. To illustrate this attitude, mention of Zellig S. Harris and the young Noam Chomsky will suffice. At a very early state of life, virtually during my first years as a stu­ dent in Copenhagen, semantics for me became a subject which you would talk about with great visions, and leisurely, because what was at that time meant to be the real thing in linguistics — besides morpho­ logy, of course — was phonology. Semantics would come into discussion in the fullness of time. Ferdinand de Saussure had spoken about a pla­ ne of content (signifié) which was truly believed to exist, but it was im­ possible to say anything sensible about this phantom. And at the same

12

PAPERS IN THEORETICAL LINGUISTICS

time phonology promised so great and prosperous results that scholars would find it anything but boring. This, however, was exactly what happened to many of us: an immen­ se boredom with phonetic detail befell us. In linguistics the climate du­ ring the sixties had become so dry that a glass of water from the well in Massachusetts was, to the taste of many dehydrated souls, like the most delicious nectar. So, in one night structural linguistics fell a vic­ tim to its own descendant, generativism. Once again poor semantics had fallen into a trap. Although with the time passing it was legalized by the generative party and even taken up by some of its adherents, semantics was handled without the slight­ est trace of universality. This is due to a false interpretation of the term 'universal', as I have already noted earlier, and also due to the fact that transformations operating on atomistic postulates lead to ex­ cessive generalizations. The way to semantics is in the opposite direction. And that road is not a macadam highway. It is a thorny one which offers its explorer no rest. It knows but one commandment: "Don't give up but try to get a good grasp of things!" If at this very moment you feel sick by the smell of empiricism, take my advice and stop reading this paper. You can trust your sensation, and experience tells me that you and I will never agree. This is a case of morality. Should we grossly misuse linguistics? Some people say no and some say yes, but most people normally are uncertain what to do, or they are prepared to discuss the case until things calm down. Science must a priori be supported by empirical observations. This is a simple statement of fact. That this is the way how things are, I will prove later in this paper. A glance at the world should convince us that false competence has given rise to a lot of problems for mankind. The study of the universal semantic entities started out in the late fifties. Only fifteen years later some sort of first model was seen. The creation of a model, however, is a great step forward, because a model, however insufficient, can be used as the starting point for discussions. The examination of semantic entities in human language has very soon shown the existence of some ever occuring universal entities. They are called sememes and can be illustrated in a two-dimensional paradigm. A sememe combines a zonal content and a row content. The semantic zones are represented by various A (capital alpha). For instance, the personal zone is represented by German wer, English who, French qui, etc. Similarly, the zone of species is represented by German was, English what, French que, etc., and the temporal zone is represented

ON THE UNIVERSALITY OF LANGUAGE

13

by German wann, English when, French quand, etc. The total list of A in a specific language we call the Constitutional Axis of that lan­ guage. The constitutional axis of German looks like this:

idiogenous

wer was welch/welcherlei wann wohin wo welcherseits woher wie wieso warum wessenthalben weshalb wessentwegen weswegen wessenungeachtet wesbezüglich proportional wieviele wieviel inwiefern inwieweit

zone of person species designation quality time goal locus alternative sides origin modality causality cause concession reference dissipated quantity compact quantity extension

The struggle for the establishment of the optimal constitutional axis was extremely hard and required great patience (see Danielsen 1975). Zones and rows intersect. Each row is established through the prin­ ciple of exclusion, i.e., a row is defined by a specific semantic category which is found in no other row. If somewhere in the system of rows a new semantic entity turns up, this automatically and unconditionally calls for the establishment of a new row. Besides the idiogenous and proportional zones we also may have sentence-generating (mostly verbal) zones. The rows are divided into identificatives (I) and dispositionals (D). The identificatives may be demonstrative, conjunctional, relative, subcentric, etc. and the disposi­ tionals may be potential relatives and conjunctionals, eventuals, faculta­ tives, etc.

14

PAPERS IN THEORETICAL LINGUISTICS

Let us demonstrate how easy it is to work practically with this very simple delimitational system. We choose as an example the zone of lo­ cation Zloc and the row of existence (DN = dispositional row N):

DN Zloc

X

This is the notation of a universal entity, which manifests itself in Ger­ man irgendwo, English somewhere, Russian  , but has to be para­ phrased in Danish as et eller andet sted and in French as quelque part. However, it makes no difference to our graph as a true picture of the totally independent universal, which is part of all languages, that it is in some languages expressed by way of a specific material entity and in other languages by way of a materialized paraphrase. Our system makes it possible to isolate a great number of sememes so that they in a clear and often very telling way stand out from one another. We may have for instance a verbal zone of locomotion (this A being represented by Mongolian xaitsi- = "where go/-") and a demonstra­ tive row (ID = identificative of row D):

ID Zvb.locom.

X

This universal designates just one single semantic entity, which we may know from Gothic as the element hirj/- where it occurs six times in the text of Wulfila 4 and remains a permanent semantic mystery to Germa­ nic philologists. Gothic hirjl- is, if looked at through universal glasses, simply the demonstrative identificative of the first person in the verbal zone of locomotion. 4

Mark 1.17, 10.21, 12.7; Luke 18.22; John 11.34, 11.43.

ON THE UNIVERSALITY OF LANGUAGE

15

Or let us consider the following example: We may have a verbal zone of existence (this A being represented by Quaday-Arabic chenho = "what is?") and a demonstrative identificative row (ID):

ID 7 Zvb.exist.

X

This universal symbol, if stress is laid on pure existence, designates the semantic entity found in Latin ecce which has its counterpart in several languages. So Italian ecco is not the imperative as suggested by Palle Spore (1975:§207,5), but clearly the demonstrative of the verbal zone of existence. I have to return to the human 'negation' which I discussed at the be­ ginning of this paper. In the system of sememes we in fact do find a row that can easily be called a negative one. It contains elements like (German) niemand, nichts, kein, keinerlei, nie, nirgends, etc. Their row carries the mark M and consequently they are named DM because they are dispositionalia of row M, i.e., negative dispositionalia, for instance:

DM Zpers

X

niemand

In connexion with the genesis of DM-elements two other dispositional rows are of a certain importance, i.e., the existential row DK (v. Eng­ lish someone, something, some, sometime, somewhere, sometimes) and the restrictive row DL (v. English anyone, anything, any, ever, anywhere). Just to make things clear: all languages possess DM, DL, and DK. On the plane of expression (signifiant) they mingle with each other to a very large degree. In Sanskrit syncretism is complete: (a) (b) (c)

Ucchritam sahata kascit ("someone can bear elevation") Ucchritam sahata kascit? ("can anyone bear elevation?") Nocchritam sahate kaścit ("nobody can bear elevation")

16

PAPERS IN THEORETICAL LINGUISTICS

In (b) the element kascit together with the interrogative intonation of the sentence establishes the DM-sememe. In (c) the sign of the recusative polarity na (amalgameted with ucchritam to nocchritam) refers with the effect of a negation to kascit. So the recusative sign in a sen­ tence may be either 'neutral' or 'engaged', the latter being the case if, besides its recusative function, it has a specific negative influence on some part of the sentence. The German word nicht originally was a pure sememe meaning "to no extent" and so it was a negative dispositional DM in the zone of extension corresponding to modern German inwiefern and inwieweit. As is well known nicht in earlier times was used to put a specific negative weight to the sign of recusative polarization, but this feature gradually became obscured. The same development also took place in other lan­ guages such as Welsh, French, and Arabic. The DK and the DM indicate an opposition X-Z which implies Y (the restrictive dispositionals DL): irgendwann

nie jemals

The principle of equality between diametrical opposites plays an impor­ tant role whenever we try to identify exactly the basic semantic entities of language. Besides the two-dimensional linguistic elements there are onedimensional elements (signs of status like German daß, ob, Polish czy, Russian  , conjunctions like German und, oder, aber, denn, elements representing a whole row like German gerade, sogar, nur etc.) and complex elements (determinations like German Haus, Hoffnung, Kom­ mentar, quietschen, meckern, trinken, blau, gewinnsüchtig, generell etc.). The complex elements are conventions and should be handled with great care. Almost every attempt to interprete them semantically will inevitably result in a poor paraphrase. I have said enough about the basic semantic entities of language. An exhaustive treatment of them still belongs to the future and seems to be the strait gate through which the analyst must pass if he will have a thorough grasp of human language. The unavoidable study of these entities constantly discloses new universals. One of the most important universals is the universal of the nuclei (cf. Danielsen 1975:61-65). A contrastive analysis of sentence-generating

ON THE UNIVERSALITY OF LANGUAGE

17

A's leads to this universal. Let us consider the following example of a verbal zone of existence and an A-row crossing each other and resulting in this universal sememe:

A Zvb.exist.

X

In Quaday-Arabic this sememe is represented by the specific A chenho, the German correspondence of which is the paraphrase "was gibt es". The contrastive analysis of Quaday-Arabic and German reveals the astonishing fact that German can only materialize the Quaday-Arabic A chenho by secreting a verbal element, i.e., a nucleus. In addition to the nucleus [GEBEN] German, in the same function, also has the nuclei [SEIN] and [HABEN]: Es gibt kein Wasser hier im Haus Was ist denn? Ist was? Heute hatte es viele Zuschauer im Theater This is not a mere coincidence. To be sure of this, you just have to ob­ serve the internal connection between the verbs of the following con­ structions: Ich gebe Ihnen das Buch Das Buch ist für Sie Sie haben das Buch and [HABEN] can be seen as X and Z, i.e., as a pair of opposites, and the corresponding Y (the implied entity) is [SEIN]. At the same time [SEIN] can be seen as a Y between an other pair of opposites, i.e., X

[GEBEN]

[WERDEN] and Z [BLEIBEN], and so we get the diagram:

18

PAPERS IN THEORETICAL LINGUISTICS

These five nuclei are universals. They emerge in a great number of languages as auxiliary verbs or some other sort of nuclei. This complex matter needs to be studied thoroughly. For the moment we can only suggest the direction such a study should take by listing a limited set of examples: Es gibt viele solche Kleinbürger There are many different sorts of people Dus i out e demeines (Old French "dukes were there, and lords") Il y a beaucoup de problèmes Es werde Licht! Der bliver mangel pâ juletrseer (Danish "there will be a lack of Christmas trees") Er wurde geschlagen Er blieb stehen Han bleu forelsket (Danish "he fell in love") To X [GEBEN] we can establish [BEKOMMEN] as an extra Z. The Y between them would then be [BEHALTEN]:

Er bekam das Märchen erzählt Han fik ham overtalt (Danish "he had him persuaded") He kept whining Indeed we can establish also [NEHMEN] as a Z to X them we have [HALTEN] as Y:

[GEBEN],

and between

The sailor took ill Tiene escribida la carta (Spanish "he has written the letter") To this group of nuclei also belong positional verbs like [STEHEN], [SITZEN], [LIEGEN] and [HÄNGEN], which very often function as auxiliary verbs: Drumān bhedayämäsatuh (Sanskrit "they let the trees fall") Sto andando (Italian "I am walking") Hij ligt te dromen (Dutch "he is dreaming") De kleren hangen te drogen (Dutch "the clothes are drying")

ON T H E UNIVERSALITY O F L A N G U A G E

19

A contrastive analysis of Mongolian and German shows us two further partial groups of nuclei. The German equivalents of the Mongolian A's ja- and xaitsi- are was tul- and wohin gehl- respectively and so the elements [TUN] and [GEHEN] are seen to be nuclei of human language. Syngenetics of the nucleus [TUN] are especially [MACHEN] and [LASSEN], but also entities like [BEGINNEN], [FORTSETZEN], and [AUFHÖREN]: 7 do not understand Er tut singen Er machte mich zweifeln Er ließ ihn laufen Er fins an zu weinen Han blev ved at hamre i bordet (Danish "he kept hammering the table") El hombre que acaba de llegar es un primo mio (Spanish "the man who just arrived is my cousin") Amekwisha kuja (Swahili "he has already come") Syngenetics of [GEHEN] are [FORTFAHREN], a n d [FALLEN]:

[KOMMEN]

and

[FAHREN]; [SICH ERHEBEN],

He went completely mad Il va venir Il ladrone venne ammazzato

dalla ragazza (Italian "the thief was killed by the girl") Hún fór alt slóra (Icelandic "she began to chat") He went on whispering He fell ill

All the nuclei that we have found in this way we will call A-nuclei. They are verbal extracts of universal sentence generating A's and their syngenetics. They turn up in a great number of languages as auxiliary verbs or some other sort of nuclei, with distributional vari­ ations but with striking regularity. Besides the A-nuclei, human language has two further types of nuclei: modal nuclei and polar nuclei. Polar nuclei occur in Finnish and have already been dealt with in the first part of this paper. The modal nu­ clei simply are the so-called modal verbs. From a universal point of view they seem to be most closely related to the concept of 'status' (see below).

20

PAPERS IN THEORETICAL LINGUISTICS

Nuclei are monodynamic (Gott ist), syndynamic (er ist gekommen), diadynamic (dies ist zu ergänzen), taxidynamic (er is¿ gut), or exodynamic (er tu¿ Zucker in den Kaffee). As will be seen from this, a nu­ cleus is monodynamic if it stands alone. A nucleus is syndynamic if it is connected with an operative verb. A nucleus is diadynamic if it is connected to an infinite verb by means of an indicator. A nucleus is ta­ xidynamic if its syntactic function is that of a copula. A nucleus is exodynamic if it takes a direct objekt. Very closely related to the universal of nuclei we find the universal of t h e nomic sentence structure. All human sentences follow a uni­ versal nomos according to which every human utterance ideally shaped falls into three separate parts. Without these delimitations mankind would be unable to communicate. We would be unable to learn foreign languages. These delimitations seem to be the only universal regularity. Without them there would be no rules at all. We are now going to commit a mortal sin. We start out with an Eng­ lish sentence. But we do this because the rest will then work much mo­ re easily: John has paid the bookseller five dollars for the book several days ago This sentence can be considered ideal because it contains all three no­ mic parts and no further division of the sentence will be necessary. It consists of an endonomic part (has), a mesonomic part (John has paid), and an ectonomic part (the bookseller five dollars for the book several days ago). The mesonomic part of a sentence is the main verb plus the subject, if any. The mesonomic part may have in it an endo­ nomic part that will then be a finite nucleus. All the rest lying outside the mesonomic part makes up the ectonomic part of the sentence. The theory of the nomic sentence structure is a semasiosyntactic one, as semantics is never ignored in the analysis. From this also follows that the structure of the German counterpart to our sentence conforms exactly to our analysis, and the only thing that must be mentioned is the placement of the ectonomic part between the endonomic part and the operative verb according to the peculiar German nomos: Hans hat schon vor mehreren Tagen dem Buchhändler Dollar für das Buch bezahlt

fünf

ON THE UNIVERSALITY OF LANGUAGE

21

It must be stressed that the mesonomic part of the sentence has been defined in such a way that, besides the main verb, it may perhaps hold a subject. A nomic analysis deals with all sorts of human utterances without causing logical harm to any of them. It must take into account that legions of sentences do not show even the faintest trace of a socalled subject. Let me just mention a few German examples: Ihm wurde von seinen Parteigenossen Dessen nimmt mich wunder Mir ist daran äußerst gelegen

geschadet

In some languages subjectless sentences are very common and may even be the norm. Some people submit themselves to the foolish mis­ conception that a sentence is bound to have some kind of subject. It is astonishing that still today we very often hear people talk about 'subject' although they have never considered what is exactly meant by that term. I find it very alarming to read the following which has been formulated by the greatest contemporary Oxford philosopher: The sentence 'It is snowing', for example, has no grammatical subject; it would be senseless to ask 'What is snowing?' But, apart from the fact that ordinary grammar is an unreliable support here — for one can also say, for example, 'Snow is falling' — it may well be said that this appeal is quite superficial. 'It is snowing', spoken in a suitable context, has some such force as 'It is snowing here and now'. What is to prevent us from reckoning 'here' and 'now' as subject-expressions denoting a time and place, and the rest af the sentence as a predicate-expression ascribing a character to this subjectpair? (Strawson 1971:215) So according to Peter F. Strawson, the sentence it is snowing has no grammatical(!) subject. But, so Strawson says, the general notion of 'snow' functions as subject, because it may appear as subject in a quite different sentence, viz. snow is falling. And furthermore the real sub­ jects of the sentence are said to be 'here' and 'now'. I ask myself: Will there ever be a time when philosophers and linguists will communicate in a sensible way with each other. In fact the sentence it is snowing does have a grammatical subject. I find it much more important and interesting to examine such senten­ ces that do not have any trace of a subject, grammatical or 'logical'. The intense study of the universal of the transversal relations ta­ kes us a step further in our investigation of the nature of subjectless sentences.

22

PAPERS IN THEORETICAL LINGUISTICS

A transversal relation is a -relation between two implied entities whereby  symbolizes the semantic content (Greek ) of the (sim­ ple or complex) verb of the sentence: Hans schlägt Peter Peter wird von Hans geschlagen Both sentences are simple transversal relations. They may look rather different but they to a very great extent express one and the same thing. Graphically they can both be represented in the following way:

Stasis

Epistasis

From a semantic point of view Hans plays the same role in both sen­ tences. He is the [A] = principle of content (A from Greek "begin­ ning"). Correspondingly Peter in both sentences is [O] = terminal of content (0 from Greek "terminal"). If the semantic [A] is made the stasis of the sentence we get: Hans schlägt Peter (S-relation) If the semantic [0] is made the stasis of the sentence we get: Peter wird von Hans geschlagen (R-relation) In German the latter is established by means of a nucleus. Other lan­ guages make use of a specific passive inflexion of the verb, for instance Latin, Greek, Danish, etc. Some languages such as Vietnamese only know of S-relations, others like Basque only know of the R-relation. It makes no sense to talk about a linguistic passive-transformation, not even in English, because English has not got a true passive but only a nuclear paraphrase. The basis of an R-relation is the stasis (i.e., Peter in the sentence above). The basis of an S-relation is the epistasis (i.e., once again Peter). The basis always remains one and the same, the semantic [O]. Many languages know of a third relation which we call the Q-relation. In the Q-relation the basis, i.e., the semantic [O], appears as epi­ stasis, and indeed as the epistasis of an S-relation. At the same time the semantic [A] appears as the epistasis of an R-relation. The two epi-

ON THE UNIVERSALITY OF LANGUAGE

23

stases we shall call epistasis 1 (or scopos) and epistasis 2. The Q-relation has no subject at all, a fact that has been totally ignored by tradi­ tional grammar. Scottish Gaelic has all three simple transversal relations: Tha Dömhnall air sgriobhadh an litir (S-relation) ("Donald has written the letter") Tha an litir air a sgriobhadh le Dömhnall (R-relation) ("the letter has been written by Donald") Sgriobhadh le Dömhnall an litir (Q-relation) ("written by Donald the letter [object]") Let us once again have a look at some German transversal relations (simple S- and R-relations) and let us add the Scottish Gaelic Q-rela­ tion because German does not have the simple Q-relation: Sein Bruder hat ihn unterstützt (S-relation) Er ist von seinem Bruder unterstützt worden (R-relation) Chuidicheadh le a bràthair e (Q-relation) ("him was helped by his brother") This shows the true nature of the German Q-relation. It is never a simple relation. The sentence ihm wurde von seinem Bruder geholfen is a semi-relation (consisting of the  wurde geholfen and the epistasis 2 sein Bruder), i.e., a semi-relation that has been extended by a diastasis (= dative object), the diastasis being ihm. The sentences ihm wurde von seinem Bruder geholfen contains no subject at all. It can be derived (or let us say 'transformed') neither from an S-relation nor from an R-relation in a simple and unambiguous way. Generative transformational grammer is the most severe contradic­ tion in linguistics for the last twenty years. There are languages that practically work with Q-relations only (such as the Northeastern Caucasian languages Bats and Agul). In Urdu and Hindi the Q-relation is the sovereign of the syntactic field. In modern Icelandic it is very common and in German it appears regularly: Mir wurde von dem griechischen Polizisten Mir ist daran gelegen Mich dürstet

gedroht

24

PAPERS IN THEORETICAL LINGUISTICS

What is missing in all such sentences is not the subject but rather the scopos or the epistasis 2. It should be noticed, by the way, that the sentence verb in a Q-relation may equally well be in the active or the passive (if the verb does not have a specific neutral form for use in the Q-relation — as is the case in Finnish and Polish). The discovery of the Q-relation was delayed two thousand years by the pupils of Aristotle. Generativists don't find it worth mentioning. But with the discovery of the Q-relation the syntactic world seems to us to have become a bit more coherent. It has, finally, opened up a syntactic universe. If we make the S-relation and the R-relation a pair of opposites (X and Z respectively), then the Q-relation is the implied neutral Y ("neither-nor" / "both-and"). The three transversal relations of human language we can picture in the following way:

S-relation:

R-relation:

Q-relation:

From this a question arises both naturally and spontaneously: Why finally should a Q-relation have a subject? It already contains all that a relation between two entities could wish for (O1 stands for any basis that can be placed in the ectonomic part of the sentence, and Ω stands for the epistasis of an R-relation and for the epistasis 2 of a Qrelation). A subject is the epistatis 2 of a Q-relation that has been stabilized in the mesonomic part of an S- or R-relation. The 'ergative' case desig­ nates such an entity that is becoming stabilized. This is our argument and we welcome any attempt to refute it. Transversal relations occur as simple, extended or reduced trans­ versal relations. We have already learnt a lot about the simple ones.

ON THE UNIVERSALITY OF LANGUAGE

25

Transversal relations may be extended by diastases (O2), metastases (Ω) or apostases (O3). A diastasis is a dative object: Er gab dem Jungen ein schönes Buch

s

o2

O1

A metastasis is a genitive object: Er bezichtigte ihn des S

Landesverrats O1 Ω

An apostasis is a complement in a case other than the stasis/epistasis case, the dative and the genitive: Scipio nudavit castrum S 1

defensoribus O3

These are the terminals of transversal relations. They may all (stasis, epistasis, diastasis, metastasis, apostasis) be expressed by means of in­ dicators (i.e., postpositions and prepositions). Indications for stasis and epistasis are called representations: Watashi wa shigoto wo oemashita (Japanese "I have finished my repr:S repr:O 1 work") Er wurde von seinem Vater bestraft repr:Ω Indications for apostasis are called transactions: Der Richter schätzt die Strafe auf zwanzig Mark ab ind:0 3 If in a language transactions emerge in the domain of diastasis or me­ tastasis we call them expanding transactions: Ich erinnere mich an meine Er lachte über sie

Urgroßmutter

We speak of reduced transversal relations when an epistasis is lacking or if the relation is obtuse:

26

PAPERS IN THEORETICAL LINGUISTICS

1. An epistasis is lacking: er singt er repariert den ganzen er beruhigt

Tag

2. The transversal relation is obtuse ('intransitive'): er schnarcht sie arbeitet er schmeichelt Intransitive verbs very seldom are totally intransitive. For instance, we might say was arbeitet dein Bruder? In this sentence the  arbeitet is combined with the epistasis was. In most instances the so-called intran­ sitive verbs are hiding their epistasis inside themselves. This epistasis being the nominal content of  is called π and its case is the partitive. In German it emerges as basis through a transformation of an S-relation into the corresponding R-relation: man singt

→

wird

gesungen

The entity man is not visible in the R-relation because according to German nomos it cannot occur as an epistasis. Some languages have got double semi-relations: Mis vantar bókina (Icelandic "I am missing the book") And so this most fascinating game goes on and on. Its richness is in­ exhaustible. One of the most important tasks lying before constitutional linguistics is to deliver a competent universal case theory by means of the transversal relations. We have already more than once discussed this very complicated but interesting problem and we will come back to it later, but it can only be solved if we make use of the universal of i n t e r f e r e n c e . This universal encompasses possessional entities in the genitive and dative plus similar entities with for and the interplay between them (cf. Danielsen 1976c; 1974b). We can for the moment only sketch it out, as it constitutes an enormous field of study. By the way, we must warn everybody not to pay greater attention to the case theories of recent years than they really deserve. I am refer­ ring first of all to Charles Fillmore's 'conceptual framework' (1968) and John M. Anderson's localistic case theory (1971), the former being pure-

ON THE UNIVERSALITY OF LANGUAGE

27

ly intuitive to a degree that equals the lack of universality characteri­ stic of the latter. I am highly doubtful as to how useful a case theory will be that is almost entirely based on the study of case-less langua­ ges. I will give just one example: In The Grammar of Case (p.129) John M. Anderson — like every Briton would do — considers the NP from John an ablative in the English sentence: Mary bought the book from

John

This point of view is not so that it could not easily be excused; on the other hand, it is of no interest because it has no universality. It re­ mains a specific English statement and manifests an extremely superfi­ cial English grammar. The French equivalent of this English sentence is, as everybody knows, the following, which again has equivalents in many other languages: Marie acheta le livre à Jean (cf. German: Marie hat ihm das Buch abgekauft, where Anderson's ab­ lative automatically turns into a dative). Or, still worse: Consider the English sentence: Mary borrowed the book from

John

In this case Anderson would probably again intuitively speak about an ablative (from John), The Chinese equivalent shows us something ra­ ther shocking: Max-liz hsiangz Yue-hanz chiez-le cheiz-penx shu A word by word translation gives us this: "Mary in the direction of John borrowed the book". Anderson's ablative is transformed into some­ thing that is very similar to an allative. I mention this just in order to cool down all the minor prophets of my generation. Universal case theo­ ries can be developed only on the basis of a huge empirical material drawn from languages with a large number of cases. Louis Hjelmslev was aware of this prerequisite. Not long ago I myself had the unexpected opportunity to show what can be put together by means of a so-called 'conceptual framework'. In my booklet "Fokus på syntaksen" (Danielsen 1974b:62) I have demon­ strated how an innocent instrumentalis in Sanskrit is reduced to a dati-

28

PAPERS IN THEORETICAL LINGUISTICS

ve, just to support some subjective assumption. In science we should be extremely carefull with subjective assumptions because the reader in all critical moments is apt to trust the author. The penultimate universal in my list is the universal of the criti­ cal field of distribution. I have dealt with this most important uni­ versal before (Danielsen 1976c:147-156), but it has very often been mis­ understood. The critical field of distribution of any sentence is exactly the field in which a subject (maybe two subjects) in relation to a pos­ sible verbal protone (vP) is stabilized. The term 'verbal protone' indi­ cates the unit which is normally known as the finite verb. But when we speak of a verbal protone of a sentence, we know what part of the sentence under discussion we are referring to. The verbal protone is the verbal part (i.e., the finite verb) of a critical field of distribution. Let us consider a contrastive pair of sentences: Gheobhaidh tú freagra, is dócha (Irish) Du wirst wahrscheinlich eine Antwort bekommen

(German)

They represent two languages none of which yields to the other in re­ spect of whatever value, including especially linguistic value. It ought to be superfluous to say such things, but in times where peculiar pragmatic-ethnic statistics play their tricks on so many linguistis it is not. Both sentences, one German and one Irish, have exactly the same meaning. As a contrastive pair they are ideal, the German sentence being the best and the most correct translation of the Irish counterpart, and vice versa: The Irish sentence is the only adequate counterpart of the German one. The Irish sentence begins with the finite verb. Most Irish sentences do that. The Irish finite verb corresponds to the full verb of the German construction: gheobhaidh = [wirst ... bekommen]. In other words: The Irish finite verb gheobhaidh corresponds to a German nucleus (wirst) plus an operative verb (bekommen). In the Irish sentence gheobhaidh is the verbal protone. In the German sentence only wirst ist the verbal protone. The German sentence, however, does not begin with the finite verb. Most German sentences do not start with the fini­ te verb if they are neutral enuntiative. The initial generative operation S → NP + VP is of no interest because it represents what would in the theory of knowledge be classed as a tautology. And it is of no interest because it forms a theory about nothing, due to the fact that it undeservedly favours the syntax of English which is regarded as the master syntax.

ON THE UNIVERSALITY OF LANGUAGE

29

Let us now consider another contrastive pair of sentences: Cor ne darvāze ko torā (Urdu) Der Dieb hat die Tür aufgebrochen

(German)

The Urdu sentence is a Q-relation. This implies that only the sentence verb (or the finite part of it) stands in the critical field of distribution, because a Q-relation, as is well known, contains no subject at all. The Urdu sentence has an agentive (= epistasis 2) in the form of a repre­ sentation (cor ne "by the thief'), and a scopos (darvāze ko) which also appears in the form of a representation, and a fully neutral sentence verb torā ("opened" / "was opened"). By all this the generative initial operation which I have critized above is proved to be totally false. We have no choice. We must list all the critical fields of distribution that we find in human language. The result is amazing (Fig. 1, p. 30). The critical fields of distribution appear in three rows, and so we have ß-sentences, y-sentences and 8-sentences. The sign 0 stands for 'zero', S for 'subject', vP for 'verbal protone', P for 'predicative', DIE for 'demonstrative indication of existence' (v. Latin ecce, French voilà, voici, Russian ,  etc., v. supra) and $ for an 'unmaterialized subject'. The paradigm shown above needs to be explained in detail which I have already done before (Danielsen 1974b; also 1976c). It is very important for me to emphasize the fact that human lan­ guage possesses a limitid number of critical fields of distribution and that we can easily count them. With special joy the linguist recognizes fully the existence of the ysentences. yY-sentences are found in all languages: Hierher mit der Kanone! This probably is true also of the yZ-sentences: Gleiche Arbeit gleicher

Lohn

γX-sentences, however, hardly occur in German, but surely in many other languages: É agradável passarmos a noite juntos (Portuguese "it is nice that we spend the night together")

30

PAPERS IN THEORETICAL LINGUISTICS

* This group of CFDs is to be considered as predicative explicitations of the CFDs Y:δZ and X:δZ.

Fig. 1: Critical field of distribution

ON THE UNIVERSALITY OF LANGUAGE

31

Sikerült elmenekülniük (Hungarian "they succeeded in escaping") Oldu olacak (Turkish "the inevitable has happened") Let us consider a few more German examples of CFD-distributions:

ßX:

0

S

0

Soweit der Überblick

ßY:

S vP

S

Es wurde polnisches Bier getrunken

SZ:

0

0

Mich dürstet Ihm wurde von seiner Schwester geholfen Mir ist unwohl

ßY:

P

vP

S

Phantastisch dieser Torschuß!

For me it is of special importance to point out that the CFD-manifestations 8Z and ßX are of course fully equal. Some languages will prefer to use one of them more often than the other one, whereas other lan­ guages are rather indifferent as to the choice between the two that are related to each other like a pair of diametrical opposites. One implies the other:

ßZ:

Er kommt jeden Tag zu uns Tāv atamänau nadīm upeyatuh (Sanskrit "strolling around they both came to the river") Nuliara narqup iluanĩpoq (Eskimo "my wife is in the cellar")

8X:

Bleibt noch der Zweifel Udeti savitä (Sanskrit "the sun rises") Vann kann sin heimaverk af skyldurækni (Icelandic) Tháinig sé chúinn (Irish "he came to us") Ist das ein Mädchen!

I feel this is the moment where I must present to the reader a little Finnish verse:

32

PAPERS IN THEORETICAL LINGUISTICS

Tuli tuli tuli tuuli tuuli tuuli tuli sammui

"came the fire" "came the wind" "the wind blew" "the fire went out"

The second line is clearly seen to be a 5X-sentence. The fourth line is a ßZ-sentence. But what then are the first and the third line respec­ tively? This question can be answered unambiguously only by the man whose mind first thought out this poem. And as long as this is the case, we will remain doubtful about a fruitful generativist solution to the analysis of sentences. The last universal to be mentioned in this fundamental description is the universal of the formal garment of sentences. Here we have to do with sentence semantics and its formal materialization. We will have to recapitulate what we have already said in other publications (Danielsen 1968a,b; 1972a, 1974a, 1975, and 1976c) about this highly interesting and far-reaching topic. A language has a limited set of status forms. It also has a limited set of constitutional (or semaphoric) sentence types. Finally, most languages also have a limited set of prolepsis constructions. Status forms are sentences without a sentence marker or with a (one-dimensional) sentence marker (which is also called a status mar­ ker): 1. 2. 3. 4. 5. 6. 7.

Er kommt Kommt er? Ja / Nein Komm, Peter! Möge er kommen! Da komme (, was da wolle!) a. b. . d.

daß er kommt ob er kommt daß er komme! ob er auch kommt obwohl er kommt obgleich er kommt obzwar er kommt obschon er kommt

neutral enuntiative status interrogative status responsive status affective status exoptative status licitive status voluntative status neutral subordinate status interrogative subordinate status affective subordinate status concessive subordinate status

ON THE UNIVERSALITY OF LANGUAGE

e. ob er kommt ob er nicht kommt kommt er (, dann ...) f. kommt er auch noch so oft

33

conditional subordinate status

potential subordinate status

Other languages show some additional variants of subordinate status forms. The status forms seem to be manifestations of a universal regu­ larity which in the different languages is observed with great accuracy. In many languages each status form has its own marker (phonematicly realized status signals like specific words or specific morphemes). In many languages the sign of the recusative polarization seeks out the place of the status marker, and this brings status/polarization markers into existence: Te rogo ne venias Procedamus

/ ne omnes

occidamur

This will sometimes lead to a shift of polarization: Timeo I ne veniat Such a shift, however, will sometimes lead to a restoration of the status sign: Je crains I qu'il ne vienne And a new and clear polarity is established: Je crains I qu'il ne vienne pas Semaphonic sentences carry a sentence semaphor, i.e., a sememe (an A, a relative or a conjunctional word): Wann kommt er? Du kannst kommen, wann du willst Wenn er kommt, werde ich froh Wer so was sagt, ist dumm Der die braune Hose trägt, ist mein Bruder Er liebt den Wein, der aus dem Rhonegebiet

kommt

34

PAPERS IN THEORETICAL LINGUISTICS

Prolepsis sentences are subordinate and consequently are not found in languages that are void of sentence subordination. They are introduced by a prolepsis followed by a status or a semaphoric sentence: 1. Er sieht so aus, als ob er lebensmüde wäre 2. Tu wie wenn du zu Hause wärest! Man speaks in sentences. The human brain is constantly seeking for new transversal relations that will make it possible for a man to for­ mulate and to express his own world and the world around him in a clear and coherent way of thinking and so communicate with other people. By now we should be able to define human language. A human lan­ guage is the nomic combination of a limited number of critical fields of distribution with a limited number of realizations of three simple transversal relations. This operation obeys the univer­ sal laws of the formal garment of sentences. The materialization takes place according to the empirically registered phonological rules. Constitutional Linguistics is facing the task of studying human lan­ guage on all levels stretching from units of the fullest extent to units of the very smallest order. The most superior units are the universals. The very highest universal is the universal of human language. Universals are as numerous as the stars in the sky. We have only mentioned some of the most important linguistic galaxies. There are more of them. Categories such as diathesis, tense, aspect, the predica­ tive sentence (which is also a transversal relation), the adjective, per­ son, gender etc. can and must be studied and explained from a univer­ sal standpoint. Too often we are told that human language possess no universals. It is a kind of tragic irony that this opinion serves as a private bul­ wark against the numerous fashionable ideas of today's linguistics. Tho­ se who cannot see universals cannot see the wood for the trees. They should go and find them. There are lots of them.

ON THE UNIVERSALITY OF LANGUAGE

35

REFERENCES Anderson, John M. 1971. The Grammar of Case: Towards a localistic theory. Cambridge Univ. Press. (2nd ed., 1973). Anderson, Olov Bertil. 1975. "Ett utvecklat psyke ger effektivare språk". Forskning och framsteg 4.24-29. (Stockholm). Bowen, John T. & T. J. Rhys Jones. 1960. Teach Yourself Welsh. London: English Universities Press. Chomsky, Noam. 1970. "Linguistic Theory". Readings in Applied Transformational Grammar ed. by Mark Lester, 52-60. New York: Holt, Rinehart & Winston. Chomsky, Noam. 1972. Language and Mind. Enlarged edition. New York: Harcourt Brace Jovanovich. Danielsen, Niels. 1968a, 1968b, 1972a, 1974a, 1974b, 1975, 1976c. See his List of Publications (pp.viii-x) above, for details. Fillmore, Charles J. 1968. "The Case for Case". Universals in Linguistic Theory ed. by Emmon Bach & Robert T. Harms. New York: Holt, Rinehart and Winston. Fowler, Roger. 1971. An Introduction to Transformational Syntax. London: Routledge & Kegan Paul. der modernen Linguistik. Gardner, Thomas. 1973. Hauptströmungen Göttingen: Vandenhoeck & Ruprecht. Giuliani, Maria Vittoria. 1969. "Grammatica trasformazionale e gram­ matica correlazionale". La sintassi: Atti del III convegno internazionale di studi, 113-124. Rome: Bulzoni. Hovdhaugen, Even. 1971. Transformasjonell generativ grammatikk. 2nd rev. ed., Oslo: Universitetsforlaget. Popper, Karl R. 1959. The Logic of Scientific Discovery. London: Hutchinson. Spore, Palle. 1975. Italiensk grammatik. Odense: Universitetsforlaget. Strawson, Peter F. 1971. Individuals: An essay in descriptive metaphy­ sics, 3rd ed., London: Oxford Univ. Press.

DIE RELATIVA IM NEUHOCHDEUTSCHENUND ANDEREN SPRACHEN Eine Vorstudie zu einer konstitutionellen Sprachtheorie Das Neuhochdeutsche weist die folgende Konstitutionsachse auf (unter Konsti­ tutionsachse verstehen wir das jeweilige Inventar von spezifischen pronominalen, verbalen und adverbialen Interrogativa bzw. Exklamativa einer gegebenen Sprache 1 :

idiogene

WER

Personenzone

WAS

Spezieszone

WELCH/-er, -e, -es (Sg.) /-e, -e, -e (Pl.)

Designationszone

WELCHERLEI

Qualitätszone

WANN

Tempuszone

WOHIN (WO hin(-))

Zielzone

WO

Locuszone

WOHER (WO her(-))

Herkunftzone

WELCHERSEITS

Alternative Seitenzone

WIE

Moduszone

Das Neuhochdeutsche besitzt wie die meisten indoeuropäischen Sprachen keine verbalen Konstitutionsachsenmitglieder. (Als Beispiel für ein solches Element mag hier das frühmittelwalisische PIEU ( = „wessen ist") angeführt werden 2 .) Die echten Konstitutionsachsenmitglieder wollen wir A (großes Alpha) nennen. Neben ihnen gibt es in der modernen deutschen Sprache eine Menge a (kleine Alpha), die sich alle nach einem ganz bestimmten Vorgang realisieren :

1 Vgl. Niels DANIELSEN, Status und Polarität im Gotischen — im Lichte des Kymrischen dargestellt, Odense University Studies in Linguistics vol. 2, Odense 1968. 2 Vgl. J. Morris JONES, A Welsh Grammar (Historical and Comparative), Oxford 1913; John STRACHAN, An Introduction to Early Welsh, Manchester 1909.

38

PAPERS IN THEORETICAL LINGUISTICS

proportionale

WIESO

Kausalitätszone

WARUM WESSENTHALBEN WESHALB WESSENTWEGEN WESWEGEN

Causazone

WESSENUNGEACHTET

Konzessionszone

(WESBEZÜGLICH)

Bezugszone

WIEVIELE

Zone der dissipierten Quan­ tität

WIEVIEL

Zone der kompakten Quantität

INWIEFERN INWIEWEIT

Ausmaßzone

Unsere Formel ist folgendermaßen zu lesen: Ein a ist jede Größe mit dem Inhalt Ind(ikator) + Aspec (= „was?"), die sich im Ausdruck als Aloc + Indikator mani­ festiert. Die neuhochdeutschen a sind demnach folgende Elemente : WObei? WOdurch? WOfür? WOgegen ? WOmit? WOnach? WOvon? WOzu? WOzwischen ?

WOr/an? Wor/auf? WOr/aus? Wor/in ? WOr/über? WOr/um?

Die a wollen wir im Gegensatz zu den echten A als unechte Konstitutionsachsen­ mitglieder bezeichnen. Das A1-rhm der neuhochdeutschen a tritt in der Form zweier Allomorphe auf: WO/- vor Konsonanten, WOr/- vor Vokal. Die Elemente WOHIN und WOHER müssen demnach als A betrachtet werden. Ihre Inhaltsstruktur entspricht nicht ihrer Ausdrucksstruktur in der von unserer Formel geforderten Weise. Wir sind uns dessen völlig bewußt, daß die eingangs registrierten neuhochdeutschen A von unterschiedlichem usuellem Wert sind. WESSENTWEGEN und WESSENTHALBEN sind nunmehr nur in spaßhaftem Gebrauch oder in schriftlichen Zeugnissen altertümlich-schwerfälligen Charakters zu registrieren. WELCHERLEI wird fast nur in der Lateinschule verwendet. WESSENUNGEACHTET ist nur in einer Konversation nach folgendem Muster denkbar : A: „

". „

". „

". „Dessenungeachtet fahre ich morgen nach Ludwigshafen."

 (wacht auf) : „Wessenungeaehtet ? " WESBEZÜGLICH ist möglich, aber nicht gebräuchlich in üblichen Sätzen. Als selektive Alternative zu WELCH/- tritt WAS für ein/- auf. Die Kasuskategorie wird als einer dritten Dimension angebörig angeschaut. Die Akkusativform WEN, die Genitivform WESSEN und die Dativform WEM werden 4*

DIE RELATIVA IM NEUHOCHDEUTSCHEN

39

somit als erster Stock bzw. zweiter Stock bzw. dritter Stock zum Erdgeschoß WER betrachtet. Es gibt Sprachen, die prägnante Konstitutionsachsenmitglieder haben, welche den deutschen WESSEN und WEM entsprechen: poln. CZYJ russ.  (Adj. = „wessen ). Auf ähnliche Weise sind die Kasusformen des neuhochdeutschen WELCH/- in einer dritten Dimension anzuschauen. Dasselbe gilt für WIEVIEL/-. WAS hat nur eine spezifische Kasusform: WESSEN (im zweiten Stock). Jetzt wären wir so weit, daß wir eine konstitutionelle Darstellung der neuhoch­ deutschen Relativa geben könnten. Die Relativa sind in jedem konstitutionellen System diejenigen Elemente, die sich der Konstitutionsachse zunächst befinden. Sie kommen sowohl auf der identifikativen Seite als auch auf der dispositionalen Seite der sprachlichen Inhaltsstruktur vor. Als relative Identifikativa treten im Deutschen drei Reihen auf: 1. die einzelsemigen Relativa, 2. die doppelsemigen Relativa und 3. die zentrisch vorweggreifenden Relativa. Hinzu kommt noch das subzentrische Identifikativum. Auf der dispositionalen Seite der durch die Konstitutionsachse angegebenen Inhaltsstruktur befinden sich die potentiell relativen Dispositionalia (s. Tabelle S. 53). Wir können demnach folgende Analyse der in unserem Thema einschlägigen Elemente durchführen : A1 : Die Mitglieder der Konstitutionsachse haben folgende Struktur :

Beispiele : WER kommt morgen? WAS möchtest du essen ? WAS du nicht alles sagst ! WELCHES Geschäftshaus hat er gekauft ? WELCH ein Ansinnen ! 3 WELCHERLEI Blumen hast du bestellt? WANN kann ich mit Ihrer Antwort rechnen? WOHIN des Weges? WO gehst du hin? WO ist er hingefahren? WO wohnt er? WOHER hast du die Bücher bekommen? WO kommst du denn her? WIE hat sie ihm geantwortet? WIESO hast du kein Geld? WARUM hast du vorher nicht angerufen? WARUM nicht? (vgl. lat. QUIN oder QUIDNI). Dessenungeachtet bin ich nach Berlin gefahren. - WESSENUNGEACHTET? *WESBEZÜGLICH hat er diese Ausführungen gemacht? (kaum verwendbar). WIEVIELE Gläser hast du getrunken? WIEVIEL Bier hast du getrunken? INWIEFERN ist sein Benehmen ungebührlich? Das Konstitutionsachsenmitglied WIE hat zugleich als proportionales Element eine ganz besondere Funktion. Es fragt nach dem Grad von adjektivischen und ad3 Beachte die spezifische begriffliche Funktion des Neutrums WELCHes: WELCHES sind die umstandswörtlichen bezüglichen Fürwörter?, WELCHES sind die Körperteile? WELCHES sind die tiefen Vokale? usw.

40

PAPERS IN THEORETICAL LINGUISTICS

Tabelle 1 Identifikativa

Dispositionalia 





A1 + 2

1

2

3

einzelsemige Relativa

doppelsemige Relativa

zentrische sub(vorweg­ zengreifende) triRelativa sches (bzw. im­ Idenplizierende tifiIdentifi- kativum kativa)

der/die (welcher, -e)

wer; der/die

wer

das; was (welches) welcher

was

was

0

welcher

WELCH/-

welcher auch immer

0

0

welcherlei

WELCHERLEI

welcherlei auch immer

wenn als

wo (da)

(wenn, als)

wann

WANN

wann auch immer

wohin wo . . . hin

wohin wohin wo . . . hin wo . . . hin

wohin

WOHIN (WO . . . hin)

wohin auch immer (wo auch immer . . . hin)

0

il

"  3 *

was

4

was



potentiell relative Dispositionalia

WER

wer auch immer

WAS

was auch immer

wo

wo

wo

wo

WO

wo auch immer

woher wo . . . her

woher wo . . . her

woher wo . . . her

woher

WOHER (WO . . . her)

woher auch immer (wo auch immer . . . her)

wie

wie

wie

wie

WIE

wie auch immer

0

0

0

WIESO

0

warum

(weil)

warum weshalb

WARUM WESHALB

warum auch immer weshalb auch immer

0

0

0

WESSENUNGEACHTET

0

wesbezüglich

0

(wesbezüglich)

(WESBEZÜGLICH)

(wesbezüglich auch immer)

wie

wie

wieviele

wieviele

WIEVIELE

wieviele auch immer

wie

wie

wieviel

wieviel

WIEVIEL

wieviel auch immer

0

0

0

INWIEFERN INWIEWEIT (QUOAD) W I E lange W I E oft a

0

weil

\vesbezüglich

fît

(als) (sobald) (solange) (solange) (sooft) (als) (sooft) aIB2

wie lange wie oft aIB3

wie lange auch immer wie oft auch immer aAl + 2

DIE RELATIVA IM NEUHOCHDEUTSCHEN

41

verbialen Lexeminhalten im Positiv (vgl. engl. HOW, dän. HVOR u. ä.), nicht zuletzt nach solchen, die zonenangebend sind: WIE WIE WIE WIE WIE WIE

groß ist das Haus? klein du bist! lange bleibt er hier? weit wirst du gehen? oft soll ich es dir noch sagen? alt ist er?

(vgl. lat. (vgl. lat. (vgl. lat. (vgl. lat. (vgl lat. (WIE =

QUANTUM) QUANTULUS) QUAMDIU) QUATENUS) QUOTIENS) lat. QUAM)

Das proportionale WIE kann sich auch auf finite Inhalte beziehen : WIE ich dich liebe ! Ein durch um indiziertes WIEVIEL fragt nach dem komparativen Grade : Um WIEVIEL größer ist er in der Zwischenzeit geworden ? ,A 1.1 : Die untergeordneten Konstitutionsachsenmitglieder haben die folgende Struktur:

Was meinst du, WER es war? Ich weiß nicht, WER es ist. Ich weiß nicht WER. Ich weiß nicht, WAS er sagt. Ich weiß nicht WAS. Ich weiß nicht, WELCHES Buch das ist. Ich weiß nicht WELCHES. Ich weiß nicht, WELCHERLEI Blumen er gekauft hat. Ich weiß nicht, WANN er kommt. Ich weiß nicht WANN. Ich weiß nicht, WOHIN er fährt. Ich weiß nicht WOHIN. Ich weiß nicht, WO er wohnt. Ich weiß nicht WO. Ich weiß nicht, WOHER er das Buch bekommen hat. Ich weiß nicht WOHER. Ich weiß nicht, WIE er das macht. Ich weiß nicht WIE. Ich weiß nicht, WARUM er schimpft. Ich weiß nicht WARUM. Ich weiß nicht, WESBEZÜGLICH er diese Ausführungen gemacht hat. Ich weiß nicht WIEVIELE Personen noch da sind. Ich weiß nicht WIEVIELE. Ich weiß nicht, WIEVIEL Bier er trinkt. Ich weiß nicht WIEVIEL. Ich weiß nicht, INWIEFERN sein Benehmen als unzulässig bezeichnet werden darf. Ich weiß nicht, WIE groß er ist. Ich weiß nicht WIE groß. Ich weiß nicht, WIE lange das dauert. Ich weiß nicht WIE lange. Ich weiß nicht, WIE weit er gehen wird. Ich weiß nicht WIE weit. Ich weiß nicht, WIE oft er zu ihr kommt. Ich weiß nicht WIE oft. Weißt du, WIE ich dich liebe? Weißt du, um WIEVIEL größer er in der Zwischenzeit geworden ist ? Was hat sie gesagt, WO sie hingeht? Wir lassen uns nicht sagen, WO wir WANN WAS zu tun haben. Die in der rechten Reihe am Ende der Sätze stehenden Konstitutionsachsenmitglieder sind völlig unbetont. Sie signalisieren einen jeweilig zu ergänzenden untergeordneten Satzinhalt. Wo sie in dieser Stellung betont auftreten, wollen wir sie als zentrische Relativa auffassen (siehe unten).

42

PAPERS IN THEORETICAL LINGUISTICS

II. IA (Identifikativum A): das subzentrische Identifikativum Struktur:

hat die folgende

Peter, weißt du was? Jetzt bin ich das glücklichste Mädchen auf der Erde! III. I B 3 : die zentrischen Relativa haben die folgende Struktur:

Die zentrischen Relativa sind immer stark betont. Sie implizieren einen relativen Inhalt, der sich auf den existentiellen Inhalt des jeweiligen Zonennenners bezieht: Irgend jemand hat es gesagt. Ich weiß aber nicht wer. ( , wer es war, der es gesagt hat). Irgend etwas hat nicht geklappt. Ich weiß aber nicht was. Irgendein Unternehmen ist geplatzt. Ich weiß aber nicht welches. Irgendwann muß er dort gewesen sein. Ich weiß aber nicht wann. Wann er gekommen ist, ist mir völlig egal. Irgendwo muß das Buch liegen. Ich weiß aber nicht wo. Wer du bist, und wo du wohnst, (das) ist uns einerlei. Woher du kommst, und wohin du fährst, (das) geht mich nichts an. Irgendwoher hat er das Buch. Ich weiß aber nicht woher. Irgendwohin muß er das Geld geschickt haben. Ich weiß aber nicht wohin. AUS irgendeinem merkwürdigen Grund hat er das getan. Ich weiß aber nicht warum. Warum du es getan hast, brauchst du mir nicht zu sagen. Er hat viel Bier getrunken. Wieviel weiß ich aber nicht. Es hat lange gedauert. Wie lange weiß ich aber nicht. Er kommt freilich sehr oft. Wie oft weiß ich aber nicht. Viele traditionelle Grammatiker wollen es sicher nicht gutheißen, daß wir diese zen­ trischen Elemente als Relativa betrachten. Für unsere synchrone Darstellung der neuhochdeutschen Relativa sind aber gerade diese Elemente von allergrößtem Belang. In dieser Reihe von Identifikativa können wir beobachten, wie sich die in vielen Sprachen mit den A (a) formal völlig identischen Relativa auf den Weg machen, um sich als Relativa von jenen loszureißen.4 Formal fällt diese Reihe im großen ganzen mit der Konstitutionsachse zusammen (WIESO, WESSENUNGEACHTET und INWIEFERN (INWIEWEIT) lassen sich jedoch nicht in dieser IB 3 -Reihe realisieren, was mit ihrer semantischen Prägnanz zusammenhängt). IV. IB 2 : die doppelsemigen Relativa haben die folgende Struktur :

Von einem universellen Gesichtspunkt aus lassen sich die doppelsemigen Relativa in drei Gruppen einteilen: 4 In einigen Sprachen können gewisse A in untergeordneten Sätzen geradezu einen demonstrativen Charakter annehmen (vgl. ungarisch: A kalauznő csodálkozik/hogy a kül földiek már miyen jól beszélnek magyarul).

DIE RELATIVA IM NEUHOCHDEUTSCHEN A.

43

I. Das relative O1 (Akkusativ-Rectum des Relativsatzes) tut grammatisch dem ihm innewohnenden Demonstrativelement Genüge : a) Diese Konstruktion weist eine syntaktische Situation auf, in der ein nach dem Beziehungswort wiederholtes Demonstrativelement (bzw. Artikel) als hinweisende Komponente des doppelsemigen Relative hinzuzudenken ist. Ein derartiges wiederholtes Demonstrativum ist uns aus attributiven Genitivkon­ struktionen bekannt: b) II. Das relative O1 wird in den für das latente Demonstrativelement adäquaten Kasus gesetzt, wobei das dabei stehende Substantiv unter Weglassung des Artikels in den Relativsatz gezogen wird:

B.

Das dem doppelsemigen Relativum innewohnende Demonstrativelement wird im Ausdruck gänzlich unterdrückt (grammatisch und syntaktisch); es kann hierbei wiederaufgegriffen werden : Wem Gott will rechte Gunst erweisen, den schickt er in die weite Welt. Wer so was behauptet, muß unzurechnungsfähig sein.  Das dem zweisemigen Relativum innewohnende Relativelement wird im Aus­ druck gänzlich unterdrückt : Nu du er blevet voksen, må du sandelig prove at tage dig sammen. (Dänisch). Du får inte föra mig dit jag inte vill. (Schwedisch). Im Neuhochdeutschen ist nur Gruppe  vertreten : Wer so was glaubt, ist dumm. (generell) Der es gesagt hat, bin ich. (speziell) Wer lügt, stiehlt auch. Die es bemerkt hat, ist seine Tochter. Du bist nicht, der du scheinst. Die hier unten gedient, sind dort oben groß. Die ich suche, trägt den Schleier. Ein Schelm ist, der seinen Herrn verläßt. Ehre, dem Ehre gebührt. Wer aber hereinkam, das war der Wolf. Wes Brot du ißt, des Lied du singst. Was du da behauptest, ist unrichtig. Wen du anrufst, den mußt du auch höflich anreden. Wessen Sie bedürfen, das werden Sie jederzeit bei mir bekommen. Wenn sie schläft, schläft er auch. Als der Zug eintraf, fing das Orchester an zu spielen. Wo er wohnt, dort möchte ich auch wohnen. Wo du hingehst, da gehe ich auch hin. Wo das gute Bier herkommt, da kommt auch die gute Wurst her. Wie dein Vater gelebt hat, so sollst du auch leben. Wieviel du zu essen haben willst, soviel bekommst du.

44

PAPERS IN THEORETICAL LINGUISTICS

Solange du dich ruhig verhältst, wirst du kein böses Wort von mir hören. Sobald du dich umkehrst, hast du mich vergessen. Sooft er kam, ging sie zur Hintertür hinaus. V. IB1: die einzelsemigen Relativa haben die folgende Struktur:

Ich kenne Herrn Meyer, der gestern bei dir war. Ich kenne Frau Schulze, die in deiner Firma arbeitet. Ich kenne den Jungen, mit dem du gespielt hast. Ich kenne die Frau, mit der er verheiratet ist. Ich kenne das Mädchen, das gestern im Wirtshaus saß. (War es ein Traum, was ich erlebte?) (Es ist nicht Furcht, was mich bewegt.) Gott vergelte ihm das Gute, was er an mir getan. Das wenige Geld, was ich besaß, hat er mir genommen. Ein Baum, welcher keine Früchte trägt, ist wertlos. Er sagte guten Tag, welchen Gruß sie freundlich erwiderte. Du bist eine janz dußlige Rotzneese, wo nich in de Zeit paßt! (dialektal) Zu der Zeit, wo ich in Magdeburg wohnte, kamen sie öfters zu Besuch. Im Augenblick, da Kriemhild scheinbar Siegfried untreu wird, hebt der Dichter ihre wahre Gesinnung hervor. In der Stadt, wo du wohnst, gibt es viele Polen. Aus der Hölle, woher er kam, hörte man noch die Bomben fallen. An dem Ort, wohin du läufst, ist keiner. Ich mag nicht die Art, wie er dahergeht. Ich kenne nicht den Grund, warum er weggefahren ist. Er fand keine Veranlassung, weshalb er sie verachten sollte. Das Problem, wesbezüglich er sich geäußert hat, ist ein sehr aktuelles. Du brauchst mir nicht soviel Geld zu geben, wie du mir versprochen hast. Soviele Menschen., wie wir gestern gesehen haben, habe ich nie zuvor in meinem Leben gesehen. Die ganze Zeit, solange du hier warst, habe ich mich gelangweilt. Die vielen Augenblicke, sooft ich daran denke, haben mich an den Rand der Verzweiflung gebracht. IV. IC 3 : Relative Konjunktionalia:

Was die Stellung des Adverbs anbetrifft, so ist sie einigermaßen frei. Damais, als sie noch jung war, war sie furchtbar unerfahren. Er schlug ihn nur (dann), wenn er ungehorsam war. Ich fand ihn (dort), wo ich ihn am wenigsten vermutet hätte. Stelle das Buch dahin, wohin es gehört ! Sie gingen schweigend (dorthin), wohin ihre Pflicht sie rief. Die Hilfe kam (dorther), woher man sie nicht erwartet hatte. Von dort, woher er kam, hörte man die Bomben fallen.

DIE RELATIVA IM NEUHOCHDEUTSCHEN

45

Dort, wohin du läufst, ist keiner. Ich glaube ihm deshalb nicht, weil er mich schon des öfteren belogen hat. Bist du deshalb aufgestanden, damit du ihn vor Tagesanbruch findest ? Sie war ihm gerade deswegen zuwider, weil sie ihn früher betrogen hatte. Er macht seine Arbeit schlecht und recht, so wie er es versteht. Du brauchst mir nicht soviel zu danken, wie du es früher getan hast. Ich bleibe heute nicht so lange, wie ich das letzte Mal geblieben bin. Er kommt nicht mehr so oft, wie es früher der Fall war? Ist er ebenso klein, wie es früher der Fall war? Er fährt ebenso schnell, wie sein Hund läuft. So oft als es regnet, hat sie Kopfschmerzen. Sobald als die Unglücksbotschaft eingetroffen war, setzten sich die ersten Hilfsmannschaften in Bewegung. Wir werden das nur solange anerkennen, solange es uns notwendig erscheint. Ich werde das sooft sagen, sooft es notwendig ist. Die Lösung dieses Problems ist umso wichtiger, als von ihr die Zukunft meiner Familie abhängt. VII. IC 2 : Neutrale Konjunktionalia:

Wenn die Blätter fallen, werden die Tage kürzer. Als ich ihn kürzlich besuchte, war er schon krank. Wie ich eine Weile so dasitze, hüpft ein Eichhörnchen auf meinen Schoß. Während sie ihre Einkäufe machte, wartete er in einem Café auf sie. Seit er seinen Beruf aufgegeben hat, geht es abwärts mit ihm. Wenn er sich etwas erholt hat, ist er wieder arbeitsfähig. Als er sein Unrecht eingesehen hatte, entschuldigte er sich. Ehe du dich umdrehst, hat er dich schon betrogen. Bevor ich ausgesprochen hatte, war er schon mit seiner Antwort bei der Hand. Bis du zurückkommst, werde ich mit der Arbeit fertig sein. Ich gehe nicht zu Bett, ehe ich (nicht) mit dieser Arbeit fertig bin. Kaum ließen sich die Löwen blicken, als ihnen auch schon ein Kugelregen entgegengesandt wurde. Er bestand sein Examen nicht, weil er nicht vorbereitet war. Ich gehe nur aus, falls du damit einverstanden bist. Wenn du gekommen wärest, hätte ich dich freundlich aufgenommen. Wenn auch Berge und Täler, Ströme und Meere uns trennen, so werde ich dir dennoch ein treues Gedenken bewahren. Er wird mehr Erfolg haben, als sein Bruder gehabt hat. Meine wirklichen Gründe sind andere, als du glaubst. Du hast mich ebensoviel geschädigt, als ich dir genützt habe. Tu, wie wenn du zu Hause wärest. Sie tat, als hätte sie nichts gehört. Ich kann ihn nicht verstehen, zumal er Irisch spricht. Je mehr er hat, je mehr er will.

46

PAPERS IN THEORETICAL LINGUISTICS

Je höher der Turm, desto höher der Fall.} um so höher der Fall. VIII. IC1 : Demonstrative Konjunktionalia:

Da ich ihn sah, fiel mir mein Versprechen wieder ein. Indem er sich bückte, fiel ihm seine Brieftasche aus dem Rock. Sowie ich seine Anschrift erfahre, teile ich sie dir mit. Indes die Eltern sich im Zimmer unterhielten, spielten die Kinder im Freien. Indessen die Kinder im Garten spielten, unterhielten sich die Erwachsenen im Haus. Solange ich mit ihm verkehre, hat er sich als ein guter Kamerad erwiesen. Nachdem ich deinen Brief erhalten hatte, machte ich mich sofort an die Erledigung deiner Aufträge. Sobald ich Einzelheiten erfahre, werde ich euch den Hergang genau schildern. Seitdem die Siegesnachricht bekanntgeworden ist, herrscht bessere Stimmung. (Er machte seine Arbeit schlicht und recht, «so wie er es verstand.) Wir behalten uns vor, unsere Preise abzuändern, je nachdem sich die Weltmarkt­ preise gestalten. Er zeigte sich als wahrer Freund, indem er mich in der Not unterstützte. Wir können die Arbeit in dieser kurzen Frist nicht fertigstellen, da wir Strom­ mangel haben. Sie darbten, damit ihre Kinder anständig erzogen würden. So ihr mich vom ganzen Herzen suchet, so will ich mich finden lassen. Ich freue mich seiner guten Meinung, sofern sie redlich ist. Dieses Buch ist gut, soweit ich es beurteilen kann. Wir können Ihnen die versprochene Lohnerhöhung zugestehen, insoweit Sie das für den Plan abgezweigte Geld richtig zu verwalten verstehen. Herder wünscht, daß ich bloß als Redakteur etwas darüber sagen möchte, insofern auch die Horen mitgetroffen werden sollten. Soviel ich weiß, geht es ihm gut. Er ehrt die Wissenschaft, sofern sie nützt. Füge ihn nicht, «so sehr du ihn auch liebst. So weit mein Auge reicht, sehe ich nur Trümmer. So oft ich Heinz sehe, fallen mir unsere gemeinsamen Jugendstreiche ein. Solange ich mit ihm verkehre, hat er sich als guter Kamerad erwiesen. So dumm er ist, so anmaßend ist er. Je gründlicher Sie sich in die Sache vertiefen, um so größeren Nutzen werden Sie davon haben. XL DA 1 + 2 : (die potentiell subordinierenden Dispositionalia und die potentiell relativen Dispositionalia): die potentiell relativen Dispositionalia (DA 2) fallen im Neuhochdeutschen formal mit den potentiell subordinierenden (DA 1) zusammen:

DIE RELATIVA IM NEUHOCHDEUTSCHEN

47

DA. Wer auch immer anklopft, ich stehe nicht auf. Was auch immer geschehen mag, ich bleibe bei dir. Welcher Mensch auch immer getötet wird, so seid ihr dafür verantwortlich. Welcherlei Pflanzen du auch immer anbaust, du mußt den Boden gut düngen. Wann er auch immer kommen mag, so ist er willkommen. Wann auch immer er kommen mag, so ist er willkommen. Wohin du auch immer gehst, ich folge dir. Wo er auch immer sein mag, meine Gedanken sind bei ihm. Woher auch immer der Wind weht, morgen segeln wir hinaus. Wie er auch immer reagieren mag, mich kümmert das wenig. Warum auch immer er das getan haben mag, es ist mir schleierhaft. Wieviele Gäste du auch einladen magst, ich komme in jedem Fall. Wieviel Bier er auch trinken mag, man merkt es ihm nicht an. DB. Wer auch immer anklopft, dem wird aufgetan. Man schlage eine Grammatik auf, welche man will. Was auch immer geschehen mag, ist mir völlig gleichgültig. Welcher Mensch auch immer getötet wird, der ist nicht vergeblich gestorben. Welcherlei Pflanzen du auch immer anbaust, die mußt du gut begießen. Wann er auch immer kommen mag, dann werden wir zu Hause sein. Wann auch immer er kommen mag, dann werden wir zu Hause sein. Wohin du auch immer gehst, dorthin folge ich dir. Wo er auch immer sein mag, da hat er sicher Erfolg. Woher auch immer der Wind weht, daher kommen die Wolken. Wie auch immer man sich bettet, so liegt man. Warum auch immer e das getan haben mag, darum ist er noch lange nicht zu ver­ achten. Wieviele Bilder Picasso auch immer malt, soviele werden ihm auch abgekauft. Wieviel du auch immer produzierst, soviel wird man dir abnehmen. Im Neuhochdeutschen gibt es demnach folgende Statusformen :5 A. Übergeordnete : B. Untergeordnete : I. Der neutral enuntiative Status: 1. Der neutral subordinierte Status: Er kommt. /daß er kommt II. Der interrogative Status: 2. Der subordinierte interrogative Status: Kommt er? ¡ob er kommt III. Der responsive Status: 3. Der subordinierte affektive Status: Ja./Nein. /daß er sofort komme ! IV. Der affektive Status : 4. Der konditionale Status : Komm ! a) /ob er kommt/ (oder) /ob er nicht kommt b) kommt er/ (dann) . . . V. Der exoptative Status: 5. Der konzessive Status: Möge er kommen ! /obzwar er kommt /obschon er kommt 5 Vgl. Niels DANIELSEN: Status und Polarität im Gotischen — im Lichte des Kymrischen dargestellt, Odense 1968, S. 138: „Ein Status ist ein Satz . . ., dessen syntaktische oder semantische Klassifikation nicht von einem vollwertigen konstitutionellen Element be­ stimmt ist." Ein Status ist ein Satz ohne konstitutionelles Semaphor.

48

PAPERS IN THEORETICAL LINGUISTICS

/obgleich er kommt /obwohl er kommt /ob er auch kommt (-zwar, -schon, -gleich, -wohl, auch sind Modifikatoren) VI. Der lizitive Status : 6. Der potentielle Status : (Da) komme, (was da wolle !) /(und) sei er auch noch so untauglich. Die folgenden subordinierten Satztypen sind als Prolepsisgefüge zu betrachten: Dieser Aufsatz sieht aus, als ¡ob Sie ihn nicht selbst geschrieben hätten. Anstatt / daß er mir Dank gewußt hätte, redete er nur Böses von mir. Ehre deinen Vater und deine Mutter, auf / daß du lange auf der Erde leben mögest ! Es ist zu schön, als / daß es wahr sein könnte ! Er blieb nur bis sechs Uhr, so / daß wir den frühen Abendzug erreichen konnten. Unsere konstitutionelle Analyse hat zu folgendem Ergebnis geführt : 1. Wir müssen im Neuhochdeutschen mit mindestens fünf verschiedenen Reihen von Relativa rechnen: IDENTIFIKATIVA a) einzelsemige b) doppelsemige c) zentrische d) relative Konjunktionalia

DISPOSITIONALIA e) potentiell relative

Diese Einteilung ermöglicht es, die Relativa scharf von den A (sowie vom subzentrischen Identifikativum) abzugrenzen. Die zentrischen Relativa, die in der tra­ ditionellen Grammatik (wenn sie überhaupt erörtert werden) bald als fragend, bald als relativ gewertet werden, finden hiermit ihren eindeutigen Platz in der inneren Struktur der Sprache. 2. Bei einer völlig erschöpfenden Analyse der neuhochdeutschen A und Relativa kristallisiert sich zusehends unabweislich ein prägnantes subzentrisches Identifika­ tivum heraus, dessen Eigenwertigkeit in früheren Darstellungen kaum beachtet wurde. 3. Die jeweilige Produktivität der A innerhalb des konstitutionellen Inhaltssystems und deren Manifestationen im Ausdruck lassen sich in bezug auf die Relativa direkt und unzweideutig ablesen. Eine weitere Analyse in bezug auf das totale Inhaltssystem erfordert eine weitere Einteilung in Reihen nach links (Identifikativa) und rechts (Dispositionalia). 4. Der etwaige Konflux zwischen Einzelfeldern einer Zone und entsprechenden Feldern anderer Zonen läßt sich ebenso deutlich und semantisch aufschlußreich ablesen. (In der Reihe der einzelsemigen Relativa (IB1) dringt zum Beispiel das Element wo der Locuszone in immer weiterem Ausmaße in die Tempuszone. Es kann sogar in die Spezieszone dringen.)* 5. Der etwaige Konflux zwischen Einzelfeldern einer Reihe und entsprechenden Feldern anderer Reihen läßt sich in ähnlicher Weise übersichtlich ablesen. (Dabei ist es von ganz besonderem Belang, die Produktivität der Elemente einer jeweiligen Konstitutionsachse genau zu fassen und zu beschreiben, um auf diesem Wege zu einer * die, wo immer den Ton angeben (AUERBACH).

DIE RELATIVA IM NEUHOCHDEUTSCHEN

49

erschöpfenden komparativen Beschreibung von den basalen Inhaltseinheiten der menschlichen Sprache zu gelangen). 6. Die funktionale Korrelation zwischen den konstitutionellen Einzelelementen las­ sen sich senkrecht und waagerecht eindeutig ablesen. (Das Resultat einer solchen Analyse ist die frappante Tatsache, daß auch Elemente wie wohin (wo . . . hin-), woher (wo . . . her-), wie, warum, (wesbezüglich) nebst wo in gewissen Fällen als un­ umstritten und unwiderlegbar relativ gewertet werden müssen. Diese letzte Schluß­ folgerung scheint manchen herkömmlichen Grammatiker schockieren zu wollen). 7. Unsere dreidimensionale Einteilung der Inhaltselemente scheint als Basis einer erneuten Erforschung der Kasuskategorie dienen zu können. Die Frage muß jetzt lauten : Was ist als Kasus (einer dritten Dimension angehörig) anzusehen, und was ist als konstitutionelle Basis (einer bloß zweiten Dimension) zu werten? Die Antwort ist für jede Sprache verschieden. (Zum Beispiel: deutsch WESSEN, Kasus Genitiv; polnisch CZYJ (=„wessen"), Anzeiger einer Inhärenzzone). 8. Unsere zweidimensionale Einteilung der Inhaltselemente scheint Wichtiges über das Genus der nominalen und pronominalen Elemente sagen zu können. Das Element das (einsemiges relativ, demonstrativ, best. Artikel) gehört primär der Spezieszone an. Die entsprechenden Elemente der und die können demzufolge als binare Alternativen betrachtet werden, die primär der Personenzone angehören (wie es aus der doppelsemigen Reihe ersichtlich ist). Die letzteren dringen aber weitgehend in die Spezieszone ein, wo sie einen rein grammatisch-funktionalen Wert bekommen. 9. Überhaupt lassen sich sämtliche grammatischen Kategorien von unserem zweiund dreidimensionalen System aus in einem ganz neuen Licht betrachten. Inter­ polationen zwischen den einzelnen Zonen, Reihen und Feldern werden aufschlußreiche Informationen über die innere Struktur der Sprache ergeben. Nicht zuletzt die ver­ balen Kategorien werden dabei klarer und treffender behandelt werden können. Die Moduskategorie muß dabei in einer dritten Dimension veranschaulicht werden, die übrigen Kategorien des Verbums lassen sich durch Querinterpolationen bzw. zwei­ dimensional erklären. Im deutschen Inventar von A-Größen ist kein verbaler Zonen­ anzeiger vorzufinden. In anderen Sprachen gibt es mehrere Beispiele für verbale Zonenanzeiger (z. B. Khalkha-Mongolisch JA- ( = „ w a s machen?")). Im Deutschen ist die verbale Zone diejenige, die ausschließlich aus Determinationen besteht. (Jede Zone ist durch eine größere oder kleinere Anzahl von Determinationen gekennzeich­ net, z. B. Mistinguette, . . . . (Personenzone), Mann, Haus, Maus, Laus, Schmaus, Zuckerfabrikdirektor, . . . . (Spezieszone), grau, widerlich, großsprecherisch, mutter­ seelenallein, schön, . . . . (Qualitätszone) usw.). (Die Hauptaufgabe der Phono­ logie wird es demnach sein, die Wurzelplereme gegenüber den morphematischen Pleremen abzugrenzen; die Anzahl von Phonemen, die Flektions- und Konjugationsmorpheme bilden können, ist in einer jeweiligen Sprache begrenzt (im Englischen sind das praktisch nur [-s/-z] und [-t/-d] (father's; cats; looked, loved), im Deutschen sind das s, t/d, r, n, m) usw. für jede Sprache) ; die in einer jeweiligen Sprache vor­ kommenden morphembildenden Phoneme bilden eine Teilmenge des jeweilig zu re­ gistrierenden Phoneminventars). (Die Syntax läßt sich mittels optimaler Gliedketten behandeln (wir stehen hier in scharfem Gegensatz zu dem von Noam CHOMSKY vorgeschlagenen logisch-intuitiven Verfahren) :

50

PAPERS IN THEORETICAL LINGUISTICS

Die Numerierung der Einzelglieder, die nach Rekursivitätsprinzipien einem Min­ destmaß an syntaktischen Teilketten zugeordnet werden, vollzieht sich prinzipiell nach folgendem Schema: Schema : I. Status (-Polarisations)zeichen II.A

ja : nein (1) (y(r) : nid1; 0: nicht7; -di-: 08 . . . ) (2) (WE R?, WAS? WELCH/-?, WANN?, . . .)

I I I . Identifikativa A   D E (3) (der/die, das, derselbe, ich . . .) IV. Dispositionalia A   D E (4) (einige, jemand, alle, kein . . .) (5) (an, hinter, nach, bis, wie . . .) V. Indikatoren (6) (Katze, wild, wieder, steh/-, neun , VI. Determinationen (7) (wohl, mal, zwar, freilich . . .) VII. Modifikatoren (8) (und, aber, oder, denn . . .) VIII. Konjunktionen IX. Kategorien in der dritten (9) (Kasus, Modus) Dimension X. Dependenzen und Interdependenzen (10) (Rektion, Nexus usw.) XL Interpolationen (11) (-est, -e, -te, -en usw.) XII. Intonation (mit Phonologie) (12) Die deutschen Relativa sind demnach identifikative (I)B, bzw. dispositionale (D)A2-Elemente, die vorwiegend als erstes Glied des Relativsatzes diesen als relativ kennzeichnen. Dieser Tatbestand muß in der optimalen syntaktischen Gliedkette in Rechnung gestellt werden. 6

kymrisch: yr wyfi wedi gwneud hynny : nid wyf i wedi gwneud hynny = 7 lateinisch: 8 feci ( = ich tat es) : non feci ( = ich tat es nicht) = kanaresisch: m'adidenu: madenu.

DIE RELATIVA IM NEUHOCHDEUTSCHEN

51

In anderen Sprachen werden Relativsätze oft nach anderen Verfahren gebildet. Im Kymrischen zum Beispiel geschieht es durch satzeinleitende Status-Polarisationszeichen : llyfr / a oedd ar y bwrdd ( = „das Buch, das auf dem Tisch lag") llyfr / nid oedd ar y bwrdd ( = „das Buch, das nicht auf dem Tisch lag") Im Kopenhagener Dialekt („GrØnnegade-Mundart") wird ein Relativsatz durchweg in der Form eines subordinierten Status realisiert, wenn dem'Statuszeichen im Reichs­ dänischen ein relatives Objektspronomen bzw. ein prädikatives Relativpronomen ent­ spricht : manden / at vi så i går / var min fætter ( = „der Mann, den wir gestern sahen, war mein Vetter"). Ist das Relativpronomen (som) im Reichsdänischen Subjekt des Relativsatzes, dann folgt im GrØnnegade-Dialekt auf das Statuszeichen at direkt das verblichene konstitutionelle Element der Locuszone der: Ham manden ¡at der står udenfor og venter/ er min faetter. Im Englischen fungiert als Statuszeichen (nach gegebenen Regeln) das devaluierte konstitutionelle Element that: I know the man /that you mentioned (vgl. I remember / that I saw him). Im Englischen (sowie in vielen anderen Sprachen, so z. B. im Dänischen) kommen unter bestimmten Bedingungen sehr oft merkmallose Relativsätze vor : I remember the man / you mentioned. Jeg kan huske manden / du nævnede. Die untergeordneten Konstruktionen treten jedoch kraft ihres ungesättigten Cha­ rakters deutlich als relativ hervor. Im Dänischen sind Sätze dieser Art in der rekusativen Polarisation immer zugleich durch die besondere Wortstellung als relativ cha­ rakterisiert : Den bog / jeg ikke fik med / ligger på mit skrivebord. Ähnliche dänische Relativsatzkonstruktionen treten überhaupt als deutlich rela­ tiv hervor, wenn der relative Status mit einem Adverbiale versehen ist : Den tale / j e g snart skal holde / koster mig meget hovedbrud. Im Chinesischen werden relative Inhalte auf der Ausdrucksebene attributiv realisiert : jin tian shang wu dao zhe li lai kan ni di nen shi cong li lai di? ( = heute morgen ankommen hier kommen sehen Sie/(Partikel te)/ Mann von wo gekommen?). (= von wo ist der Mann gekommen, der heute morgen gekommen ist, um Sie zu besuchen?). Wenn das Objekt eines vorhergehenden Verbums mit dem Subjekt des relativen Inhalts identisch ist, wird die folgende Konstruktion gebraucht: wo you yi ben hao shu shi zhuo zi shang tou ( = ich habe ein Stück gut Buch ist Tisch auf) ( = ich habe ein gutes Buch, das auf dem Tisch liegt). Die Konstruktion muß als eine Verkettung von zwei neutral enuntiativen Status­ formen angesehen werden. Die Relativität klingt in dieser Verkettung unwill­ kürlich mit. Manche Sprachen haben überhaupt nicht die Möglichkeit, Relativsätze zu bilden, so z. B. das Türkische: Talebeye vermis, oldugum kitap yeni idi.

52

PAPERS IN THEORETICAL LINGUISTICS

( = „Student-(Dat.) gegeb-en worden-sein Buch neu war") = das Buch, das ich dem Studenten gegeben hatte, war neu. Universell gesehen scheint es demnach gerechtfertigt zu sein, mit folgenden Reali­ sationen eines relativen Inhalts zu rechnen : I. Partizipialkonstruktionen. (Beispiel: Türkisch) IL Attributivkonstruktionen. (Beispiel: Chinesisch) III. Statuskonstruktionen 1. mit direkt kommutablen Status-Polarisationszeichen (Beispiel: Kymrisch) 2. mit Statuszeichen a) mit echtem Statuszeichen (dän. (Kopenhagener Dialekt) : manden / at jeg sa i går/ . . .) b) mit devaluiertem konstitutionellem Element als Statusmerkmal (engl. the man/that I saw/ . . .) 3. ohne Statuszeichen (dän. manden / jeg så i går; engl. the man / I saw) IV. Konstitutionelle Relativsatzkonstruktionen. (Beispiel: Latein, Deutsch) Die deutschen Relativsätze gehören alle zur Gruppe IV. Im Deutschen sind keine anderen Relativkonstruktionen möglich.

Anhang Anschließend folgt in übersichtlicher Form ein erster vollständiger Versuch, die basalen Inhaltskategorien (die konstitutionellen Elemente) der menschlichen Sprache zu verkarten. Der Weg zu einer solchen Verkartung wurde schon oben betreten und skizziert. Wir sind der festen Überzeugung, daß der semantische Knoten der Sprach­ wissenschaft ein Scheinproblem ist, um das man sich mit gravitätischem Ernst bewegt wie die Katze um den heißen Brei: Man läuft ungern Gefahr, wegen angeblichen „Sprachphilosophierens" angeprangert zu werden. Ich wage den Sprung in einer Zeit, wo einerseits das Wort Semantik wieder langsam salonfähig wird, und wo anderer­ seits maßgebliche Kreise innerhalb der modernen Linguistik ohnehin von einer be­ dauerlichen philosophistischen Myelitis9 befallen sind, vor der man kaum die Augen niederzuschlagen braucht. 9 Ganz gefährlich wird diese tragische Krankheit, wo sie sich mit verstocktem Struktu­ ralismus vermählt, um ehrliche wissenschaftliche Arbeit zu diffamieren (vgl. die neulich in Oslo entlarvte Machtverzerrimg der sogenannten „Grand-Hotell-Kommission", die im Herbst 1969 der jungen dänischen Sprachwissenschaft das Recht zum Leben aberkannte, s.in: Odense Universitets Årsberetning 1969—70, S. 11—18). Siehe auch die glānzende Rezen­ sion Ernst PULGRAMS von Noam CHOMSKYS Language and Mind in : The Modern Language Journal, Vol. LV, No. 7, November 1971, S. 479: „. . . if universale are things that are possessed by all languages, then they must be sought and analyzed by an examination of all languages (or of as many as linguists can reasonably handle and may consider statistically adequate). It seems therefore that the call for empiricism as expressed above leads to another kind of universale and not to the universal grammar as defined by the transformationists." Und S. 480: „Thus transformational grammar has, especially among its more inept though no less vociferous practitioners, come to talk increasingly about itself rather than about language and languages : the game is the thing, solipsism is the message, and omphaloscopy is the method."

5

Z.Phonetik 1/75

D I E R E L A T I V A IM N E U H O C H D E U T S C H E N

53

Unsere Darstellung vollzieht sich in folgenden drei E t a p p e n : I . Versuch einer Skizzierung eines optimalen I n v e n t a r s von Konstitutionsachsen­ mitgliedern : A. Idiogene Konstitutionsachsenmitglieder : 1. Nichtverbale Mitglieder (S. 54-55) 2. Verbale Mitglieder (S. 56-57) B . Proportionale Konstitutionsachsenmitglieder (S. 67-68) II. Versuch einer Skizzierung der Identifikativa in der menschlichen Sprache (S. 58-59) III. Versuch einer Skizzierung der Dispositionali a in der menschlichen Sprache (S. 59-60) Als A u s g a n g s p u n k t wird die französische Konstitutionsachse b e n u t z t . Diese wird d a n n u n m i t t e l b a r mit einer außerordentlich wohlgerüsteten Konstitutionsachse, u n d zwar der lateinischen, u n m i t t e l b a r konfrontiert. W o diese beiden K o n s t i t u t i o n s ­ achsen zu k u r z k o m m e n , werden p r ä g n a n t e Zonenanzeiger a u s a n d e r e n S p r a c h e n herangezogen. Vidyaiva param narasya bhūšanam, heißt ein altes indisches Wort. Ob die generativen offenders es jemals in seiner anmutenden Schlichtheit verstehen werden? (vgl. R. B . N o s s , The Underground Transformer, in: Language Sciences No. 23, 1972, S. 8ff.): „The leading offender here is certainly Noam CHOMSKY. His Syntactic Structures began the flight from reality t h a t is still going on, and t h a t has turned a whole generation of potential linguists into logicians. From the safe havens of their labyrinthine formulae, the generative gram­ marians have consistently ignored the real problems of linguists, both in language teaching and in the description of languages."

54

PAPERS IN THEORETICAL LINGUISTICS

Die Idiogenen Konstitutionsachsenmitglieder Deutsche Korrespondenzen

Latein

WER

QUIS

Französisch  QUI

zone de la personne

WAS

QUID

QUE/QUOI

zone de la notior

(WESSEN)

CŪIUS/-a/um

0

zone de l'inhérence immanente

(WEM)

P Yδ I W (frühmittelwalisisch)

0

0

zone de l'adhérence immanente

(FÜR WEN)

WISTWIS (georg.)

0

0

zone de la prohérence

der W I E V I E L T e

QUOTUMÜS

(LE QUANTIÈME) zone de l'ordinal

WER von beiden

UTER

0

zone de l'alternative

WELCH/er

QUÏ/QUIS

LEQUEL

zone de la désignation

WELCHERLEI

QUÄLIS

QUEL

zone de la qua­ lité

0

0

zone de l'apparence •

aus WELCHem Lande

CŪIĀS

0

zone de l'origine

WOvon (von WEM)

0

de QUI/de QUOI duQUEL

zone de la dénotation

für WAS (für WEN)

0

0

zone de la pronotation

zu WAS (zu WEM) an WAS (an WEN)

0

à QUI/à QUOI auQUEL 0

zone de l'annotation

WTE aussehend

zum W I E V I E L T e n Male

NOLAKO (baskisch)

HÁNYADSZOR (ung.)

0

QAQUGO (grönländisch) WANN

zone de l'ordinal temporal zone du futur

QUANDŌ

QUAND

„PRÄSENS LINIE" zone du passé

QANGA (grönländisch) WELCHER Tag

IKKA (japanisch)

0

0

zone du date

bis WANN

 (russ.)

0

0

zone de la durée visée

seit WANN

MIÓTA (ung.)

0

0

zone de la lurette

55

DIE RELATIVA IM NEUHOCHDEUTSCHEN

Deutsche Korrespondenzen

Französisch A

Latein QUORSUM

0

auf W E L C H E von beiden Seiten

πoτέQωσε (griech.)

0

0

zone du but alternatif

bis W O H I N

SANAMDIS (georg.)

0

0

zone du but visé

WOHIN

UBI

WO

!

I

OU

zone du but

zone du lieu

auf WELCHer von beiden Seiten

πoτέQωθι (griech.)

0

0

zone du lieu alternatif

in WELCHer Richtung

MERRE (ung.)

0

0

zone de la direction

UNDE

(d'OU)

zone du point de départ

WOHER von WELCHer von beiden Seiten

πoτέQOθεv (griech.)

0

0

zone du point de départ alternatif

aus WELCHer Richtung

MERRÕL (ung.)

0

0

zone du point de départ visé

WOZU

QUORSUS

0

zone de l'inten­ tion

WIE

QUŌMŎDO (QUI)

COMMENT

zone de la manière

0

0

zone de la la manière alter­ native

0

0

zone de la causalité

(QUĀRĒ)

0

zone de l'indice

WARUM

CUR

POURQUOI

zone de la cause

auf WELCHem Wege

QUĀ

(par OÙ)

zone de la route

WESSENUNGEACHTET

0

0

zone de la concession

0

0

zone de l'égard

0

0

zone de la condition

0



auf W E L C H E von beiden Arten

πoέΤQωs (griech.)

WIESO an W E L C H e m Merkmal

BIHWE (gotisch)

QUIN

(WESBEZÜGLICH) unter WELCHer Bedingung, in W E L C H e m Fall statt W E S S E N

(πη)

P O U R Q O U I ne

zone de I la substitution

56 Die idiogenen rakters

P A P E R S IN T H E O R E T I C A L LINGUISTICS Konstitutionsachsenmitglieder

verbalen

(bzw. satzgenerierenden)

Cha­

La­ Französisch tein A

Deutsche Korrespondenzen WAS gibt's? WAS ist (da)?

CHENHO[1] (Ouaday-arabisch)

0

0

zone verbale substantielle (ou bien existentielle)

W E R ist . . . ?

NDJENI[2] (subija) YINI[3] (zulu)

0

0

zone prédicative de la per­ sonne

0

0

zone prédicative de la notion

WE:SE:KEW2[4] (menomini) P I E U [5] (frühmittelwa­ lisisch)

0

0

zone prédicative de la qua­ lité

0

0

zone prédicative de l'inhé­ rence immanente

W O H I N gehen?

XAITŠI-[ 6 ] (mongolisch)

0

0

zone verbale du but de la marche

WAS tun? WAS machen?

JA-[7]

0

0

zone verbale proprement dite

WAS ist geschehen? WAS geschieht (wird) ?

CE šbd-[8] STA šů (ischkaschimisch) STÓXi[9] (ischkaschimisch)

0

0

zone verbale du procès

0

0

zone verbale spécifique de l'action actuelle

WAS ist . . .? von W E L C H E R A R T ist . . . ? W E S S E N ist . . .?/ WEM gehört . . . ?

WAS machst du?

(mongolisch)

Hierzu kommen noch zahlreiche prädikative A wie zum Beispiel die folgenden aus dem Menomini Eskimoischen und aus dem Menomini (vgl. Leonhard BLOOMFIELD : The Language, Yale Univ. Press 1962): WANN ist . . . ?

TA.Q-PEH

0

0

zone verbale du temps

WO ist . . . ?

0

0

zone verbale du lieu

W E S H A L B ist . . . ?

NAUK (eskimoisch) A.Q

0

0

zone verbale de la cause

W I E ist . . . ?

TA . Q

0

0

zone verbale de la manière

in W E L C H E R Richtung liegt . . . ?

A . Q-NAKAH

0

0

zone verbale de la direction

W I E V I E L ist . . . ?

TA . Q-NEKOH

0

0

zone verbale de la quantité

W I E V I E L E MALE sind . . . ?

TA . Q-TAHNE . NOH

0

0

zone verbale de la fréquence

und andere mehr, die noch genau zu erforschen sind. [1] chenho, Ouaday-arabisch, = „was gibt es?" (Die Etymologie ist unsicher), (vgl. Henri CARBOU : Méthode pratique pour l'étude de l'arabe parlé au Ouaday et à Vest du Tchad, Paris 1954 (Nouveau tirage) S. 30). [2] ndjeni, = „wer ist . . . ?", Subija (vgl. E . JACOTTET: Grammaire soubiya (Études sur les langues du Haut-Zambèze), Paris 1896, S. 33, § 54). („wer?" = ani). [3] yini, = „was ist . . . ?", Zulu (vgl. Clement M. D O K E : Text Book of Zulu Grammer, 5. Ausg. Cape Town 1954, § 134b). („was?" = ini).

D I E R E L A T I V A IM N E U H O C H D E U T S C H E N

57

[4] we:8e:kew = „welcherlei ist . . . ?", „von welcher Art ist . . . ?", Menomini (vgl. Leonard BLOOMFIELD : Language, London 1961 (6. Ausg.), S. 260) ;  = „von welcher Art ist . . . ?", russisch. [5] pieu, = „wessen ist . . . ?, „wem gehört . . . ?", frühmittelwalisisch ; kié, = „wem ge­ hört . . . ?", ungarisch. [6] xaitsi-, = „wohin gehen?", mongolisch (vgl. Nikolaus P O P P E : Khalkha-mongolische Grammatik, Wiesbaden 1951, § 192). [7]ja-, = „was t u n ? " , „was machen", mongolisch (vgl. Nikolaus P O P P E : Khalkha-mon­ golische Grammatik, Wiesbaden 1951, § 192). [8] ce šbd, = „was ist geschehen?", sta šů, = „was geht vor sich?" (sonst überall: „was?" = ciz?), ischkaschimisch (vgl. T. H . : ,   ,Moskau 1959, S. 48-49). [9] stóxi, = „was machst d u ? " , ischkaschimisch (vgl. T. H . : ,   , Moskau 1959, S. 48-49). Die proportionalen Konstitutionsachsenmitglieder Französisch A

Deutsche Korrespondenzen

Latein

WIEVIELE

QUOT

WIEVIEL

QUANTUM

W I E groß

QUANTUS

(QUEL)

zone de la quantité dense

W I E klein

QUANTULUS

0

zone de la quantité dense exiguë

zone de la quantité dissipée COMBIEN zone de la quantité conjointe

0

zone de la prix

(EN a-t-il de la chance ! — Wie glücklich ist er nicht!)

0

EN

zone de la quantité partitive

W I E lange

QUAMDIU

0

zone de la durée

innerhalb W I E langer Zeit

QUOAD

0

zone de l'expiration de la durée

seit W I E langer Zeit

0

(depuis quand)

zone quantificative de la lurette

W I E schnell

0

(en quel) délai)

zone du délai

W I E teuer

Ё (russ.) 0

QUATENUS1 JUSQU'OÙ

W I E weit

QUOTIENS

W I E oft W I E V I E L E Male

INWIEFERN/ INWIEWEIT

0 zone de la fréquence 0

HÁNYSZOR MENNYISZER (ung.) QUATENUS2

zone de la distance e de l'étendue

zone quantitative de la fréquence

COMBIEN/ COMME zone de la mesure QUE/

58

P A P E R S IN T H E O R E T I C A L LINGUISTICS

Deutsche Korrespondenzen

Latein

Französisch A

WIE

QUANTOPERE

COMBIEN/ COMME zone de l'intensité QUE/

WIE

QUAM

COMBIEN/ COMME zone du dégré neutre QUE/

um W I E V I E L

QUANTO

(de) COM­ BIEN

zone du dégré com­ paratif

0

zone du dégré équatif

P A gyn (wa­ lisisch) (Äquativ) Die Identifikativa (I) : A   D

Das subzentrische Identifikativum Die relativen Identifikativa (1, 2, 3) Die konjunktionalen Identifikativa (1, 2, 3) Die demonstrativen Identidieser,/ der/, jener, derjenige fikativa da, dort, hier, damals, jetzt, deswegen, derwegen, um deswillen, desfalls, der­ maßen E Die deskriptiven Identisolch, solcherlei, solchergestalt, solchermaßen, fikativa solchenfalls, derartig, dergestalt, dergleichen F Die suppletiven Identifikativa der eine . . . der andere ; einerseits . . . andererseits G Die akkurativen Identifikativa (gerade* er) georg. ager ( = „gerade hier"), eger („gerade da") a) exklusive (seule la forme est détestable) Lui seul le sait. Les femmes seules s'y entendent. (engl. only* . . . , deutsch nur* . . .) H Die limitativen Identifikativa b) nicht exklusive (la seule forme est détestable). La seule pensée de l'embrasser m'est en exécra­ tion. (engl. mere, deutsch bloß) c) prähibitive :         . d) präklusive (nur temporal) Alors seulement je compris son dessein. (engl. only, deutsch erst) I Die monovalenten Identider einzige Il est mon seul ami. Elle était ma seule joie. fikativa (engl. only, deutsch einzig) J Die isolierenden Identifikativa allein Il l'a fait tout seul. Il avait appris tout seul, rien qu'à regarder les lettres. (engl. by himself, alone; deutsch allein)  Die isosemischen Identifikativa derselbe, der gleiche (derselbige) Il reste toujours le même (engl. the same, deutsch derselbe, der gleiche) L Die injunktiven Identifikativa selbst er hat es nicht verstanden. Même lui nous a déçu. (engl. even* . . . , deutsch selbst* . . .)    ,  ,     . M Die epitomischen Identifikativa sie ist die Güte selbst (selber). Elle est la bonté même. (engl. itself, deutsch selbst) * Elemente wie gerade (genau, ausgerechnet; nur, lediglich; selbst (— sogar) nennen wir Identifikatoren (bzw. Reihennenner).

D I E R E L A T I V A IM N E U H O C H D E U T S C H E N N

Die konzentrischen Identifikativa



Die syntonischen Identifikativa

P

T U V W X Y

Die dignifikativen Identifikativa Die delimitierenden Identifikativa Die autosemischen Identifikativa (engl. Die autosemischen possessiven Identifikativa Die possessiven Identifikativa Die reziproken Identifikativa Das reflexive Identifikativum Das annotative Identifikativum Das denotative Identifikativum Das neutrale Identifikativum

Z

Die persönlichen Identifikativa

Q R S

59

il est mort ici même we met each other t h a t very night          (superpositiv)        (positiv) (ne) . . . quidem (negativisch) selveste kongen er tilstede (dán.)        je l'ai rencontré sur l'escalier même. il but à même le goulot. er wird es selber tun (selbst). Il le fera lui-même. himself, deutsch selber, selbst) eigen mein / dein / sein / ihr / unser / euer / einander (sich) sich (frz. y: je ne m'y intéresse pas) (frz. en: nous en avons déjà parlé) es regnet } . .. primare ich / du / er / sie / wir / ihr [Identimeinerseits, deinerseits usw. J

Die Dispositionalia (D) : (Äquativ) Æ (kymr. cymaint â = ebenso gross wie) (Superlativ) Z (am meisten, am wenigsten usw.), (kymr. mwyaf = der größte)  (Komparativ) Y (mehr, weniger usw.), (kymr. yn fwy na = größer als)  (Positiv) X a (viel(e), oft, groß, lange, weit, schnell) ß (ein wenig, selten, unweit, klein, langsam, bald) W (wenig(e), selten, (nur) langsam)  V (genug)  U a (zu viel)) frz. trop ß (zu wenig) türk. eksik  (ziemlich (viel)) ein ziemlicher Idiot S a (schon) ß (endlich) R a (fast, beinahe, nahezu) dän. næsten, frz. presque ß dän. knapnok, frz. à peine, (deutsch: kaum) Q Indefinite Dispositionalia (man, es; frz. en, y) (man sagt, es habe einen Krach gegeben) (tu m'en veux?) (com­ ment s'y prend-on pour le faire dispa­ raître?) P Universelle Dispositionalia (all, ganz, beide u. dgl.) (nicht exklusive) lauter (exklusive)  Multiple Dispositionalia (manch; mehrere u. dgl.) N Diversitive Dispositionalia (verschiedene u. dgl.) M Negative Dispositionalia (engl. nobody usw.) (niemand, kein usw.) L Restriktive Dispositionalia (engl. anybody usw.) (jemand, je usw.)  Unbestimmt existentielle Dispositionalia lat. quidam  Existentielle Dispositionalia (engl. somebody usw.) (irgendeiner, einige, mitunter usw.)

60

P A P E R S IN T H E O R E T I C A L LINGUISTICS

J I

Superadditive D i s p o s i t i o n a l i a Alternative Dispositionalia

H G F E D

Partikuläre Dispositionalia Solitäre Dispositionalia Indizierende Dispositionalia Individuelle Dispositionalia Indeterminierte Dispositionalia

 Fakultative Dispositionalia  Eventuale Dispositionalia  2 Potentiell relative Dispositionalia A1 Potentiell subordinierende Dispositionalia

(ein weiterer, weitere, noch u. dgl.) ((einige) . . . andere; sonstwo u. dgl.) (engl. other / else) (ein besonderer, besonders u. dgl.) (ein einzelner, einzelne, vereinzelte u. dgl.) (ein gewisser u. dgl.) (jeder, jedesmal usw.) (ahd. niweizhwër usw.) (ich weiß nicht wer) (frz. n'importe qui usw.) (egal wer) (ein etwaiger)

D u r c h die Artikel werden in einigen S p r a c h e n substantivische oder substantivisch verwendete Determinationen der identifikativen bzw. der dispositionalen Inhaltsseite der Sprache zugeordnet.

ON THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE Preamble to a Semasiosyntax During the sixties syntax became the main theme of linguistics, while semantics was still widely ignored. It is only of late that it has been taken up in leading linguistic circles. It has even been claimed that it would be sensible to give up entirely the difference between semantic and syntactical characteristics1. That does not only look sensible, it appears really to be the royal ladder to a comprehensive understanding of human language, which has, alas, for so long been split by speculative dichotomic linguistic theory. And as interest grew for semasiosyntax, it was only natural that interest for universal linguistic features was aroused too. Linguists must again slowly adjust themselves to work with lan­ guage in the plural, even though such may still to a great extent be regarded as out of fashion or even unnecessary. Here, as elsewhere, fashion can retard the development of a branch of science for years or even decades, almost paralyzing it. In an earlier article2 we have presented the following model of a constitutional description of language: Identificatives A Dispositionals idiogenous

not sentence-generating sentence-generating

proportional Under A (capital alpha) we understand, as is known, interrogative, respectively exclamatory sememes, consisting of a zonal content and a row content [WER? WAS? WELCHER? WANN? WO? et cetera (idiogenous/not sentence-generating), NDJENI ...? YINI ...? PIEU ...? WE1 Cf. J. Bechert - D. Clément - W. Thümmel - K.-H. Wagner, Einführung, p. 63. 2 Cf. N. Danielsen, Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsfor­ schung 28 (1975) pp. 50-73.

62

PAPERS IN THEORETICAL LINGUISTICS

SEKEW ...? et cetera (idiogenous/sentence-generating) (see below) and WIEVIEL(E)? INWIEWEIT? INWIEFERN? et cetera (proportional)]. Sentence-generating A's are very rare in the Indo-European languages. We shall try to show in this essay that these A's are of the greatest import as universal features in connexion with the syntactic study of the sentence in human language. Let us consider one after another the sentence-generating A's in several languages where they appear in a particularly clear manner: A specific predicative personal zone is indicated in the Soubiya lan­ guage through the highly significant A NDJENI3 :

The German language has no exact equivalent to the predicative A NDJENI in Soubiya. As a German counterpart functions instead the A of the personal zone (WER) + the nucleus ist. We say that the German A produces a nucleus: Muntu uzo

A Ndjeni Wer ist

dieser Mann?

As a rule, most European languages produce a corresponding nucleus in this syntactic situation: English: French: Polish: Albanian: Modern Greek: Cymric: Danish: Armenian:

WHO is he? QUI est-il? KTO on jest? KUSH ësht ay? ΠOIO∑ εωvai aύτó; PWY ydyw ef? HVEM er han? OV e an? et cetera.

A few European languages do not do this, however: Russian: Hungarian: Cf. Kanarese:

 ? KI az? Avanu YÂRU?

3 Cf. E. Jacottet, Grammaire soubiya, p. 33, par. 54.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

63

However, if the content of the sentence refers to a time which specifi­ cally falls outside the realm of the present, these languages will also produce a nucleus: Russian:

  ?   ?   ? KI volt az? KI lesz az?

Hungarian:

The A YINI in Zulu corresponds to a German A of the zone of species (WAS) with the nucleus ist4 : Ku YINI WAS ist

lokho?

 YINI

na .?

WAS sind U YINI WAS bist

das?

sie? wena? du?

The subject concordance used at any given time is placed in front of the verbal form YINI. In Malayalam, a corresponding verbal A may emerge: Idu ENDÄKUNNU? (= 'WHAT is this?' [this what's]). In this syntactical situation most European languages produce a nucleus corresponding to the German one: English: Spanish: Czech: Welsh: Albanian: Modern Greek: Danish: Lithuanian: Basque: 4 Cf. . M. Doke, Text-Book, § 134b.

WHAT is that? ¿ QUE es eso? CO je to? BETH yw hynny? SHKA ësht ajo? TI εωaι aύτό; HVAD er det? KAS tai yra? ZER hori da?

64

PAPERS IN THEORETICAL LINGUISTICS

In some languages this production of a nucleus does not take place: Russian: Hungarian:

 ? MI ez?

Outside the sphere of the present these languages will also realize a nucleus syntactically: Russian: Hungarian:

  ?   ? ( ?) MI volt ez? MI lesz ez?

A-questions are in certain languages characterized by interrogative overcharacterization: Chinese: Basque: Armenian:

Zhèi shi SHÉMmo [Dies ist was- '?'] ZER othe da? INC e as?

The interrogative overcharacterization is in this connexion of no importance and we shall not go further into this, neither here nor in the following5. A specific predicative zone of immanent inherence is signified by the early Middle Welsh A PIEU6 :

The early Middle Welsh A PIEU corresponds to a Hungarian KIE: Kié Wessen __ ist

ez a haz? dieses Haus?

5 Cf. N. Danielsen, Die Frage, p. 12. 6 Cf. J. Morris-Jones, An Elementary Welsh Grammar; J. Strachan, An Introduction.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

65

Most European languages produce a nucleus here, just like German; several of them thereby refer to a specific inherence-A: Polish: CZYJ jest ten dom? Latin: Illa domus CUIA est? Czech : ČÍ je onen dům? Serbo-Croatian: ČIJA je ta kuca? S p a n i s h : C Ú Y A es esa casa? Modern Greek: Bulgarian: Others use the genitive of the A of the personal zone (resp. the zone of species) : English: Danish: Greek: Latin: Turkish:

WHOSE is the house? HVIS er det hus? TINOS CUIUS est illa domus? Bu ev KIMINdir?

Still others use an indicated A (denotations or annotations) in con­ nexion with the nucleus produced: Dutch: Van WIE is dat huis? Rhaeto-Romanic:.De TGI ei tschella casa? QUI est cette maison? French: Rumanian: Italian: Portuguese:

A CUI e casa asta? Di CHI è quella casa? De QUEM é esta casa?

Russian may be used as an example of languages which do not produce a nucleus in the given syntactical situation: Russian: Outside the sphere of the present a nucleus comes clearly to light:

66

PAPERS IN THEORETICAL LINGUISTICS

In Welsh the sentence 'WESSEN ist das Haus (da)?' runs like this: l.Eiddo PWY yw'r ty hwn? [Eigentum wer ist das Haus jenes] 2. PWY biau'r tŷ hwn? [Wer/(wessen, rel. →), dessen-ist das Haus jenes] In both cases the PWY of the personal zone is in the nominative. In our first example the PWY stands as the second member of a status constructus. In our second example the PWY has been fronted so as to begin the sentence. Right after the PWY comes a relative status. These two Welsh sentences present an ideal starting point for a linguis­ tic explanation of the relation A between the two semantemes sein and haben. Both sentences are dealing with the same factual content. Many generative grammarians would claim that both sentences have the same content. The matter is not that easy however. Instead, our two Welsh sentences are to be considered as two different semasio-syntactical forms having the same factual content. We are here dealing with a content model, which, roughly speaking, has three main constituents: first, there is a possessor (A), second, there is a thing to be possessed (B), and, third, a possessive relationship (C) between A and  is more or less explicitly expressed. Human language manifests a clear need for the ability to express whatever is important at a given time under point three in the form of an uncomplicated and clearly prominent subject. Each of the three content constituents, ,  and C, acquire in subject position a special pregnancy and actuality within the given factual situation. The facts of the case can be presented schematically in the following way:

la) Wessen ist das Haus? b) Eiddo Pwy yw'r ty hwn? II Wer hat das Haus? Pwy biau'r tŷ hwn?

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

67

As the previous discussion makes clear, the constituents A and  will mainly show up in subject position in the given content model. C, how­ ever, refuses for some reason to take the subject position. Over this one might philosophize long and drily. What is important, however, is that A and  like to appear as subjects. For this,  requires the nucleus ist, a form of the nucleus verb No. 1, sein. A requires in this content situa­ tion a finite form of the verb haben. This verb, haben, is from this viewpoint a laterally reversed sein, whose nucleus function as predicator in an inherence expression changes into verbal emanence: A C WESSEN ist

A C WER hat

The sign refers to non-emanence, while ( signifies verbal emanence. (The valence of the verb is the syntactical manifestation of the verbal emanence, which we regard as the major all-inclusive semasiosyntactical concept). In other words: Against the non-emanent verb sein we have its ema­ nent counterpart haben. Their own value come to light in several languages, where they are used widely as 'auxiliary verbs', distributed according to their deepest sense: sein is preferably used with intransi­ tive verbs, haben with transitive. Er ist lange bei uns geblieben. Er ist wieder aufgetaucht. Er hat ihn sehr oft geschlagen. Er hat das Wasser ausgeschüttet. Il est arrive hier. Il s'est enfui. Il a rendu le livre. Il Va regardée pendant des heures. Lui è stato qui tante volte. Lei è partita alle sei. Lei è arrivata alle sette. Loro Vhanno visto parecchie volte. Lei Vha battuto col bastone. As any linguist knows, this is no hard and fast rule. That these two auxiliary verbs, sein and haben, have the very same syntactical function

68

PAPERS IN THEORETICAL LINGUISTICS

in the nucleus position, blurs the boundary between emanence and nonemanence, and their mutual distribution within the system often goes haywire: Compare:

German: Danish: German: Danish:

Er ist hier gewesen. Han har været her. Er ist mir begegnet. Jeg har mØdt ham.

In Spanish it has gone so far that the verb haber in the nucleus func­ tion here considered has completely outmaneuvered its potential counterpart ser: Mi hermano no ha llegado a Madrid. El tren ha salido de la estación. Ha habido mucha fruta este año. Norteamérica se ha hecho señora del mundo. España ha producido grandes hombres. Yo había leído ya la carta. Consequently the verb tener appears as a counterpart to haber in this nucleus situation, namely to emphasize the result of the action present­ ly expressed in the participio pasivo: Tengo escrita mi carta. Tengo pensado ir a Valladolid. In English the nucleus have gains the upper hand everywhere in these Constructions:

I have been in Yorkshire. He has left at daybreak. He has arrived already. He has fallen asleep. We have bought a new house. I have seen him lately. I have lost my last illusion. He has gone to Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch. But: He is gone.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

69

And now we have a suitable reason to pass on to our further consider­ ations: Certain languages present a specific predicative adjective zone. Such a zone is, for instance, expressed in the Menomini language through the A WESEKEW? 7 : QUALIS

WESEKEW

est

WE.SE.KEW wigwam? = What kind of house is this? Most European languages present a nucleus like the Latin A QUALIS. The verbal WESEKEW of the Menomini language corresponds to a nonverbal adjectival A with a nucleus in the other languages. Now, in an optimal constitutional axis there are various adjectival A's. They may be illustrated this way: A a) QUOTUS lat. b) QUOTUMUS

Ó∑0∑ (Greek)

Ordinal zone

lat. UTER

HWApAR (Gothic)

Alternative zone

lat. QUI

ZOIN (Basque)

Zone of designation

lat. QUALIS

ZER (Basque)

Zone of quality

lat. (CUIUS apparitionis)

NOLAKO (Basque)

Zone of appearance

lat. CUIAS

CUIAS (Latin)

Zone of origin

7 Cf. L. Bloomfield, Language, p. 260; L. Bloomfield, The Menomini Language, pp. 480-481.

70

PAPERS IN THEORETICAL LINGUISTICS

Examples: Greek: (= Which one (of a number) is this?) Gothic:

HWA AR ist sa? (= Which (of the two) is he?)

Basque:

ZOIN etche da? (= Which one is the house?) ZER etche da? (= What kind of house is this = How is the house?) NOLAKO etchea da? (= What kind of house is this? What does the house look like?)

Latin:

CUIAS est? (= What countryman is he?)

In all six examples a nucleus is to be noted. Most European languages require a nucleus in sentences of this kind. A few however do not produce such a nucleus: Russian:



Hungarian:

MILYEN az udvar? (= What kind of farm is it? = How is the farm?) MELYIK az udvar? (= Which one is the farm?)

Outside the present the nucleus is consistently expressed: Russian:

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

Hungarian:

MILYEN volt MILYEN lesz MELYIK volt MELYIK lesz

71

az udvar? az udvar? az udvar? az udvar?

In some languages the nucleus can be omitted in a given sequence: Filia tua UTRA (est)? Filius tuus QUOTUS (est)? (= Which one (in the number) is your son? (i.e. 'with how many men is your son coming?')) QUALIS in circuitu ascensus? et cetera. A similar optionality may be observed in differently realized predicat­ ive adjectival constructions: QUIS clarior (est) Themistocle? (cf. Gothic: Du amma HWAS wair s? (2 Cor.: 2:16)) Tarquinio QUID (est) impudentius? Omnia praeclara rara (sunt)? The omission of the nucleus is especially symptomatic for rhetorical questions with a negative content. Facts are, however, that most European languages make use of a copula: According to the factual content of the constitutional axis predicative substantives and adjectives with a nucleus will be produced. The syntactical position of this nucleus differs from language to lan­ guage. Let us look at the position of the nucleus in a few European predicative sentences with exactly the same content (the predicative complement is printed in italics) : E n g l i s h : A wise man is honest. Georgian: Irish: Is macánta fear feasa. Breton: Onest e vez eun den fur. Welsh: Y mae dyn doeth yn onest. Manx: Ta dooinney creeney onneragh. Gaelic: Is e duine onorach duine ciallach. As can be read out of this material, the nucleus will in most cases lean heavily against the predicative complement. That this can in no way be construed as a cathegorical imperative is shown by the two standard

72

PAPERS IN THEORETICAL LINGUISTICS

sentences from Welsh and Manx, where the nucleus is removed as far as at all possible from the predicative complement. According to the factual content of the constitutional axis predicative clauses in the present will in several languages have no nucleus: Russian: Hungarian: Arabic:

Egy okos ember bescületes.

Occasionally polarized nuclei may be observed. Each polarization has its own nucleus in the 3rd pers. sg. pres. ind. Thereby it becomes poss­ ible to represent separately the signs of recusative polarization and the nucleus (1), or they may be expressed by one and the same sign (2): (1) Manx:

Ta dooinney creeney onneragh. (= a wise man is honest) Cha vel dooinney creeney mee-onneragh. (= a wise man is not dishonest)

(2) Czech:

Rozumny muž je poctiv. (= a wise man is honest) Rozumny muz není nepoctiv. (= a wise man is not dishonest)

(3) Lithuanian: Ismintingas vyras yra doras. (= a wise man is honest) Ismintingas vyras nèra nedoras. (= a wise man is not dishonest) In certain other languages only the propositive polarization is furnished with a nucleus in the present: Irish:

7s macanta fear feasa. (= a wise man is honest) Ní mímhacánta fear feasa. (= a wise man is not dishonest)

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

73

Scottish-Gaelic: Is e duine onorach duine ciallach. (= a wise man is honest) Chan e duine neo-onorach duine ciallach. (= a wise man is not dishonest) A modern American logisticistic-psychologic linguistic school draws up the following deep structure for the English sentence a wise man is honest8 :

a man / man is wise / is honest They do it in dead earnest: But why they do it becomes more and more incomprehensible to me with every day that passes. One thing stands out as a clear fact, not least according to the above amplifica­ tions: This sentence stemma has absolutely no universal value. When seen in a larger context it looks as though it has been taken right out of thin air. From a universal viewpoint it has no importance for the mechanics of human speech. 8 Cf. N. Chomsky, Language, pp. 22-24 (the Massachusetts magician and his mostly monoglot followers hardly ever seem to bother about language[s]; they make themselves guilty of the gravest transitus ab intellectu ad rem ever seen in linguistics; cf. E. Coseriu, Sincronía, p. 9).

74

PAPERS IN THEORETICAL LINGUISTICS

And with that this question arises: Indeed, what elements are really used as carriers of nuclei, respectively as non-carriers? Let us just review what has been proved thus far. I. First, we are concerned with the elements which can be used as an apocrisis to an A of the personal zone: Wer ist ... Wer ist dieser Mann? Er ist Reinhold Schröder. Wer ist diese Frau? Sie ist Mechthild. Wer ist dieses Mädchen? Sie ist die Tochter Reinholds. In Polish the   is nominative and asks for the person, while its instrumental KIM asks for his profession or his whereabouts:  on jest? KIM on jest ?

On jest Zdzis aw Kochanowski. On jest profesorem.

In Russian the nominative asks both about the person and the pro­ fession:

Outside the sphere of the present the distribution of the nominative and the instrumental case is like in Polish:

II. Secondly, we are concerned with elements which are suitable for use as an apocrisis for an A of the zone of species: WAS

ist...

WAS ist das?

Das ist ein Mann. Das ist eine Frau. Das ist ein Mensch.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

75

Das ist ein Tier. Das ist Onkel Otto. Das ist ein Buch. Das ist Wasser. Das ist Krieg. Das ist eine Hoffnung. Das ist ein Nichts. Das ist Liebe. Das ist Arbeit. III. Thirdly, we are concerned with elements which are suitable for use as an apocrisis for an A of the immanent inherence zone: WESSEN

ist...

WESSEN ist das? dan. HVIS er det? engl. WHOSE is that? Hung. KIE az?

Es ist meins. Det er fars. (Det er pokkers). Det er vores allesammens. It's mine. It's father's. Az az apámé. Az enyém.

IV. Fourthly, we are concerned with elements which are suitable for use as an apocrisis to an A of the adjectival zone. Let us take another look at the schematic presentation of the idiogenous adjectival. A Which one of a number Which one of a row

lat. QUOTUS QUOTUMUS

ΠÓ∑TO∑

Ordinal zone

Which one (of two)

lat. UTER

HVAþAR

Alternative zone

Which

lat. QUI

ZOIN

Zone of designation

What kind/'how'

lat. QUALIS

ZER

Zone of quality

'How' it looks

lat. (CUIUS apparitionis)

NOLAKO

Zone of appearance

from What country

lat. CUIAS

CUIAS

Zone of origin

76

PAPERS IN THEORETICAL LINGUISTICS

Presently we are interested in finding out how the idiogenous adjectiveA's, which may be used at any given time, behave in regard to predica­ tion. They may be used adjectivally all six, but there are, however, two kinds of adjectives: Those which preferably appear attributively, and those which preferably appear predicatively. Besides, there is a vast number of adjectives, which may be used in both ways. This fact can be stated outright however: The A of the zone of designation is in all languages struggling against being used predicatively. In this fact we may already see an indication regarding the often rather capricious dis­ tribution attributive : predicative: An adjectival apocrisis to the six possible A's manifests itself accord­ ingly either only, respectively preferably, attributively (1), or only, or predominantly predicatively (2), or both predicatively and attribu­ tively (3) whereby they can exhibit certain differences in meaning according to their function at any given time (4). 1. Eine völlige Freiheit ist nie zu erreichen. Bares Geld ist erwünscht. Die mutmaßlichen Attentäter sind eingesperrt. Die obere Tragmulde ist kaputt. Die bisherigen Versuche sind gescheitert. Der eingefleischte Junggeselle ist unempfänglich für Schmeicheleien. Der linke Flügel ist restauriert. Seine äußerliche Ruhe ist irreführend. Sein angeblicher Vormund ist knapp bei Kasse. Etwaige Beschwerden sind zurückzuweisen. Die eigentlichen Interessenten werden vorstellig (et cetera). 2. Ein gleichgültiges Mädchen war fehl. Ein alter Freund ist uns abhold. Die beiden Kritiker sind wett. Der alte Schauspieler ist wohlauf. Der kleine Junge ist perplex. Unser langjähriges Mitglied ist pleite. Unser neues Motorboot ist klipp und klar. Das bayrische Bier ist alle. Die bösen Frauen sind schuld (et cetera). 3. Ein großes Unglück ist geschehen. Das Unglück war groß.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

77

Ein schönes Mädchen war zugegen. Das Mädchen war schön. Ein langer Weg ist erschöpfend. Der Weg ist lang. Der gute Arzt ist getrost. Der Arzt ist gut (et cetera). 4. Die arme Frau ist blind. Die Frau ist arm. Eine tüchtige Ohrfeige wäre angebracht. Der Junge ist tüchtig. Ein hoher Würdenträger ist gestorben. Der Würdenträger ist hoch. Eine natürliche Tochter ist erbberechtigt. Seine Tochter ist sehr natürlich. Der blaue Himmel ist wolkenlos. Der amerikanische Student war blau. Ein starker Raucher ist wagemutig. Der Schmied ist stark. Ein alter Freund von mir ist gekommen. Mein Freund ist alt. Das ist glatter Selbstmord. Die Fahrbahn war glatt (et cetera). It is an established fact, that there are many apocrises among the adjectives, which cannot be combined with a nucleus. A semantic category not to be underestimated is the adjectival attributivum, which, in German, for instance, makes its grammatical independence apparent in the fact that it must be declined separately. The Slavonic languages teem with specifically attributive adjectives, which unite with whatever substantives they are added to, to form conceptual units, which in many other languages would be expressed by means of compound nouns:

78

PAPERS IN THEORETICAL LINGUISTICS

Russian:

  (= chained dog. Germ. Kettenhund)   (= powder magazine)   (= dirge)   (= grapevine)   (= metre [in poetry])   (= light effect)   (= glottis) et cetera.

The above survey of the idiogenous adjectival A's of the human language shows how the German language tends to utilize the adverb­ ial A WIE from the zone of way and manner within the zones of quality and appearance: WIE ist er? WIE sieht er aus?

Er ist sehr gemütlich. Er ist dunkelblond und blaß.

There are also other adverbial A's which produce a predicative nu­ cleus: WO ist er? WANN ist es?

Er ist hier. Es ist morgen.

(In Greenlandic, the sentence-generating A NAUK means 'WHERE is'). Also, certain substantival and adjectival proportional A's produce a nucleus (in this they do not differ from the other substantival and adjectival elements): WIEVIELE sind sie? WIEVIEL ist es?

Sie sind zehn. Es ist eine Mark.

In the same way detached adverbial proportionalia produce a nu­ cleus: WIE oft ist es?

Es ist zweimal in der Woche.

Human language realizes widely 'topical' and 'iterative', (respectively 'habitual') alternatives to a predicative nucleus:

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

79

This figure appears to reflect a linguistic law, which repeatedly be­ comes manifest in the arrangement of the human speech mechanism. It concerns the binary opposition between an X and a Z, which, if need be, may be supplemented with a third factor, a Y, where necessary, or because the Y makes the categories X and Z superfluous. This law may be expressed in this way: (1) There is something, which is X. (2) There is something, which is Z. (3) There is something, which is neither X nor Z, or which may render both X and Z superfluous. This law we can observe, for example, in the use of the anaphoric pronouns in German: X: Er liebt sie. Y: Es regnet. Z: Sie liebt ihn. Or, we can take the category of grammatical gender: In Danish you operate with an X (commune) and a Z (neutrum). In French we also have an X (masculin) and a Z (féminin). In German, on the other hand, we have an X (Masculinum), a Y (Neutrum) and a Z (Femininum). The English language makes do with a Y. In the case of a predicative nucleus German makes do with a Y-element, just like many other languages: Er ist charmant. Er ist außer sich. Er ist immer geduldig.

80

PAPERS IN THEORETICAL LINGUISTICS

Several languages have actualizing predicative nuclei: Breton:

Ar bugel a zo eveziant. (= the boy is bright) (Y) Eveziant eo ar bugel. (= the boy is bright) (X)

Spanish: El alumno es atento. El alumno está atento. El alumno está en la escuela. In several languages specifically iterative, respectively habitual predi­ cative nuclei may be observed: Breton:

Eveziant e vez ar bugel. (= the boy is bright)

Russian:      ,    .     . Already our latest Spanish example (just above) indicates the necess­ ity of establishing one more category, namely, the category of the situational predicative nucleus; indeed, it is very relevant to consider this category as a new Z, which stands in opposition to X, that is, as seen from a universal viewpoint:

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

Breton:

81

Er skol ema-en. (= he is in the school) Er skol ema-hi. (= she is in the school) Hizio emaint eveziant. (= today they are bright) (Z)

We may consider the Verbum substantivum as a Y to the latest intro­ duced Z (always from a universal viewpoint):

The Verbum substantivum appears in some languages as a predicative nucleus: Scottish-Gaelic: Tha duine ciallach gu h-onorach. (= a (particular) wise man is honest) Irish:

Tá fear feasa macánta. (= a (particular) wise man is honest)

When the Verbum substantivum does not appear as a predicative nucleus it signifies mere existence (see below). Presently it is of particular importance to investigate the function of the predicative nucleus elements here discussed in passive constructions. The English language, for example, uses a Y-nucleus throughout in passive constructions:

82

PAPERS IN THEORETICAL LINGUISTICS

He is detested by his pupils. He was killed. The verb get can be used as an emanent passive-nucleus: He got killed. (cf. Get going!) The passive durativity is expressed through a transformation of the Y through a Y: He is being killed. In French a Y-nucleus functions similarly as a realizer of the passive: Il est admire par le public. Il fut tué. In Spanish the passive is formed by means of a Z-element: La puerta es abierta. El pájaro fue muerto por el muchacho. Against this an X-nucleus takes care of the copula function, when the participial predicate is to be understood as a pure adjective, i.e., when the participio pasado in the predicative position must be considered a mere apocrisis to an adjectival A of idiogenous character: La puerta está abierta. In this connexion we have only one thing left, which is to enter once again shortly upon the subject of the Verbum substantivum left a short while ago. Let us consider this very important content element again, now when it stands exposed in its most subtle and intrinsic function, namely to signify pure existence. A specific A of the predicative description of pure existence is found in Ouaday-Arabic9 : áENHO

WAS resp.WAS

9 Cf. H. Carbou, Méthode pratique, p. 30.

ist} gibt es?

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

83

Let us consider what corresponds to this very interesting A in a few illustrative languages: Turkish: Spanish: German: Serbo-Croatian: Danish: Swedish: French:

NE var) i QUÉ hay? WAS ist? ŠTA ima? HVAD er der? VAD finns det? QU' y a-t-il?

Apparently constructions of emanence and expressions of nonemanence vie with one another to express the condition of pure ex­ istence: Turkish employs a construction with a species-A + a specific exist­ ence nucleus (which, by the way, is polarized): NE var NE yok?

(= What is there? What isn't there: What's news?)

Spanish uses a construction with the species-A and a separated specific emanence nucleus. The sentence has no subject. German uses a species-A + a separated nucleus denoting pure exist­ ence: Was ist jetzt? Serbo-Croatian employs the A of the species zone with a separated emanence nucleus (ima is 3rd pers. sg. pres. ind. of the verb imati 'haben'). The same phenomenon can be observed in corresponding Polish constructions, in the recusative polarization: Nie ma nic nowego. (= there is no news) In Danish the Y-nucleus er is produced from an A of the species zone and furnished with the indeterminate dispositional of the locus zone. The A is the subject(→predicative10).

10 Cf. N. Danielsen, Sprachwissenschaft 3 (1978) pp. 184-224.

84

PAPERS IN THEORETICAL LINGUISTICS

In Swedish the A of the species zone produces the specific expression of existence finns, which then is furnished with the indeterminate dispositional det of the species zone. The sentence has two subjects. In French the species-A produces the emanent nucleus a. This nucleus is furnished with the indeterminate dispositional il of the species zone as well as with the indeterminate dispositional y of the locus zone. Il is the grammatical subject. It is well known that the German sequence of words 'WAS gibt es?' is much used. Here the verb geben is used as an emanent nucleus (cf. the above English passive constructions with the nucleus get, e.g. He got killed) just like the nucleus constructions found in certain other lan­ guages, which also employ finite forms of the verb 'to give*, cf.: Japanese:

Furukawa san wa boku wo tokidoki sūkiyaki wo tabe ni Ginza e tsurete itte kuremasu. (= Furukawa San often takes me to the Ginza, to eat sukiyaki.)

In the above German construction the grammatical subject is the indeterminate dispositional of the species zone: es. In Southern German (in Bavaria, Austria and Switzerland) corresponding constructions are realized with the verb of emanence haben: Heute hatte es wenige Zuschauer auf den Pisten. In several languages the non-emanent verb of existence appears as nu­ cleus in sentences, which express a possessive relationship. Thus the thing possessed becomes subject of the sentence: Welsh:

Y mae cap gan y bachgen. (= the boy has a cap) Nid oes cap gan y bachgen. (= the boy does not have a cap)

Hungarian:

Gábor bácsinak nagy családja van. (Uncle Gabor has a large family) Gábor bácsinak nagy családja nincs. (Uncle Gabor does not have a large family)

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

Turkish:

85

Büyük bir odam var. (= I have a large room) Büyük bir odam yok. (= I do not have a large room)

It is interesting to note the expressions of existence with opposite polarization. In Russian the propositively polarized expression of existence  is only used thus, when it receives special emphasis:  .    .  e. (the 'logical subject', i.e. the electron, is in the genitive of superimposition) Cf. Polish:

On jest tutaj. (= he is here)

But:

Nie ma go tutaj. (= he is not here)

The propositive Polish sentence is constructed just like the one that corresponds to it in German; its recusative counterpart has an emanent expression of existence with the 'logical subject' in the genitive. By way of summarizing we can establish these points: 1. Some sentences have a subject and a verb. Ex. : Ne var} or Was ist? 2. Some sentences have no subject. Ex.: Šta ima? 3. Some sentences have no verb. Ex.: Quid mihi? 4. Some sentences have neither subject nor verb. Ex.: Senho? 5. Some sentences have two subjects. Ex.: Vad finns det?

86

PAPERS IN THEORETICAL LINGUISTICS

6. Some sentences have two verbs. Ex.:     . I find this worthy of some consideration. These are the linguistic facts. The science of linguistics cannot get around it but must consider these very simple and unavoidable facts. The logics of Aristotle will not be of any help at all. Logics is a product of language and so it cannot from itself produce a living language, let alone explain it. Thus far we have dealt with nine sentence nuclei. They may be illus­ trated graphically, to wit, in a tricolic way: I. Emanent nuclei:

II. Immanent nuclei:

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

87

In Ishkashimian there are a couple of elements which are bound to interest us in this connexion11 :

Clearly we are here dealing with interrogative constructions which ask about the process itself. At first sight the presently discussed syntactical situation in Ishkashi­ mian may appear to be quite trivial. The matter is more complicated, however, than one might think at first. The A of the Ishkashimian species zone (= WHAT?) is neither CE nor STA, but ČIZ: CÍZ-bt zůγd? (= What have you taken?) The exclusive function of the element CE is to establish specific zones outside the species zone. It only emerges in three connexions: 1. CE ári? 2. CE-rang?

= =

3. CE

=

šbd?

Why? How? What kind? (amí Ce-ráng odám? = What kind of man is he?) What happened?

The A STA appears only in two combinations: 1.STA-s góxi? (›Stóxi?) = What are you doing? 2. STA su? = What happens? With that a distinct zone of the predicative process has been realized. It is divided into an X (preterite) and a Z (present/future). The differ­ ence in tense is made clear in the nucleus. In German the verb werden appears in its various forms as nucleus of the pure process. Er wird Mann. Er wird klug. Er wird besser. (with the comparative) Er wird bestraft. (passive) Er wird kommen. (future) 11 Cf. T. H.

pp. 48-49.

88

PAPERS IN THEORETICAL LINGUISTICS

The process verb werden may be considered as part of a tricolic diagram:

X and Z are in this diagram active, resp. inactive in relation to Y, which stays neutral towards both of them. Z-elements can also appear as nuclei: Er bleibt stehen. Er blieb ihr treu. Er blieb ein Narr. The left Z-elements require a fuller explanation: In Khalkha-Mongolian we find the following important A 12 :

Here we are also concerned with [gehen], but this time in a very concrete sense. We also know the zone displayed by the A XAITSIfrom the Indo-European languages. Gothic displays quite distinct demonstrative identificativa belonging to this zone; they are admittedly only found in the imperative mood: Hiri! Hirjats! Hirjiþ!

(= come here!) (= come here (both of you)!) (= come here (all of you)!)

The verbs of coming and going play quite a large role in many areas as finite nucleus-morphemes in more complicated verbal constructions: French: English: German: Italian: German:

Il va venir. (very near future) He goes mad. Er geht fischen. et cetera. Lui viene ammazzato. (passive) Das kommt viel schlimmer.

12 Cf. N. Poppe, Khalkha-Mongolischc Grammatik, par. 192.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

89

Presently it is quite warranted to incorporate the elements gehen and kommen as steps in an expanded tricolic diagram. Thereby we shall be able to differentiate between concretizing and abstract elements:

The Z grow (increasing) and the X go (initiative) stand as intransitive opposite the transitive Y keep (continuative). Z He grew mad. Y He keeps crying. Tengo una casa. Tengo hambre. (The Spanish constructions have become expressions of [have]) X Don't go saying anything to Peter! Allons voir le film italien! Allons, allons! I am going to see him tomorrow.

90

PAPERS IN THEORETICAL LINGUISTICS

The Z komm/en and the X geh/en stand as expressions of movement with opposite orientation over against the neutral Y steh/en (bleiben): X Vamos a ver! Y A: Vieni! Vieni! - B: Sto andando! (= I come!) (Italian) Z Il ragazzo viene bastonato. (= the boy will be thrashed) (Italian) To the Y [keep] correspond a Z = [give] and an X = [get]: Es gibt viele Deutsche in der Schweiz. Donoŋ sypahi ə1 die. (= 'both the sepoys went off') (Urdu)  ! (Russian) He got drunk. He got killed. Accordingly, the Y [hold] is matched in one dimension by an X = [put] and a Z (= [take]), and in another dimension by the X [let (go)] and the Z [catch]:

Vélin hélt sér gangandi. (Icelandic) Tengo escrito el artícolo. (Urdu) (= 'all the trees were cut down') The captain took ill. əŋ ne sara go∫t kha lia. (= 'I ate up all the meat') (Urdu) Het let him run.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

91

Meg fogjuk beszélni a nyaralást is. (Hungarian) (= 'we shall discuss the summer-holidays, too') An extremely productive row of nuclei is made up by the verbs

Die Mutter schickte sie schlafen. Er legte das Kind schlafen. Sie setzte den Schmöker ins Regal Staub fressen. Ich stellte den Mauersteinroman aufs Bücherbrett Staub sammeln. El rey se puso furioso. Le juif se mit à pleurer. Er warf den Happen in die Butter schmoren. Xara görözö aladzi orxiko. (Mongolian) (= 'the bear killed him') Las mujeres echan a correr. The Y [stand] has an X = [fall] and a Z = [get up/rise]: Sto venendo. (Italian) He fell ill. Lərki ro pəri. (= 'the girl burst out crying') (Urdu) Meri valdə bol wthi. (= 'my mother suddenly spoke') (Urdu), and the continuative Z [go on] has the Y [stay] and its X [stop] and Z [get going/start]: They went running towards the station. He went on singing. Məyŋ kha cwka. (= 'I have already eaten') (Urdu) Becari rone ləgi. (= 'the poor woman started crying') (Urdu)

92

PAPERS IN THEORETICAL LINGUISTICS

At long last there is only left for us to discuss the verbal A over all verbal A's; it is found in Khalkha-Mongolian13 :

The Mongolian A has a nearly complete conjugation and behaves for the most part just like any other Mongolian verb 14 . In German the A of the species zone produces the Verbum vicarium, the pure verbal con­ tent of the zone. The Verbum vicarium is very important as a sentence-morphematic nucleus in several languages: English: Whom does he love?

Do love me!

He loves me. He does love me.

He does not love me.

Does he love me?

Does he not love me?

(According to the rule given, the sentences within the right-angled amblygon are to be realized with the nucleus do), German: Was tut er trinken? (colloquial) Er tut Mundharmonika spielen. Er tut dich leiden. Italian: Mi ha fatto piangere. Fammelo vedere! Le posso far vedere del rossetto? French: Il fait froid. 13 Cf. N. Poppe, Khalkha-Mongolische Grammatik, par. 192. 14 In Tagalog the verbal prefix um- + the A ano (= WHAT?) gives the A UMANO (= 'do WHAT?'): UMAANO ka? = what do you do?

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

93

The Italian examples point to a tricolic division into an X (licitative) and a Z (causative verbs), vis-à-vis a neutral category Y:

All three are used as nuclei in German: (colloquial) Er tut sie lieben. Er läßt ihn laufen. Er machte mich weinen. Please note the difference between: 1.Ill'laissé peindre. = Er hat ihn malen lassen. and 2.Ill'fait peindre. = Il a fait faire son portrait. = Er hat ihn malen lassen. In German and Danish the syntactical structures of these sentences do not differ in the least, which often leads to misunderstanding. The licitative X may evidently be split into two categories: X (to supple­ ment the active) and Z (to supplement the passive):

94

PAPERS IN THEORETICAL LINGUISTICS

The latter Z belongs semantically to the field of the above discussed tricolic diagram 'give'/'have'/'get'. With this we are in a position to define a syntactical nucleus in the following way: A nucleus is a verbal kernel which has been produced by a sentencegenerating A-universal, and elements corresponding to it, to which it stands in a topological (dicolic or tricolic) semasio-syntactical inter­ relationship. Modal verbs are also nuclei, but of a different sort. As modal nuclei they belong to a third dimension of the verb, in the modal sphere. This is however a very different matter, into which we cannot enter present­ ly, for lack of space. Finally, there is a third category of nuclei, which play an exceedingly important role in certain languages. These are the polar nuclei, which we know, for example, from the Finnish language. The sign of recusative polarization (= the word not) is in Finnish really a nucleus, which is conjugated like a verb: en palannut heti = I did not come back at once. et palannut heti = You did not come back at once. ei palannut heti = He did not come back at once.

THE NUCLEUS CONSTRUCTIONS IN HUMAN LANGUAGE

95

emme palanneet heti = We did not come back at once. ette palanneet heti = You (pl.) did not come back at once. eivät palanneet heti = They did not come back at once. An exhaustive semasio-analytical analysis of the nuclei has yet to be made. It is however an indispensable prerequisite for anyone who wants to put something meaningful down on paper about the syntax of a human language and its universals.

A FIRST CONSTITUTIONAL STEP TOWARDS A UNIVERSAL SYNTAX A constitutional approach to a syntactic description of human language sets out on the assumption of seven maximes whose general validity finds its only epistemological support in well-founded empirical linguistic investigations: (1) (2) (3) (4) (5) (6) (7)

There There There There There There There

are are are are are are are

sentences sentences sentences sentences sentences sentences sentences

with a subject (S) and a finite verbal predicate (vP). without a subject. without a finite verbal predicate. witht neither a subject nor a finite verbal predicate. with two subjects. with two finite verbal predicates. in which a finite verbal predicate is the subject.

A constitutional deduction follows a procedure which makes due allowance for three simple and consistent linguistic principles of occurrence and availability: 1. There is something which is X. 2. There is something which is Z. 3. There is something which is neither X nor Z (i. e. Y), where Y is either in a given system or where Y supersedes X and Z.

called

for

The relation X : Z is a dicolic relation. The relation X : Y : Z is a tricolic relation. Y alone implies default of a category in question. This simple theory seems to give a new point of departure to linguistic analysis whether it be semasiosyntactic or phonological or morphological, and whether it be synchronic or diachronic. Used reasonably and with systematic care, its notion of the third possibility might turn out to be a life-giving asset in any linguistic field which has for years been a fit subject for rigid dichotomic hypothesizing. In this ar­ ticle we shall try its capacity in the field of universal syntax, a subject which till recently we would have shaken our head at as hardly susceptiple to any serious re­ search. Taking our stand on the tricolic relation 0 S 0(X) : 0 0(Y) : 0 vP 0(Z), where S signifies 'subject', vP 'finite verbal predicate' and Ø 'none' of the two, the following diagram evolves, so to say, all by itself (P stands for predicative, vn stands for verb nucleus, and DIE stands for demonstrative indication of existence). The resultant diagram represents a model which is extremely easy to handle : 1. The general survey of basic syntactic possibilities is-divided into three vertical rows ß,γ, and δ. 2. Every where in the system a Y implies an X and a Z. 3. X , Y and Y, Z indicate predicative sentences without a subject and without a verbal nucleus respectively.

98

PAPERS IN THEORETICAL LINGUISTICS

The critical fields of distribution

thus achieved of the human language are the follow-

(vP indicates 'verbal protone' [ = the finite verb in a critical field of distribution], S in­ dicates 'subject', P indicates 'predicative', DIE indicates 'demonstrative indication of existence', vn indicates nucleus [of existence, process, or remanence], Ø indicates 'zero' and $ indicates unmaterialized subject). 4. The model has four appendants : a) Y and X : (read to) 6 Z. b) Z and Y : ß X. Y X c) X and Y : ß Z. d) : X : ß Z (the only dicolic relation in the system ; future research may make out the case of a tricolic relation even here). 1 This group of CFDs is to be considered as nuclear predicative explicitations of the CFDs Y:δZ and X:δZ.

25 Z. Phonetik 4/77

A FIRST STEP TOWARDS UNIVERSAL SYNTAX

99

In the following we shall take a look at the contents of every single box in the system. We shall start with the tricolic relation which makes up the vertical middle balk of γ-sentences and then follow the diagram joint by joint :

γX:

yY:

yZ:

Portuguese : Hungarian : Turkish: Manx : Russian : Russian: Greek : German : English : Danish: Polish: German : German : Hausa : Swedish : Latin :

βX:

δZ:

English : German : French : Greek : Seres : Italian : Russian : IcelandicIcelandic : German : Latin :

É agradável passarmos a noite juntos. ( = „ i t is pleasant t h a t we (can) spend the night together") Sikerült elmenekülniök. ( = „ t h e y succeeded in escaping") Oldu olacak. She dooinney creeney t a onneragh. ( = ,.a wise man is honest")

Hierher mit der Kanone ! Down with the t y r a n t s ! H i t med pengene ! Nie znaleziono dzielnego podróznika. ( =,,the brave explorer wasn't found") Gleiche Arbeit, gleicher Lohn. Der — (und) Inszenierung! Quanqanen yaro nan, qarfi gare shi. [small boy this, power with. him] ( = „ t h e small boy is strong") J a g en kvinna. ( = , . I — a woman") Esse quam videri.

That idiot! Der Trottel ! Le voyou ! (Matth. I I I , 3) Samba diñkeñkatu. [Samba's (his)-having struck me] ( = „Samba struck me") (Guinea-Bissau) Piove. Vindinn lægði. ( = „the wind calmed down") H a n a vantar bókina. ( = „she is in lack of the book") Hier wird gearbeitet. Bello utendum est nobis.

100

PAPERS IN THEORETICAL LINGUISTICS German: Russian: Polish:

I h m wurde von seinen Schwiegereltern geholfen.   . Znaleziono dzielnego podróznika. ( →Y : ßZ?) ( = „a valiant explorer was found") Scottish-Gaelic: Chuidicheadh le Dômhnall a bhràthair. ( = „his brother was helped by Donald")    . Russian : Urdu: Lərke ne meri ( = „ t h e boy palled my leg") 2

βX:

Y:βX Z : βx if: Z βZ

German : English : French : Seres : English : Latin : German : Arabic: Russian:

Der Trottel ! That idiot ! Le voyou! Samba dinkenkatu. [Samba's (his-)having struck me] ( = „ S a m b a struck me") (Guinea-Bissau) Even he!3 Et tu, Brute ! Selbst mein eigener Bruder ! Al-hamdu li-lläh. -  !

2

To all sentences in the indicative correspond different categories in a third dimension: the subjunctive (Y) and the imperative (Z), the iussive (Y : Z) and the optative (X : Z), and so on:

Pigaiuk? (Eskimo) ( = d o you use it?) : pigât ( = y o u use it) 3 Three categories are attached to Y/Z : βX and βX in a third dimension: on one hand the objective (Z in a dicolic relation) and on the other hand the inhesive (Y) and the vocative (Z in a tricolic relation) : 25*

A FIRST STEP T O W A R D S U N I V E R S A L S Y N T A X X : Z : ßZ :

ßY:

Latin : Russian : Welsh: Lithuanian:

German :

X:βY:

Swedish: Russian : Japanese :

Z

:ßY:

Icelandic :

ßZ:

Scottish-Gaelic English : German : French : Spanish : Russian : Ibo:

101

Ecce homo !   ! Dacw'r d v n ! ( =,,Iook, t h a t ' s the man"!) Štai Kaunas!

Es war einmal ein König. Es folgt die Presseschau. Es wird vieles getan. Det var engâng en konung.   ? Ano onna-no-hito wa san-nin kodomo ga arimasu. ( = „ t h a t woman has three children") ( = „it aroused something of a stir") Is e duine onorach duine ciallach. A wise man is honest. Colorless green ideas sleep furiously. Ein Unglück kommt selten allein. L'amour pardonne tout. Cristóbal Colón ha descubierto a América.      . Nwokea malu ife bu onye eziokwu. ( =,,a wise man is honest")

continuation of footnote 3 German : Junge ! Junge !/ Water!/Charles !

vocative inhesive subjective investitive objective

investitive :

Danish: Pokkers også! Water! ( = there is water) Water! ( = bring me some water!) Latin :   miserum ! German : Guten Tag! = Hungarian : J ó napot !

X genitive

Y

Z

—

Y

dative (giboganemo kneue fora imo (Tatian, Matth. 27, 29)').

z

estando la señora en el campo

accusative ablative (den Kopf (Cicerone „ gesenkt) consule) ♦The investitive can be considered as being situated in a third dimension to γY.

102

P A P E R S IN T H E O R E T I C A L LINGUISTICS

δZ:

Welsh: Arabic : Italian : German : Icelandic : Irish : Baruan :

δY: X : δY: Z : δY:

ßY:

German : Russian : Hebrew : Russian : Arabic : Russian : Hungarian Russian:

Y mae dyn doeth yn onest. ( = „ a wise man is honest") Wasalū t-talāmidhatu 'ilā 1-madrasati. ( = „ t h e pupils arrived in to the school") Vengono gli altri. Bist du ein blöder Typ ! Hafði þá tekizt að jafna ágreining Ira og Breta. ( = ,.a compromise between the Irish and the British had then come about") Bhris sé an claí orm. ( = „he broke m y fence") Sa vinaka na vale ni n. [is good the home of eat] ( = „ t h e restaurant is good") (Fiji Islands) Laß ich es lieber!  . ' 'elohim et ha-'ares. ( =„God created the world")    . Yak ūnu zaidun kataba. ( =,,Zaid will have written")    . Mennek-jönnek az emberek.    .

G e r m a n : P h a n t a s t i s c h dieser Sprung! Sanskrit: samudrasya. Malgassian : Maitso ny ahitra. [green the grass] ( = „the grass is green") Irish : Ní mímhacánta fear feasa. ( = „a wise man is not dishonest") Russian :

A FIRST STEP TOWARDS UNIVERSAL SYNTAX

Y:X:iSZ:

Italian : French: Russian : Hungarian : Latin : German : English : German : French : Turkish: Breton : Scottish-Gaelic :

Z : X:j5Z:

Russian :

X : ßY Z : ßY: ßZ:

Y:]5Z: X:)SZ:

Eccolo. Le voici.   . Egy okos ember becsületes. Omnia praeclära rāra. Ich - ein Dieb ? Í Excellent ! Ausgezeichnet! Magnifique! Çok güzel! Brao! Chan duine neo-onorach ( =..a wise man is not dishonest")

:X:j?Z:

δZ:

Y : ôZ :

X:δZ:

Arabic :

German: Mich friert. Latin: Bello utendum est nobis. Latin: Duce nobis opus est. Gaelic: Chuidicheadh le Dömlmall a bhràthair. Icelandic: Honum hnignar. ( = „ i t is going worse with him") I c e l a n d i c : N ú n a er því lokið. (,.now it is finished") Russian : Ukrainian: ( = t h e general public was informed on the proposal") Latin : Laboramus pro patria. Greek: Czech: Mluvím cesky. Italian: Faremo qualche cosa di nuovo. Spanish: He leído a Calderón. Polish: Będziemy zawsze razem. Eskimo: Tikíput. ( = „they have come") German: Kommst morgen? Danish: Stol på mig! ( =,,trust me!")

103

104

δX:

PAPERS IN THEORETICAL LINGUISTICS Breton: Greek: Italian: Spanish: Portuguese : German :

Brao eo 'Ayavòv

έστιv.

E buono. Es bueno. Está bueno. É bom. Está bom. Ist schon gut.'

Our roving expedition through the basic sentence patterns of human language is finished. What did we learn from this expedition? The refractory reader will say: nothing. Because he always will. The neutral reader or the indifferent one will admit that a couple of crucial misunderstandings in traditional syntax call for correction. The positive and interested reader will, as I myself, feel induced to avow that he knows a lot more now than he and I knew a fortnight ago, and endlessly much more than any logical and psychological metalinguistics can tell us. Let us sum up just a few results of our purely linguistic analysis: 1. If we still don't know what a sentence is, we know now what a sentence can be. 2. We must realize that a finite verbal predicate can be a subject (cf. γX). 3. We know a lot more about what a vocative is and what an imperative is. 4. We realize that human language operates with a set of special sentence cases. 5. We see that sentences without a finite verbal predicate are of different kinds. 6. We notice that sentences without a subject are of different kinds. 7. We advert to the fact that if a sentence has two finite verbal predicates this may be due to widely different semantic requirements. 8. We take note of the fact that if a sentence has got two subjects this may be due to completely different semantic requirements. 9. We must distinguish sentences with an implicit subject from sentences without any imaginable subject whatsoever. 10. We must distinguish sentences with an implicit finite verbal predicate from sentences with a semantically blurred-out implicit predicate. One thing ought to be quite clear by now: ARIsTOTELian logics will never serve our turn. Any linguistic theory which is based upon ARISTOTELian, Cartesian or any other logics is foredoomed to failure. The outcome of such a theory will be false truisms and uninteresting tautologies. Some modern linguists believe that our sentence type ßZ generates all other sen­ tences. What entitles them to assume that this sentence type has precedence of all others? As a linguist I cannot with good conscience give any priority to any special sentence type. They all have an equally just claim. Every one of them is a possibility. As possibilities they constitute the apparatus of the cerebral mechanism of human speech. Every language operates with a specific set and distribution of possibilities; it operates with a specific set and distribution of syntactic norms (of possibilities) as it operates with a specific set of cases, times, modes, genders, phonemes, intonation patterns, etc. — The rest is polarity and semasiosyntactic subdivision into status forms vs. constitutional sentences. 4 For more detailed information on these two groups of sentences, cf. Niels DANIELSEN: An Essay on Nomos and Human Language, Copenhagen 1976, p. 152.

THE CASES UNRAVELLED (OR: BACK TO LINGUISTICS!) The mere formulation of the title of this article shows two things clearly: 1) the authors special love and care for his subject which, as time has shown, is not an easy one, and 2) his more-than-willingness to try again, in a last desperate endeavour, to link together the more outstanding points of recognition harvested during many a tough research period within the last ten or fifteen years. I think that it is rather a token of invanquishable entelechy than an unforgivable treading waters to challenge again, and hopefully for the last time, an epistemological quarry of knowledge whose stories seemed for a long time appeased and well-disposed. I have seen, during the last few years, the enormous advantages of investigating basical semasiosyntactic problems in the light of Qrelations (sentences without a subject). This is exactly what our forefathers forgot to do. They all believed that a human sentence, in order to be a proper sentence in human language, must have a subject. Every now and then they had been warned by people who had not ever been taken in by this Aristotelian (and Chomskyan) hypnosis and who had again and again pointed at the subjectless sentence as the tertius inter pares (alongside with active sentences (S-relations, more or less) and passive sentences (R-relations, passive in their nature)). The Qrelations seem to give us new tools in our exploration of the human cases never quite ready to loosen the grasp they have got of their never tiring persecutor. A couple of days ago this challenge was enhanced into an almost insupportable state of confrontation when in my daily Icelandic

106

PAPERS IN THEORETICAL LINGUISTICS

newspaper I noticed two almost alike sentences on page 28. This latter state of affairs is of no importance whatsoever, as everybody can see, but the emerging of the two sentences and their baffling impact on the reader are more valuable in this connection. There were less than ter lines between them. The first one read:

(= I thought of my father) and the second one was worded in the following way: (= I also thought of my mother) In the course of less than seconds every rest of localistic superstition still remaining in my mind frittered down to zero. Both these sen­ tences are expressive of equally good Icelandic. They both mean practically the same. They are uttered with the same ease in any part of the 'Land of Ice', and they convey more or less the same sentiments with regard to the expressed act of thinking. But what is dative in the first sentence is nominative in the last one. You let your thoughts go to the shores of Old Hellas. There you found parallelisms of an equal imprint. I am thinking of sentence pairs like

(= I have worked well) (nominative!) There are at least three thousand years between our two sentence pairs. Their structures are amazingly alike. They both smash any belief

THE CASES UNRAVELLED

107

in localistic case theory. The resemblance in clothing cannot possibly be accidental. It seems to say something extraordinarily important about the structure of casing-mechanisms. With the help of another two North-Western European languages we should be able now to set up a simple initial paradigm for our following - equally simple - case considerations. Let us take a German example as it looked more than thousand years ago and confront it with its modern German version. And let Irish be the garniture of the linguistic arrangement: 1) Briseadh é. (= 'he was broken')

(Irish)

2) Mih ist ës wuntar.

(OHG)

Dessen nimmt mich wunder.

(NHG)

(= 'it astonishes me') 3) Mér varð hugsað til föður míns. 4) Briseadh air. (= 'he was defeated')

(Icelandic) (Irish)

Our 4 (5) sentence specimens, all taken from a geographically strongly limited area, that of North Western Europe, reflect, so to speak, all human language structures. They have one extremely important thing in common: they are all Q-relations. They don't have the faintest shadow of a subject. They all have an (impersonal!) verb which is only keen on one thing: to govern. It doesn't care about subjects at all. Instead it runs into all sorts of nice objects. This state of affairs is the norm in some languages, it is widespread in several others, it is wellknown in a lot of tongues, and no human language is completely free of this highly interesting phenomenon. So our sentence No. 1 (from Irish) has a finite verb, impersonal in its nature, governing a direct object expressive of apolepsy, i.e. concerned only to take or grasp what runs into it. In sentence No. 2 we have an archaic state of affairs confronted with a modern one. The object is in the genitive in both sentences. The grasp of sentence No. 1 has obviously turned into a

108

PAPERS IN THEORETICAL LINGUISTICS

clinch in this specimen. What has happened between OHG and NHG is obviously that the original nucleus [BE] has been substituted by a more modern [TAKE] and that the original genitive object (metasta­ sis) ës being without possibilities of survival in NHG, leaves it to its partner dessen to express the metastatic idea of integration. In our 3rd sentence a dative emerges, most unexpectedly, some will say. The main thing is that it is there and that it is quite O.K. that it is there. It is, like any other dative, expressive of syntony

(Daniel-

sen 1980:129), and it is highly questionable, whether the syntonicity in question is rather possessive or π-related: does it refer primarily to the verbal π (the nominal entity of the sentence verb, or does it flirt with the 'father', or both? Our last example (sentence No. 4) is highly interesting: it only differs from sentence No. 1 by the fact that the object has turned into an ordinative. The verb hasn't undergone any change. An ordinative is any object case which is not an accusative, a genitive, or a dative (or still less a proessive, cf. below). The cases considered so far (the accusative, the genitive, the dative, and the ordinative(s)) are the governed cases:

(= 'he was broken')

scopos (O1 ): apolepsy

THE CASES UNRAVELLED

109

Dessen nimmt mich wunder.

Beside the four governed cases human language has at its disposal a consolidating case (the nominative) (the 'investor'), and an energetic couple of cases: the ergative and the instrumental. By using the investor expedient (or the 'nominative resort') you turn a Q-relation into a corresponding S-relation, thereby coordinating the sentence verb (the ) with an investor:

110

PAPERS IN THEORETICAL LINGUISTICS

Ihm wurde von seinem Bruder geholfen,

(Q-relation)

Er wurde von seinem Bruder unterstützt.

(S-relation)

By introducing nominatives, very often in the beginning of the sentence, you obtain a sentence structure known to most of us as 'the sentence'. By introducing ergatives in a Q-relation you get a sentence structure which is still astonishingly unfamiliar to the great bulk of contemporary linguists (Anderson 1971; Fillmore 1968; Heger 1966; Helbig 1973; Jakobson 1936; Kuryłowicz 1960; Robinson 1970); we know it, you and I, from languages like Georgian, Bacbian, Basque and many others. Some famous European Q-relations: a) the subjectless Polish constructions built on the impersonal  in -ono: widziano męzczyznę na drodze (= 'a man was seen on the road' (in the Polish sentence the 'man' is the object)); b) the corresponding Finnish constructions with the impersonal  in -ttiin: avattiin ikkuna (= 'the window was opened' (the 'window' in the Finnish sentence is the scopos of a Q-relation; although it is in form nominative, Finnish grammar­ ians are forced to call it accusative II (!))); 3) the absolute German constructions with an accusative depending on an (infinite!) perfect participle: den Arm emporgehoben, trat er zur Tür herein. And a lot of others from Ukrainian, Estonian, Icelandic, and others, e.g. Urdu: Pwlis ne cor ko

(Urdu) (impersonal)

(= the police caught the thief/ the thief was caught by the police) In this Urdu sentence the energetic case ergative has been introduced in a Q-relation. It is more unwilling than willing to invest anything in the sentence  (the sentence verb). In a North Western European language like Scottish-Gaelic the role of the ergative is tacitly taken over by an indication:

THE CASES UNRAVELLED

111

Chuidicheadh

(=(was) helped by Donald his brother) The indication le Domhnall is expressive of ergativity in this simple Qrelation. In English we are forced to say either:

or:

Donald has helped his brother.

(S-relation)

(His) brother was helped by Donald.

(R-relation)

in which case the diathesis of the S-relation changes into an R-relation diathesis (the passive), and what ought to be an instrumental is represented by an indication (i:

) just as we saw it a moment ago in

connection with the ergative. The ergative is the case of the initiator (instigator) first of all. Although, in some languages even non-human beings may be put into it: Zuen migek ez duen chilintcharik.

(Basque)

(=(by) our heifers not are worn jinglebells) Muthikoak erran daut. (=the boy told it to me)

The two energetic cases, the ergative and the instrumental, differ only in their relational affiliations: the ergative belongs to the sphere of the Q-relation and doesn't counterbalance with anything else syntacti­ cally. The instrumental is a typically supplementary entity: it counter­ balances with the subject of an R-relation, it is the epistasis corre­ sponding to the stasis of any such R-relation:

112

PAPERS IN THEORETICAL LINGUISTICS The father loves 3ohn.

(TR-S)

3ohn is loved by the father.

(TR-R)

The instrumental in many languages which mirror the universal order (Russian, Canarese, etc.) is represented in English and a lot of Western European languages by an indication for the epistasis (i: ).1 These are the main criteria of the essence of an instrumental: it is energetic (and therefore often takes over for the ergative in languages which are without the latter), and it is an epistasis case (which leads to the often observed phenomenon that the TR-epistasis wanders into other trans­ versal relations, even S-relations and constitutes 'false epistases' (instrumental objects). We know this from Russian, eg.:

(=he leads the army) For an epistasis 2 in a Q-relation it is quite normal to choose an instrumental instead of an ergative in such languages which are ergativeless. Allow me an example from Russian:

(= he was killed by the lightning, or: the lightning killed him) The result of all this is that we get the following constitutional case theory valid for all human languages: In human language there are: 1) stasis/epistasis cases: ergative (case of the epistasis 2) nominative (case of the stasis) accusative (case of the TR-S epistasis)

THE CASES UNRAVELLED

113

instrumental (case of the TR-R-epistasis) essive (case of the existential epistasis) translative (case of the process epistasis) (cf. Finnish) 2) convergent cases: genitive

(the integrational case ← the case of the πrelativity) = the case of integration

dative

(the contingency case resp. the case of the Πadherence) = the case of syntony

proessive

(the case of hypersyntony (proherence))

3) transactional cases: all other cases ('ordinatives') (The partitive is excluded from this sober company as it is obviously a purely sememic (or semantic) category). Very little can be added to this case sketch. Everything seems to be in a tranquil universal order, and no single language seems to be able to break the harmony established by the transversal relations. The evidence provided by them is very strong because they represent a theory based on real living languages. A couple of points, though, need some clarification: 1) The notion of π-relativity. We consider this idea the uppermost principle in any theoretical case hierarchy. We have to go to a couple of non-North Western European languages in order to sort it out. Then, when it has been sorted out, it will be of an invaluable use for any case analysis. In Greenlandic, any noun being the 'subject' of a transitive finite verb ends up in the genitive. We find exactly the same genitive in Lithuanian in the corresponding TR-R ('passive') topos: the 'agent' of any Lithuanian passive construction automatically ends up in the genitive. How come? As mentioned many times before in several other connections, the

114

PAPERS IN THEORETICAL LINGUISTICS

work = do + π (= work) harm = do + Π (= harm) love = do + π (= love) etc. This became clear to me only after I had studied languages like Turkish, Persian, and Urdu, where extremely many verbs are expressed in their 'simple' form in exactly this way (using the auxiliary 'to do' + a (verbal) noun). It is only after many years of such exotic studies that you find out that English is structured both in a Danish way, by tradition prevailing in our Danish school-system, and a Turkish or Urdu way. You have both 1)

[work] (in 'simple' neutrally enunciative sentences)

2)

[do

work]

(in (less 'simple'?) interrogative and 'negated' sentences) So you get: 1)

He works every day. (with the  [work])

and 2)

Does he work every day? He does not work every day. (with the  [do work]) 0+ π

The Π we call the nominal content of the sentence verb. This Π, like

THE CASES UNRAVELLED

115

any other nominal element tends to establish integrational corpora of the sort: Z

H

mother's cup (H) der Hut des Vaters (∑)

consisting of hegemonies (H) and satraps (∑) to form delimitations, A well-known delimitation is amor Dei, used very much in our schools to demonstrate the difference between a subjective genitive and a corresponding objective genitive. The Greenlandic genitive in the subject position mentioned above is a product of π-relativity and mirrors the nature of ail 'subjective genitives': Angutip qingmeq pitugpa. [the man the dog he binds it] (= the man binds the dog) The integrationai ('genitival') relation which always exists between the stasis ('subject') of a simple S-relation and its π of the  (the nominal element of the sentence verb) is our most subtle mirror of the socalled subjective genitive. So, if amor Dei means [God's love] it is reflected semasiosyntactically in an extremely simple sentence like God loves John (I) (π = H) If we change the diathesis and turn round the sentence or, to say it very simply, if we turn it into a passive construction (if we transpose the S-relation into a corresponding R-relation) this in no way damages the subjective genitive and its syntactical mirroring; it is now the

116

PAPERS IN THEORETICAL LINGUISTICS

'agent' (the 'logical subject') which turns into the (genitival!) satrap: Jonas yra Dievo labai mylimas. (∑)

(Π

= H)

corresponding to:

John is loved very much by God. Accordingly, any objective genitive is mirrored in genitive objects (cf. amor Dei = [love of God] ): John loves God. (Π = H) (Z) As we see, the genitive is a highly ideal case for objects. And subjects. It is so ideal that it suffocates in most languages because of this ideal state, its own latent self-contradiction, its curse. The nexus furnishing is chosen instead, when necessary, and elsewhere specific object cases are introduced. 2) The nature of the transversal relations. This point, again, needs some clarification. Transversal relations are sentential sequences built up on   (more rarely on a missing ) and its fellow players (subject, objects). There is normally not more than one subject in a sentence, and there are never more than two objects. This is the glaring limitation of our communicative instrument. A TR, once uttered, strives for the next one, for the simple reason that it will be killed very quickly by its own idiosyncracy. It can hardly have more than one subject (if any!), and it cannot possibly have more than two fellow players in the other end in the form of objects. Every time you want to cross the limits set for you by the inventor of language, you have to embark on a new transversal relation.

117

THE CASES UNRAVELLED

But a TR is an artful guy. It may undergo marvellous internal metamorphoses and assume shapes that even our creator had hardly quite foreseen. I think of the combined transversal relations. Nobody seems to have seen through them. Every time we are told by our grammars about sentence types which put their objects in the nomina­ tive although all sorts of object cases are available, we can be sure of one thing: it isn't the cases that have gone crazy, or the syntax that has grown mad, or the grammar that lets down its true servant. In such cases it is a combined transver$al relation that seems to take us in. Combined transversal relations (TRs) are all such sentences which start as a given TR and then fizzles out in the form of a different one. So any sentence which starts as an S-relation and ends as an R- or Qrelation becomes a TR-SR resp. a TR-SQ. Equally a sentence which starts as an TR- or Q-relation and ends as an S-relation becomes a TRRS or TR-RQ resp. TR-QR or a TR-QS. Let us look at some examples. First of all, the three simple transversal relations: TR-S:

Der Vater unterstützt den Sohn.

TR-R:

Der Sohn wird vom Vater unterstützt.

TR-Q:

Dem Sohn wird vom Vater geholfen.

And now for their combinations:

TR-SR:

Domus fit a Patricio.

(Latin)

[the house comes about by Patrick] This sentence from the European Lingua Franca discloses a TR-Ssubject (= stasis) and a TR-R-epistasis. In the form of a diagram it looks like this:

118

PAPERS IN THEORETICAL LINGUISTICS

cf.

English:

They are rather impressed by his stubbornness.

German:

Ich bin ganz besessen von seiner Musik.

Greek: (= the soldier was captured by the enemy) Russian: (= she is considered my friend) (cf. Arabic:

Qadi

mina

fì j-junaynati. (= we were pleased by the revela­ tion of the dirhams in the garden)) TR-SQ:

Es gibt viele Probleme.

In the form of a diagram our German sentence would look like this (the stasis is that of a TR-S, and the epistasis is obviously that of a TR-Q, i.e. it is a scopos!):

) cf. French: Il y a des problèmes partout. Il faut chanter une chanson. In Russian, for example, we find TR-SQ-specimens built up on TR-Qepistases 2:

(= I detest alms) The structure of this sentence is:

)

119

THE CASES UNRAVELLED

TR-RS:

The boy was taught a lesson. Manden berøvedes pengene.

(Danish)

In this case we have the following sentence structures: TR-R for S and TR-S for 1 :

This combined transversal relation is exceedingly widespread in human language. We find it in Greek: (= the tyrant was conquered in the battle) Japanese:

Watakushi wa sori ni tokei o suraremashita. (= [I have been robbed of my watch by a pickpocket] i.e.: I have had my watch stolen by a pickpocket)

Our last sentence (from Japanese) is an example of a so-called conglomeration. We know them from Classical Greek, for instance:

(= the soldiers had their eyes spoilt by the snow) Both in the Greek sentence and in its English traduction we find a beautiful representation of a conglomeration:

120

PAPERS IN THEORETICAL LINGUISTICS

As we see clearly, our conglomeration is defined by having two epistases: one for the TR-R and one, demanded by the , for the TR-S. TR-RQ:

Es wird von den Studenten gesungen.

Der synges af studenterne.

(Danish)

The subject ('es') in the German sentence is semantically shallow; it does not mean anything, because it does not refer to any given object in reality. It might do, only then our TR-RQ turns into a pure TR-R:

Es wird von den Studenten gesungen.

But in the German sentence used by us to show the nature of a human TR-RQ the 'subject' 'es' means nothing. It must be there, though, because its role is a fixed one in TR-RQs of this type. We may draw the sentence structure, once again, and we get:

The stasis (the 'subject') is es (of an R-relation), and the epistasis is that of a Q-relation (a so-called epistasis 2).

TR-QS: The best example we have for our purpose is from Classical Greek:

(= we must do this)

THE CASES UNRAVELLED

121

This sentence is primarily a Q-relation, because it has no subject. No discussion: the combined TR we have in front of our eyes is a TR-QS. If we draw it we see it clearly (the  is from a Q-relation, and the role of the (epistasis 2) has been taken over by the  of an Srelatipn).

Another instance of this rarely discoverable and highly interesting combined TR is offered to us equally by Classical Greek:

(= we must do this) In this last case, a nucleus

has been introduced. It expresses

obligation, and like so many other modal nuclei in a great number of human languages it is incapable of taking a subject. The diagrammati­ cal representation of our last sentence from Greek looks like this:

TR-QR: Some people say that (North-)Western Lithuania is part of North­ western Europe. Or at any rate part of Lithuania. Lithuanian is perhaps the only European language which offers us a convincing TRQR: laikoma arklys. (= they keep a horse)

122

PAPERS IN THEORETICAL LINGUISTICS

'They'in this sentence is in the genitive (!). The horse, arklys, is in the nominative (!!). But we should never forget that it is the horse which is being kept. This does in no way change the shocking fact that the sentence verb (the ) is in the neutral gender (!!!). The neutral verb gives preponderance to the Q-relation. The fact that it is the horse which is being kept and that it is in the nominative makes us recognize in it the stasis of an R-relation. The genitive

is

no problem, really. It is the quite normal genitive for TR-R and Qepistases which is so characteristic of Lithuanian: the genitive is the case in this language which expresses the 'agent' in 'passive' construc­ tions. So our Lithuanian construction was perhaps not so amazing as we thought at first. It looks like this:

It is matched by sentences from the flower garden of languages (Caucasus). Let us look, for a moment, at an example from the King of Languages (Georgian):

[= by the students wrote letter] Here the epistasis 2 of the Q-relation is in its specific case (the ergative), the letter is in the nominative (think of the Lithuanian horse), and the verb -? The verb, or the , is split up in a funny way functionally: it invests the number of the ergative. This leads us to our first fission: we have to indicate in our diagram of the sentence that its Π has become over-exerted by drawing a line through it to indicate the border between the two transversal relations which constitute the syntagm:

THE CASES UNRAVELLED

123

This syntactic structure, again, reminds us of the normal syntagm for grammatically correct transitive constructions in Greenlandic:

[the student

a letter

he wrote it]

The student is in the genitive. No wonder, so were 'they' in Lithuanian. The letter is in the (inactive) nominative. No wonder, so was the horse in Lithuanian and the letter in Georgian. The Eskimo verb is an alloy (it tells us by its mere form who is doing what in the sentence, <

= [he wrote it]). Isn't this fantastic: the Lithuanian

confirmed once more in its

capacity of genitive! But we had to walk a long way to explain it. It doesn't become clear until you see the whole complex through Geor­ gian and Eskimo glasses. The rest is rection. This is what most case researchers overlook completely. Indicators (prepositions, postpositions, circumpositions) govern cases. What else should they do, if they have the chance, than to govern some governed case? It sounds like a bad joke, but it isn't. This is exactly what happens in all the languages which have rectional repartition problems. The governed cases are repartitioned according to certain universal main rules and with regard to certain more or less provincial school rotes, ham-fisted, from a grammatical or a linguistic point of view ham-fisted, but pragmatically pragmatic and functioning to a certain degree. . If we look for grammatical main rules concerning indicators and the cases they govern we can easily find them. The first rule is very easy. It goes something like this: in languages without cases indicators govern no cases. The second rule is this: in languages with more than

124

PAPERS IN THEORETICAL LINGUISTICS

one form for the noun the indicators govern governed cases. Or their constitutional fellow cases, their challengers, so to speak, outside the mesonomic sphere, i.e. any case in the ectonomic part of the sentence, or: anything bat the nominative. If you have special governed cases of the substantive, why shouldn't you use them in any position where the substantive is governed? The governing takes place according to fixed subsets of rules. In Arabic, for instance, all indicators select one and only one case (the genitive): Mithlu hādhā l-camali

ghayru

mumkinin

like

other

possible

this

thing

except fi

1-mujtama'i

l-carabiyyati.

[in

the society

the Arabic]

(= a thing like this is impossible in Arabic society) Kānat

qalcatun

fauqa

1-jabali.

[there was a fort

on top of

the hill]

Wasalati

1-banātu

wa

jalasna

[arrived

the girls

and

sat down

in the class]

This is an extremely simple state of affairs. And yet, there are thorns in the garden: the deictic elements do not follow the rule (cf. English where the deictic elements 'I', 'you', 'he', 'she', 'we', 'they' follow their own rules when governed by verbs or 'prepositions'). This is an enormously important main rule: the deictic elements often follow their own rules. In Arabic they refuse to be governed in the genitive like other nominal or pronominal elements. Instead their 'pronominal suffixes' melt together with the indicator leaving special ordinatives for a whole series of indicators:

125

THE CASES UNRAVELLED

'Iltafit 'ilayya! ( = pay attention to me) Wa alū minhā. ( = they arrived from it (her)) Hasibtuhu jāhilan. ( = I considered him ignorant) Prepositional con-fusions are known from several

North-Western

European languages (Irish, Manx, Scottish-Gaelic, Welsh). The onecase-governing-system is known from a lot of languages. Latvian uses it in the plural where all indicators govern the instrumental. In the Latvian singular a fixed repartition of cases takes place so that some indicators select the accusative, others take the dative, and yet others take the genitive. Bērni skrēja gar veco māju.

(acc.)

( = the children ran past the old house) (gen.) ( = this garden is after my taste) (dat.) ( = he is covered up (up) to (his) neck in mud) We have mentioned this unforeseen phenomenon elsewhere before (Danielsen 1980:85), and we shall spare the reader a detailed account of it here. German and Icelandic behave like Latvian singulars: a repartition of governed cases takes place. The genitive is found with indicators who play the role of satraps with regard to the element governed (the indicators thereby often represent old or still existing nominal ele­ ments in the general case (Schmidt 1967:132-34):

126

PAPERS IN THEORETICAL LINGUISTICS

Laut dieses Gesetzes haben Minderjährige kein Recht auf Arbeit. Zeit seines Lebens arbeitete er in den grossen Fabriken. Dank seines beträchtlichen Vermögens studiert er unent­ wegt klassische tibetanische Philologie an den vornehmsten buddhistischen Hochschulen in Lhassa. Statt dessen solltest du dich strikt an die geschriebenen Worte halten. Inmitten des Waldes wohnt ein alter Holzhacker. Wegen seines Handicaps vermag er sich kaum zu wehren. The repartition of accusatives and datives correspond to what we know from the transversal relations: the accusative is the apolephthical case, i.e. it is merely concerned with the mere taking (or taking over) function. So it is ideal for expressing the goal. And the dative expresses syntonicity, the striving together of elements. Its special capacity of expressing 'Ruhelage' therefore seems clear. If we confer our German examples with evidence from Icelandic, nothing much can shake our theoretical considerations in their essen­ tial and basic structure. Let us look at some accidental Icelandic syntagms taken from an Icelandic newspaper and my Icelandic infor­ mant: 1)

with the genitive: Bandaríkjamenn eru við öllu búnir vegna gervihnattarins. (= the Americans are fully prepared because of the satel­ lite) pað er ómögulegt að lifa án peninga. (= it is impossible to live without money) Við förum norður í stað þess a5 f ara suður. (= we shall go north instead of going south) Í miðju skógarins bjó gömul kerling.

THE CASES UNRAVELLED

127

(= in the middle of the wood there lived an old hag) Sunnan islands eru góð fiskimið. (= south of Iceland there are good fishing banks) 2)

with the accusative: þú mátt ekki stökkva upp á borðið! (= you mustn't jump onto the table!) Hann ók inn í bæinn. (= he drove into the town) Hún skreið inn undir stólinn. (= she crawled in under the chair) Núna er flugvéiin úti y f i r hafinu. (= now the aeroplane is out over the sea) þeir hlupu í gegnum stofuna. (= they ran through the room) with the dative: Hann bjó í sveitinni a l l t sumarið. (= he lived in the country ail the summer) Bollinn er á borðinu. (= the cup is on the table) Að húsabaki sá ég gamlan hest. (= behind the house I saw an old horse) Kötturinn liggur undir rúminu.

(= the cat is lying under the bed) Hann var5 fyrir miklu aðkasti. (= he was heavily criticized) Icelandic differs cardinally from German in one point only: in the way in which the dative, i.e. the pure formal dative takes over a lot of functions which are l e f t to the instrumental in very many languages. We should never forget that a t o t a l syncretism of the dative and the

128

PAPERS IN THEORETICAL LINGUISTICS

instrumental is observable in a good deal of human tongues, e.g. in Japanese: Taro ga watashi ni Nippon-go o oshiete-kuremashita. (= Taro taught me Japanese) Dorobo wa keikan ni tsukamaerareta. (= the thief was caught by the policeman) As soon as we see this it is easier for us to reconcile ourselves with what we see in Icelandic: Hann hafði framið sjálfsmorð eftir tilræðið með því að skjóta sjálfan sig tveimur skotum í höfuðið. (= he had committed suicide after the assault by shooting himself with two shots through the head) HafÔi árásarmaðurinn smám saman fyllzt hatri í garð þessara manna. (= the attacker had, little by little, been filled with hatred against these men) Öllum frístundum sat hann heima hjá ser. (= during all his free hours he sat at home in his own place) In all these instances the instrumental dative is purely adverbial. We know parallel situations from Latin: Bello utendum est nobis. (= war must be used by us) In my own school we were taught that nobis was a dative of the agent (an instrumental dative). Maybe it is. But one might as well be tempted to see in it a quite normal, undeviated  (diastasis) dative 2

([to us the implementation of war is a must]). The dative is in an

129

THE CASES UNRAVELLED

enormously exposed position as regards instrumental deviations. This is obviously a universal law. Its impact on Japanese is total, its conse­ quences are not quite so strong in Icelandic. But it is very productive in this language, though it be partial. The discussion of the Icelandic dative serving as an instrumental leads directly into the conclusive complex of our case of cases. As mentioned above, the instrumental dative in Icelandic is adverbial in its basic function. If it is escortedly (not purely functionally) adverbial or not adverbial at all, it is supplied with an indicator: 1) insulatedly adverbial: Hún var furðu lostin. (= she was wonderstruck [struck with amazement].) 2 2) escortedly adverbial: Hann sló hana meó hendinni. (= he beat her with his hand) 3) fellow player in a transversal relation (=

in a TR-R):

(Bréfið skrifaðist af manninum.) (= der Brief wurde von dem Mann geschrieben) Our last instance is especially interesting and actual at this point of our elucidations. The man' in this sentence is a purely grammatical fellow player in a transversal R-relation. In the good old days we would simply say: it corresponds to a plain subject in the corresponding non-passive construction: A)

The man wrote the letter.

B)

The letter was written by the man.

It is that simple. All that happens is a topical change of fellow player 1 and fellow player 2 from A to B, accompanied by a corresponding change in diathesis (from the active to the passive). Excuse me, but I can't say it much simpler. It is already trivial. The epistasis of TRs is replaced in some languages by an indication. Indications for stases and epistases (end-points in simple transversal relations) we have often

130

PAPERS IN THEORETICAL LINGUISTICS

called representations, and we see no reason to change this suitable term. Some languages have representations in all possible positions: Watashi wa Tai-go ga dekimasen. [= I Thai (is) not susceptible to speaking] (= I can't speak Thai) Nippon-jin wa kore o hashi to iimasu. (= Japanese people call these chopsticks) Keikan ga dorobo o tsukamaeta. (= the policeman caught the thief) Several other languages have representations in R-relations and to differing extents in S-relations:

Er wird vom Vater bestraft. She was jilted by the young man. Nu-am văzut pe Ana.

(Rumanian)

(= I haven't seen Anna) Quiero a la muchacha española.

(Spanish)

(= I love the Spanish girl) 'Ahsasnā bi-surūri fi qulūbina.

(Arabic)

(= we felt joy in our hearts)

(= the government has finally confirmed the laws of the youths) (No single Arabic TR-relation with an expressed

(epistasis) can be

considered genuinely Arabic). Equally, indications for genitives, datives and proessives are called substitutions. We know them from all North-Western European languages:

THE CASES UNRAVELLED

131

This is the rose of the gaudy night. I hate the callous look of his eyes. He resented the gaunt complexion of her face. We detest the sallow deliberation of certain contemporary Polish politicians. He gave the book to the small boy. Dette er mindesmærket for de faldne.

(Danish)

The appearance and the distribution of substitutions for the genitive, the dative and the proessive is an enormous chapter in linguistics still in the shape of an incentive for those willing to cope with it. If we compare the following simple Unkings with each other: das Denkmal des russischen Dichters = the monument of the Russian poet = das Denkmal vom russischen Dichter = monumentet for den russiske digter an enormous field of research opens in front of our eyes with a lot of hopefully not too trivial answers to a heap of good questions deprived of the imprint of truism. Substitutions are more or less widespread in all North-Western European languages. They are symptomatic of English (← Romance) syntax. It is wrong to look at them as cases, as for instance Hjelmslev (1972:114ff.)

did. They are equally symptomatic of German to a

certain degree: Er schreibt einen Brief an die Mutter. (= Er schreibt der Mutter einen Brief) Er schreibt einen Brief für die Mutter. (= Er schreibt der Mutter einen Brief)

132

PAPERS IN THEORETICAL LINGUISTICS

(Who would dream of considering 'an die Mutter' and 'für die Mutter' as specific case forms in German? Any such misuse of terms would lead to inconsistent considerations and 'jumble theories'). Dies ist der Rucksack vom kleinen Jungen. Wo liegt die Domkirche von München? Ich hasse die schrullenhaften Überlegungen von diesen seichten Pragmatikern. Constructions like these are not especially popular among German teachers in any country. But they are very popular among Germans in the sense that they are used several billion times during a day within the German-speaking community. Substitution in Icelandic is even less widespread than in German. A dative construction is often replaced by a construction with 'til': Hann skrifaði föður sinum bref. = Hann skrifaði bref til föður sins. but an indication for the genitive is almost unthinkable of in Icelandic. Substituting languages are the Romance tongues, first of all: le livre de mon père la camera di Pietro el amigo de la estudiante araba a gramática de Maria Il écrit une lettre à son père.

etc.

and where the indication fuses onto the following (pro)nominal ele­ ment, we see all sorts of substitutional ordinatives: le livre du père le livre des garçons

THE CASES UNRAVELLED il libro delle ragazze el libro del muchacho o livro da mulher brazileira Il écrit une lettre aux jeunes filles.

133

etc., etc.

All other indications for fellow players in transversal relations are transactions. A transaction is the indicational rewriting of any other case than the nominative (the case of the stases), the accusative (the case of the epistases), the dative (the case of the diastases), the genitive (the case of the metastases), the proessive (the case of the hyperstases) and the instrumental as the epistasis case of TR-Rs. One example for thousands and thousands and several more thousands: ő nevetett a tanulón.

(Hungarian)

Er lachte über den Lehrer. He laughed at the teacher. Han lo ad læreren.

(Danish)

The Hungarian superessive (tanulón) is rewritten in very many lan­ guages with the help of different indicators. In Hungarian the super­ essive is demanded by the verb nevetni as the object case (apostasis =  ) of the latter. I.e.: any 

which appears in the form of a specific case (=an

ordinative) must be rewritten in languages not possessing ordinatives in the form of a transaction. The end of this article is drawing near. A lot of detailed questions couldn't find their place on the record. And still more answers could not be given. They will be given in future, but their solution demands a collective will to go into research and not give up until the theory has

134

PAPERS IN THEORETICAL LINGUISTICS

been completely internalized into its physical railments (of a more phonological nature). Let us take one detail, for a moment, and look at it with our new glasses. An old lady asked me the other day: 'Why do verbs always take the genitive in Russian sentences when these are negative?1 My answer was that not only Russian verbs behave that way but equally the verbs in a whole series of Slavonic languages (Polish, Czech, Kashubian, etc,). The reason for this is that evidently negati­ vity is furthering integration, and consequently the heart function of integration, i.e. π-relativity. We see this in very many languages: Il n'a pas de copains.

(French)

Nie ma pieniędzy.

(Polish)

([not has money] = there's no money) (Russian) (= I don't like Poles) It is a new way in which to explain an old trivial problem. Excuse me, did I say 'problem'? Was it at all a problem, as long as you could just say that in negative sentences all Russian verbs take the genitive - it is something which you shouldn't bother about, because that is how it is, and don't think too much! 'Don't think too much' is the motto of all too much linguistics in our days. It is only surpassed by the motto of pragmatics, now demanding all the academic rights which formerly belonged to any and all humanistic fields of study. With pragmatism the greatest antihumanistic monster has grown into a roaring dragon, and we hear it and we see it formulating itself every day. 'Don't think at ail', it says. 'Don't learn anything more, speak the gospel of the shoal waters, and make thinking without thoughts a sign of nobility!' Berdyayev (1958: 318): No, we say, no, no. Back to linguistics! If we give you the right to say and write things without vision and depth, please let us keep our

135

THE CASES UNRAVELLED

old human rights to broaden out our knowledge and make our (not your) ideals richer. Let us keep our Greek, and Russian, even Arabic alphabets. You are the deniers of our age (cf. Goethe's Faust: 'Du bist der Geist, der stets verneint'). We know why you resent people knowing the wonderful alphabets. People knowing something. People knowing the pleasure of knowing. Rather than that of not knowing. What we should know now, then, would be to disentangle our cases (= Kasus) in a hitherto unseen manner. We should now be able to say a lot more about the coalescence/non-coalescence of cases (seen diachronically) than we were before. If we look at our German paradigms, once again, we see something very clearly: masc.

fem.

neutr.

der Hof

die Burg

das Schloss

den Hof

die Burg

das Schloss

des Hofes

der Burg

des Schlosses

dem Hof(e)

der Burg

dem Schloss(e)

In the feminine colomn we see a very productive syncretism ('contrac­ tion') of the genitive and the dative. We know it from several other languages, just think of Armenian or some of the Balkan languages: Zonja e së mbrëmjës së së djëlës ishin të mira.

(Albanian)

(= the women of that Sunday night were good) Nuk i përgjigjet së mbrëmjës së së djëlës.

(Albanian)

(= it doesn't correspond to that Sunday night) This 'contraction' is especially symptomatic of languages which use 'possessive datives'. In Hungarian, the possessive dative takes over all genitive functions (assisted by a possessive suffix at the end of the hegemony):

136

PAPERS IN THEORETICAL LINGUISTICS

A fiúnak a haza nagyon szép.

(Hungarian)

(= dem Sohn sein Haus (ist) sehr schön) In many Balkan languages the possessive dative is extremely produc­ tive: Lomila su j o j se pleca.

(Serbo-Croatian)

(= her shoulders broke together)

cf. Rumanian: Fiecare

cioară

lauda puii sai.

(= every crow praises its young) Albanian: I mbeti mendja atje. (= her thoughts kept hanging there) In the following example (from Albanian) the dative vacillates between the  and the possessive function: Mbase më bënë veshët. (=einen Streich spielen mir (meine) Ohren)

One example shows us the enormous force of contraction:

Kjo tronditje shpirtërore fëmijëvet ua kishte zhdukur krejt zellin për mësimet edhe dashurinë për shkollën. (= this spiritual shock had wiped away the women's entire interest in education and (their) love for the school) The element 'fëmijëvet' is dative here. If i t were genitive, i t would have had the majeutical particle e between itself and its hegemony ('tronditje shpirfërore'). This opens a quite new and unforeseen possibi­ lity

of

explaining the majeutic

(= delimitationizing)

particles

in

137

THE CASES UNRAVELLED

Albanian. They turn out to be extremely important when you want to sort out what is genitive and what is dative. Another symptomatic instance of contraction is the following: Atyre u ishte mbushur koka se çdo mësues është gjakpirës. (= their heads had been filled (with the idea) that every teacher is bloodthirsty) In Icelandic, the possessive dative clings to its hegemony: Hann hljóp niður í göngin með tugi manna á hælum ser. The same thing happens in Bulgarian, Macedonian, and Rumanian: A

Inca din scrierile din tinerete

numeroase

referitoare la propria-i viata. (= even in his early writings we find a lot of information referring to his own life)

(Macedonian) (= my brother found me) (Bulgarian) (= I call my son)

Latvian is the 'language of contraction': Vasari kokiem lapas ir zaļas. [im Sommer den Bäumen die Blätter sind grün] (= in summer the leaves of the trees are green)

In German (and in all Slavonic languages) it is used to some extent:

138

PAPERS IN THEORETICAL LINGUISTICS Geh mir aus dem Wege! (= Get out of my way!)

So, German feminine nouns follow universal rules which do not apply to the masculines and neuters (the same thing goes for Icelandic). They are primarily of a functional (grammatical, semantic), and by no means of a phonological nature. The phenomenon is just one more example of how the most interesting (actual, basic) things in linguis­ tics are found through functional semantics. Phonology is the servant for all in a well-disposed linguistics, whether syn- or diachronic. It is long past twelve now. But when it morns, let us help each other towards a much profounder understanding of all the problems which we normally serve to our students advising them not to think too much. Don't let our universities grow into the thoughtless servants of a collective social adversary of humanistics. This last remark of mine does not belong to any purely scientific contribution of any sort. I am well aware of that. Therefore I give it to the RECORD. The big RECORD of nonforgettal. Institut for germansk filologi Odense Universitet DK-5230 Odense M

THE CASES UNRAVELLED

139

Notes 1.

Its ideal ordinative locum tenens is the ablative. This case takes over the TR-R-epistasis function in Armenian, for example:

(= the nest was built by the bird). 2.

In some languages, even 'particles' of different sorts govern the accusative, cf. Arabic: 'Inna harārata s-samsi mudirratun fi š-šarqi. (= the heat of the sun is injuring in the Orient) (in this case we see how the 'particle' 'inna (the status/po­ larity indicator of the neutral enunciative status) governs 'the heat'(→accusative). Dhukira 'anna l-malika marldun. (= it has been mentioned that the king is ill) (here the 'particle' 'anna (= that, i.e. the status/polarity indicator of the neutrally subordinated enunciative status) governs the 'king'(→accusative)). Hadara l-yahûdu lâkinna l-caraba ghābū. (= the Jews attended, but the Arabs stayed away) (in this instance, the 'particle' lâkinna (= but) governs the 'Arabs' (→accusative)), etc. Deictic (pronominal) elements representing these governed nouns in the accusative are found in their governed forms (they are suffixed to the governing particle), e.g.: Dhukira 'innahu maridun. (= it has been mentioned that he is ill)

140

PAPERS IN THEORETICAL LINGUISTICS

Bibliography Anderson, J.M. 1971. The Grammar of Case. Towards a

Localistic

Theory. Cambridge: Cambridge University Press. Berdyayev, N. 1950. Der Sinn der Geschichte.

Tübingen.

Danielsen, N. 1980. Linguistic Studies. Heidelberg: Carl Winter Uni­ versitätsverlag. Fillmore, Ch.J. 1968. The Case for Case'. In: E. Bach and R.T. Harms (eds.), Universals in Linguistics, pp. 1-88. New York: Holt, Rinehart & Winston. Heger, . 'Valenz, Diathese und Kasus'. Zeitschrift

für

romanische

Philologie 82:138-170. Helbig, G. 1973. Die Funktionen der substantivischen deutschen Gegenwartssprache.

Kasus in der

Halle.

Hjelmslev, L. 1972. La categorie des cas. Internationale Bibliothek für allgemeine Linguistik. Band 25. Herausgegeben

von

Eugenio

Coserru. München: Wilhelm Fink Verlag. Jacobson, R. 1936. 'Beitrag zur allgemeinen Kasuslehre. Gesamtbe­ deutungen der russischen Kasus'. Travaux du Cercle

Linguistique

de Prague VI. Kuryłowicz, . 1960. 'Le problème du classement des cas'. In: Esquisses Linguistiques,

Wrocław-Krakau.

Robinson, J. 1970. 'Case, Category and Configuration'. Journal of Linguistics 6:57-80. Schmidt, W. 1967. Grundfragen der deutschen Grammatik. Berlin: Volk und Wissen. Volkseigener Verlag.

A FINAL WREST: ON EXISTENCE, EXISTENTIAL DISSEMINATIONS AND THEIR DELEGATIONS On the first day God said, Let there be light, and let there be day and night. On the second day God said, Let there be a firmament, and let there be water above the firmament and below the firmament. On the third day God said, Let there be earth and sea, and let there be herbs and trees on the land. On the fourth day God said, Let there be sun, moon and stars to give light on the earth. On the fifth day God said, Let there be fish in the sea and winged creatures on the land and under the heaven. And let there be cattle, snakes and wild animals on the land. On the sixth day God said, Let there be Man. On the seventh day God saw that everything was good and that he had a good rest. The Story of Creation is the ode of being. The paean of existence. Existence was the delightful undertone of the Land of Creation. This wonderful land where such basic ideas as smell and taste were given from before the beginning, and where language was used by God even before he had any human beings to speak to. After the initial wrest of God this most marvellous planet of ours set out on the most daring expedition to last for thousands of centuries. Mankind had got its luggage with it, including the gift of language. It was a many-faceted jewel. And without limits was the spirit governing the oral utterances. Through it all was twittering the undertone of existence, from the beginning to the end. Let's do it! was

142

PAPERS IN THEORETICAL LINGUISTICS

the shrill tone of the orchestrational finale. Do what? Get it down on paper before it is too late. A final wrest! Methinks the Sanskrit grammarian Yāska was the first to say it explicitly. I am thinking of the passage in his Nirukta (written 500 years before the beginning of the Christian era) where he writes the following momentous observations on the entities of being in his own language, anticipating, as it were, Aristotle, the father of Western Philosophy: Bhāvapradhānam äkhyätam sattvapradhānāni nämäni tad yatra ubhe bhävapradhäne bhavatah

pūrvāpari bhūtam bhāvam ācaste

vrajati pacati iti upakrama - prabhrtyapavargaparyantam mūrtam sattvabhūtam sattvanämabhih vrajyā paktir iti. 1 'That which has being (bhāva) as its essential element is called the verb, those which have substance (sattva) as their essential element are nouns. Further, where both have being as their essential element, one expresses being which consists of a suc­ cession by vrajati, he walks, pacati, he cooks, and being which is as it were, a substance (sattva) formed as a body from the beginning to the end, by nouns for things, viz. vrajyā, a walk, pakti, a cooking'. This passage from the oldest known Veda-commentary was refined five hundred years later by Durga in his voluminous commentary (ed. by H.M. Bhadkamkar, Bombay 1918) where we read: 'Vākye hy äkhyätam pradhānam tadarthatvät gunabhūtam nāma tadarthasya bhāvanispattāv angabhütatvät evam tävad äkhyätam vä kye pradhänam'. 'For in a sentence the verb is the essential element because it is

A FINAL WREST

143

necessary for the sentence, and the noun is secondary because it is a member of what is necessary for the sentence in the production of bhāva ('being'). This is how the verb is the essential element in a sentence'. A similar discussion in Europe is aroused through the reading of Aristotle's Organon where we find the renowned passage:

' . . . . is the co-signifier of time, whose section means nothing in itself, and it is also always the sign (token) of what is being said about something else'. Or:

'I say that it signifies its own meaning + time, viz. health is a substantive, whereas be in good health is a verb: it joins to the meaning the fact that it is in existence now'. H. Steinthal takes up the discussion in his Geschichte wissenschaft

der Sprach­

bei den Griechen und Römern (Berlin 1863). On page 233

in this important book we read: Das

bedeutet ein

, und dieses ist nicht denkbar

144

PAPERS IN THEORETICAL LINGUISTICS ohne Zeit. Also bedeutet das *

weil es die Zeit mit bedeutet,

eben das Sein. Und so ist denn welches in jedem

das reinste

enthalten ist und es dazu macht; denn ist so viel wie

Eben darum aber wird auch jedes

zum

The Frenchman 3. Tricot writes seventy-three years later in his translation of the Organon (Aristote, Organon, I Categories, II De l'interprétation. Nouvelle traduction et notes par 3. Tricot, Paris 1936): Le verbe est ce qui ajoute à sa propre signification celle du temps; aucune de ses parties ne signifie rien prise séparément, et il indique toujours quelque chose d'affirmé de quelque autre chose (p. 81). Anything in a sentence which is said about any other part of that sentence is foredoomed to be the 'verb' of the very same sentence. So we get: He cries. He my friend. (!) She in Norway. (!) The potential ejection of the entity of existence is extremely sympto­ matic of any human verb which is finite. The ejected part thereby becomes the nucleus of the syntagm. The verb to be is the most frequent nucleus in human language. It is used as an auxiliary and as a copula, and its various shadows of meaning in modal sequences are of momentous importance. The nucleus [BE] may easily be depicted as the epistemological centre of a tricolic or a constitutional derivation of all human nuclei

A FINAL WREST

145

(cf. our chapter on human nuclei in Niels Danielsen, Linguistic

Studies,

Heidelberg 1980):

The rest is description. The description starts with a constrastive analysis of sentences gene­ rated by finite

with their, mostly Indo-European, counterparts

which contain no sentence-generating or pronominal

but replace it by an adverbial

(nucleus):

Šenho?

(= WHAT is (there)?)

(Ouaday Arabic)

Muntu uzo ndjeni?

(= WHO is this man?)

(Subiya)

Kuyini lokho?

(= WHAT is that?)

Pieu'r tŷ ?

(= WHOSE is the house?)

We . se . kew wigwam?

(= HOW is the house?)

Nauk qingmeq?

(= WHERE is the dog?)

Ci yaana?

(= WHAT do you do?)

(Mongolian)

Ci XAITSI/-?

(= WHERE do you go?)

(Mongolian)

Ce sbd?

(= WHAT happened?

(Zulu) (Middle Welsh) (Menomini)

(Greenlandic)

(Ishkashimian)

> WHAT became ...?) A tricolic repartition of nuclei basic and symptomatic of human language leaves us with the following representation of entities:

146

PAPERS IN THEORETICAL LINGUISTICS 2

machen X

2

let (go) X

put y

y

y

grow 2

tun

y

X

hold - take - keep 2

place catch

x

x scheinen

lassen y

2

go -

- werden

z  throw turn X

get

  sein - bleiben 2

reynast 

stand

2

-



X

sit - lie - hang

2

come The centre of this mapping is the [BE]-entity. It is the prerequisite and the consequence of all the other focused entities. Thus, the [BE]-entity may be evaluated as the consequence of reynast (Icelandic: 'turn out to be'), having scheinen as its prerequisite, and as the consequence of werden, having bleiben as its prerequisite. Placed in a nomical sentence pattern it becomes the ever prevailing endonomic part of the sentence:

There are three main categories of nuclei: 2

147

A FINAL WREST 1)

the sentence-generating nuclei

2)

the polar nuclei

3)

the modal nuclei

The [BE] -entity is latent in all of them. Sometimes it alone takes over the modal tournure of a sentence. It may do this 1)

unexpressed

2)

in a finite manifestation

3)

in an infinite manifestation (playing the role of a 'hermit')

to 1 (without syntagmatic representation): Gjërat e dimrit për t'i shturë. (= 'the things for the winter must be collected')

(Albanian)

Excellent! Congratulations! Ude af 0je, ude af sind.

(Danish)

(= 'out of eye out of mind') Priecïgam man dzïvot! (= 'I must live pleasantly!') (Latvian) to 2 (with a finite representation): Ada Sirilgē upan dina utsavêta yanna tiyenavi.

(Singhalese)

[today Cyril's born day celebration -to go is] (= 'today I must go to Cyril's birthday party') to 3 (with a hermit representation): (Russian)

(Here the hermit

in the form of an infinitive is

connected with the predicative instrumental).

148

PAPERS IN THEORETICAL LINGUISTICS

[him indeed such a daring person not find] (='indeed, he couldn't find such a daring person') It is a revelation to realize how much human sentences are capable of expressing without any hint whatsoever of a subject. In none of our last six examples was there the slightest shadow of a subject. An honest and genuine linguistics should refrain from presupposing sub­ jects. This applies equally to a whole bunch of human sentences with a demonstrative [BE]-entity as its case-governing nucleus: Eccolo arrivato!

(Italian)

[there is him arrived] Idhā bi-rajuli!

(Arabic)

[behold a man] (cf. Le voilà!)

(French))

In a recently published case investigation I have demonstrated how it pays off to start your analysis with sentences which are totally devoid of any subject whatsoever (the so-called Q-relations) and doing this without presupposing any sort of preceding transformation(s) (implying such phony tricks as the deletional or/and insertional operations). We should never forget that in the history of human tragedies these two jesters, the deletion and the insertion, were the ever-acting foolers of men. Let us look, instead, at some disseminations offered to us by human language, exactly as they are. And let us start with the simplest of them all, those which save the subject. 1)

Bello utendum est nobis.

(Latin)

(= 'we must use war') 2)

Tev nebūs slinkot. (= 'you must not be lazy')

(Latvian)

149

A FINAL WREST 3)

Man bija jāmācās.

(Latvian)

(= 'I had to learn') 4) Bhí orm rith abhaile.

(Irish)

(= 'I had to run home') 5)

Sinun on täytettävä velvollisuutesi.

(Finnish)

(= 'you must accomplish your duty') 6)

(Icelandic) (= 'one had to get accustomed to good service')

7)

(Russian) (= 'why did fate have to throw me into the circle of smugglers?1)

8)

(Classical Greek) [that to be done is us] (= 'we must do that')

9)

Mwjhe jana hay.

(Urdu)

(= 'I must go') The great advantage of a subject-independent verb (vP = verbal protone;  = the whole verb) of a Q-relation is obvious: it can choose its own partners. So, in 1 a dative is chosen (beside the object in the ablative), 2, 3, 6, 7 and 9 all choose datives, 5 chooses a genitive, 8 an extra accusative. And 4 chooses - an allativic ordinative! Now, if instead you choose to introduce specific modal nuclei, you can obviously do this in two ways: 1)

by sustaining the subjectless construction

2)

by introducing a subject

to 1:

Trzeba mi buty wysmarowac.

(Polish)

(= 'I must polish my shoes') (Lithuanian) (= 'the father must work')

150

PAPERS IN THEORETICAL LINGUISTICS (Russian) (= 'for this you must give him a piece of candy') Neturìnt rãkto negãlima 1 atrakìnti dur

(Lithuanian)

(= 'you cannot open the door not having a key') (Lithuanian) (= 'they too have to do work') to 2:

(Rumanian)

Trebuia neapărat respectat. (= 'it is to be unconditionally respected')

etc., etc., etc. Sa pead oskama hästi lugeda.

(Estonian)

[you must be able to well read) Sa pead hästi lugeda oskama. [du musst gut lesen können] (= 'you must be able to read well') Mēs būtu varējusi iet pastaigāties.

(Latvian)

[wir (hätten) könnend - seiend gehen spazieren wären] (= 'wir hätten spazieren gehen können') (= 'we could have gone for a walk') We are moving into a realm of tricky disseminations. 3 Diagrammatically, the last one (from German) looks like this:

('quadrusection')

A FINAL WREST in which diagram

151

stands for any infinitive serving as an

infinite part of a complex verb. German has the world record: Er soll sie kaum mit dem Italiener haben spazieren gehen lassen wollen können. In this last instance our quadrusection has changed into a sexcusection, which is really the utmost overworking of a human sentence! It looks like this:

We see how the semantic order of the components in German disseminations runs contrary to the syntactical order! This is sympto­ matic of German. In Dutch, for instance, this principle is broken:

I have come to the point now, where I must send a thought to the printer(s). I shall have to draw a few more diagrams, though. But I shall certainly be aware that even the best printing offices of our time have their limitations. A last observation for today must be concerned with a tricky problem having reference to a great number of languages using finite modal nuclei followed by that-sentences (subordinated (un)charged status): Mora da je caroban grad. (= 'it must be a wonderful town')

(Serbo-Croatian)

152

PAPERS IN THEORETICAL LINGUISTICS (Bulgarian) (= 'you must knock one time at the door') Ne treba da ides.

(Serbo-Croatian)

(= 'you need not go') (Arabic) (= 'he cannot get up') Mundet sapuni të jetë me i mirë.

(Albanian)

(= 'the soap can be better') Vajzat nuk mund të dergojmë në shkollë.

(Albanian)

[the girls not can that we send to school) (= 'we cannot send the girls to school') (Albanian) (= 'I must consult somebody concerning this matter') (Mod. Greek) (= 'the town would never be conquered') (Rumanian) (= 'I hadn't wanted to ask for Tincuta's hand') The question is now: which syntactical role does the subordinate clause play in this sort of sentence? Is it the lacking subject of the verb in the main clause, or is it the complement of the main clause? We vote for the complement solution. We do this for two reasons: the sentence nucleus always strives for its own complementation. And, 2, the false subject clause is highly unwilling, not to say absolutely reluctant to enter into the initial position of the sentence which is in very many of these languages the genuine topic of the subject. In quite a few languages the function of a modal nucleus is left to a naked and often weak functive

which then represents a most subtle

A FINAL WREST

153

modal meaning of the [BE] -entity: (Russian) (= 'why should she not have stayed a little longer?')

Some languages even conjugate such functives: Nie powiem, zebym się rozgniewał.

(Polish)

(= 'I cannot say that I was angry') All these observations and reflections allow us, now, to cope with the Bulgarian syntagm:

(= 'you must not miss the lessons') The subject is tion is

The nucleus is

. The nucleus complementa­

. But from an English, German, or French

view there is something utterly wrong. The nucleus in this sentence does not and cannot enter into a grammatical interrelation between the two halves of a nexual sphere. The nucleus which ought to be the investor above all has given up this essential role and concentrates fully on its complementation:

The nucleus is the verb 'to be'. Its form is astonishingly naked, although such investments as time and mode are obviously both represented. The nexual link between the subject and the verb is shoved on to the finite verb of the complementation. And the subject is where it is supposed to be:

stands initially, leaving no doubt

154

PAPERS IN THEORETICAL LINGUISTICS

about the role of the complementation (cf. above). Our Bulgarian sentence is an astounding specimen of an SQ-relation. In order to get on (or back?) to the more unengaged role of the [BE]-nucleus as the finite centre of complex verb constructions we shall have to look at the following extremely simple syntagms from Sanskrit, Lithuanian and Latvian: Aha [I

bhavadbhir dra by you

avya

(Sanskrit)

to be seen]

(= 'I am to be seen by you', 'you must see me') Tróbos matyti.

(Lithuanian)

(= 'the houses (are) to see (be seen)'; 'can be seen') Šis vasaras siens nopļauts.

(Latvian)

(= 'this summer the hay (has been) harvested') The three sentences are built up in a very similar way. They have all skipped the nucleus, leaving the function of the modal  to the next fellow player in the row, some participle or infinitive. The participle in the Sanskrit example is absolutely unequivocal: it is a modal participle with the specific role to express obligation in the passive. In our Lithuanian example it is matched by an infinitive, whose role it is to express potentiality. And in the Latvian example the copulaless syntactic topos is occupied by a past participle in the passive expressing nothing of a more modal imprint. In our example from Sanskrit the modal power of the left-out [ BE] -entity has been entrusted to a specific participle of the wanted nature, in our Lithuanian example the modal force is left entirely to an extensive grammatical category, i.e. the infinitive (which can play an unspecified number of other roles). And in Latvian we sin ply have a predicative construction like all other predicative constructions. If, in our Latvian example, we replace the passive participle with a corresponding active entity, we get the whole story about the nature

A FINAL WREST

155

of the Russian imperfect:

(= 'he/she/they spoke very distinctly') (The iconic sign

stands for finite participle!).

Now, if we leave the nucleus unaffected we get a dissemination: Tróbos yrà matyti.

(Lithuanian)

This is a frequent dissemination with all sorts of nuclei as its finite centre: Jää seisma!

(Estonian)

(= 'keep sitting!') Du måste göra detta.

(Swedish)

(= 'you must do this') If we extend the nucleus with participles or infinitives of different sorts, we get the well-known English, German, French, etc. complex verbal constructions, in which the finite nucleus takes most of the investments (polarity and gender/number may be delegated

to the

closest participle): Ja jsem ich nevidel.

(Czech)

(= 'I haven't seen them') þeir eru búnir  gera (= 'they have finished doing this')

þtt.

(Icelandic)

156

PAPERS IN THEORETICAL LINGUISTICS (Persian) [killed not-been I was] (= 'I had not been killed')

Even aspect is delegated to some fellow player in many languages: Es apēdu trīs äbolus.

(Latvian)

(= 'I have eaten three apples') Some languages delegate the investment category of diathesis to the next following infinite fellow player in a dissemination: Detta har aldrig gjorts förut.

(Swedish)

(= 'this has never been done before') Normally, in such instances, the diathesis marking(→passive) is taken over by a specific infinite segment in the row: Er ist erschreckt worden.

(German)

He has become frightened. Han er blevet forskrækket.

(Danish)

Il a été terrifié.

(French)

If the nucleus is modal, a whole series of troubles arise. Where is the topos of diathesis in a dissemination with a modal nucleus? 1)

Some languages leave out any indication of existential dia­ thesis in such constructions, cf. Icelandic: Fargjald skal greitt í síðasta iagi 2. mars. (= 'the ticket money must be paid at the latest 2 March')

2)

Others leave it to the nucleus: Võidakse küsida, miks se niin on. [(wird gekonnt fragen) why that so is] : 'One may ask why that is so')

(Estonian)

157

A FINAL WREST Voitiin järjestää julkisia hiihtokilpailuja.

(Finnish)

[wurde gekonnt arrangieren öffentliche SkiturniereJ 3)

Others again leave it to some specific segment: It may be doubted, whether he is the best.

Extremely exquisite is the delegation of person met with in some languages: El kell válnom tőled.

(Hungarian)

(= 'I must part from you') Neki is jönnie kellene.

(Hungarian)

(= 'he, too, would have to come') Nem kellett volna elmennünk.

(Hungarian)

(= 'we hadn't had to go there') Mar ott kellene lennem.

(Hungarian)

(= 'I ought to be there already') In such instances we may talk of such a radical extra finitizing of the mesonomic section that aj

-sentence type emerges.

E agradável passarmos a noite juntos.

(Portuguese)

(= 'it is agreeable that we spend the night together') So we get the following three main types of disseminations with a [ BE| -nucleus and an 'infinitive' in German: Dies ist zu ändern.

(obligational)

158

PAPERS IN THEORETICAL LINGUISTICS Dies ist kaum zu glauben. Er ist gerade essen.

(potential) (neutral)

The two modal variants use a modal participle (zu ändern, zu glauben) which looks like an infinitive in our two examples, but the true nature of which is unmasked if we turn it into its attributive form: der zu ändernde Fehler die zu glaubende Wahrheit The nucleus [ BE] is pressed very hard in all its functions by its emanent counterpart [HAVE]. We see this already in situations where [ BE] is the nucleus of pure existence: What is wrong with you? Was hast du? Or compare Danish

Hvad er der?

with French:

Qu'y a-t-il?

In German you may hear both 1)

Heute waren viele Leute ...

and 2)

Heute hatte es viele Leute ...

and 3)

Heute gab es viele Leute auf den Pisten.

For pure existence Icelandic uses er, English uses is, Dutch uses is, Russian uses , Urdu uses hay, etc., whereas Spanish uses hay, Serbo-Croatian uses ima, Albanian uses ka, Greek uses

Chinese

uses yeou, (and Polish uses jest in propositive sentences, but ma in recusative ones!).

A FINAL WREST

159

The universal fight between [BE] and [HAVE] may be noticed in all of the sentence types we have discussed. In Austria it is normal to say Er ist auf der Bank gelegen. (= 'he has been lying on the bench') Where Reichsdeutsch (Standard German) would say: Er hat auf der Bank gelegen. Both: Er ist gefahren. and:

Er hat gefahren.

are possible in German, but they have different meanings. Some languages use only [ BE] to form the perfect, for instance Finnish, Serbo-Croatian, Latvian, and others, whereas others, like English, use only [HAVE]. A third category uses both [BE] and [HAVE] according to special rules of usage and grammatical distribu­ tion. In Lithuanian, we even find obligational constructions where [HAVE] takes over for [BE] :

[this matter also for you should be known] (= 'you too ought to know this matter') [they had with him to go to war] (= 'they had to go to war with him') In English and German the nucleus [HAVE] is used quite normally in such constructions: He has to be here at six o'clock.

160

PAPERS IN THEORETICAL LINGUISTICS Er hat um sechs Uhr hier zu sein. (cf. Spanish: A caballo regalado no hay que mirarle el diente).

It is symptomatic of both the English and the German constructions that the nucleus is joined to its closest syntactic fellow player by the help of a very common preposition (English: to, German zu). This procedure is so common in German that German grammarians simply talk about two different kinds of common infinitives: one is called the zu-infinitive, the other is the so-called infinitive without zu. In Irish, the nucleus [BE| is combined with a whole bunch of indicators (prepositions) or ordinative transactions in order to get different complex verb constructions going: 1)

Tá sé ag obair le coicís anuas. (= 'he has been working for a fortnight now')

2)

Tá sé gan bheith ar fónamh. [is he without being in health) (= 'he is not healthy')

3)

Tá sé caillte aige. [is it lost with him] (= 'he has lost it')

4)

Tá cónai air ar imeall na cathrach anois. [is living on him on the outskirts of the city now)

5)

Tá an nuachtán á leamh aige. [is the newspaper at its reading with himl (= 'he is reading the newspaper')

6)

Tá sé ina shuí in aice na tine. [is he in his sitting near the fire] (= 'he is sitting near the fire')

7)

An raith sí do mo lorg?

161

A FINAL WREST ['?' was she to my looking] (= 'was she looking for me?') 8) Bhí orm rith abhaile. [was on me run home] (= 'I had to run home') 9) Ní raibh aon rud le hithe agam. [not was one thing to eat with me] (= 'I didn't have anything to eat') 10)

Tá sé chun scéal a insint dúinn. [is he for story to tell to us] (= 'he is going to tell us a story')

Irish seems to have the world record in the use of indicated segments. The indicated segment in Irish is always a Π (= the nominal content of the verb, or the verbal noun, or even better: the naked verbal substantive). From this there is no long jump to languages using adverbial verb form (sometimes wrongly called gerunds, their best nomenclature being 'absolutives'): (Lithuanian) [I have having much money saved] (Singhalese) [I at the station having come been] (= 'I come from the station') (Singhalese) [then you in advance being having tickets buy gone] (= 'then you proceed and buy the tickets in advance!')

162

PAPERS IN THEORETICAL LINGUISTICS Piyadāsa mē gedara padimci vi innavä.

(Singhalese)

[P. in this house mansion having is taken] (= 'P. has settled down in this house') Vahina-kota mama kār ekē gedara emin hitiya.

(Singhalese)

[raining when I car in home going was] (= 'during the rain I was going home in my car') Mama  ma ka ka hitiya.

(Singhalese)

(redupl. abs. II) (=

'I was devouring the meal') 2)'I used to devour the meal')

Ganga galamin tibunā.

(Singhalese)

[the river floating was] (= 'the river was floating along') A segmentational infinitive may even end up in a specific case form in some languages: Kangas näyttää olevan tanskalaista.

(Finnish)

(= 'the material seems to be Danish') In this construction the infinitive ('to be') is in the accusative (olevan!). In a sentence with the nucleus 'to be' the next following verbal segment is either finite or infinite: 1)

Huwa kâna yadhhabu fi

l-madïnati.

(Arabic)

(imperfect) (= 'he was walking to the city') Huwa kina dhahaba fi l-madinati. (perfect) (= 'he had walked to the city')

(Arabic)

163

A FINAL WREST

(cf.

(Arabic) 

imperfect

[he was almost he made fly him] (= 'he almost made him fly') Nem kellett volna elmennünk,

(Hungarian)

(= 'we hadn't had to go there') 2)

Ahmed'in bu

olsayd1m,

(= 'if I hadn't heard this word of Ahmed's, I wouldn't have thought of anything') Kodomo ga asonde-imasu.

(Japanese)

(= 'the children are playing') Dit her önersaacht wiis maast.

(North Frisian)

(= 'this ought to have been investigated') Er ist kaum ins Wasser gegangen.

Er war fischen.

In some constructions the [BE]-nucleus is split up to an incredible extent: Amari omoshiroku nakatta desu arimasen deshita. [especially interesting isn't is not it was] (= 'it wasn't very interesting')

(Japanese)

164

(cf.

PAPERS IN THEORETICAL LINGUISTICS

Kono kōjō wa okiku nai desu. [this factory not-big is] Kono kōjō wa shizuka dewa arimasen. [this factory quiet being not-is] Mukashi wa anna-ni kirei dewa arimasen deshita. [formerly that beautiful being not-is was] (= 'formerly she wasn't that beautiful')

A similar feature we find in Chinese which is utterly non-restrictive when it comes to disseminating with the help of existential nuclei: Zhāng xiānsheng zai waitou dengzhe ni le. [Zhang Mr. is outside waiting (for) you] (= 'Mr. Zhang is outside waiting for you') If we translate a few (mainly modal) segmentations from, say, German, we run into a baffling realm of 'differentness', syntactical recalcitrance and preference of the infinite. So, if we take a German sentence like Er muss seinen Pass verlegt haben. and translate it into Chinese, we get (our system of transcription is the unvanquished one proposed and used by Dr. Olov Bertil Anderson, formerly Lund, now Uppsala, Sweden:4

[absolutely he has forgotten he (accusative) passport place is-in which place]

165

A FINAL WREST The German syntagm: Er wird uns gehen lassen müssen gets much more economic: Tha

yaoz

wox-men

chhyuz.

[he absolutely will let us go away] The German sentence Dies hätte untersucht werden müssen, (= 'this should have been investigated') emerges in Chinese deprived of at least half of its nuclear idiosyncracy:

[already that time this (piece) matter (relation) ought investigate to the bottom] The German Ich war noch kurz was essen. becomes:

(= 'I just wanted eat') and the German: Er blieb stehen. turns into: Tha

le.

(= 'he kept on standing') And the German Wir haben ihn gehen lassen müssen.

166

PAPERS IN THEORETICAL LINGUISTICS

is refound in Chinese as: Wox-men puz chiex-fangz tha puz khex! [we not let loose let go him is no good] The last one of these six nucleus-killers is really the very best example to show how you may get rid of the boring nuclei. Either you choose a completely different sentence structure, or you stuff as much as possible into single lexemes. Some day this should be investigated in all details! Our last Chinese example leads to the untying of our knot. It is an extremely troublesome untying we still have in front of us: it is the entire grammatical and syntactical chapter comprising what is nor­ mally described as 'negation'. This notion is unpardonably turbid and muddled. Recently we have proposed a new procedure of description for human sentences and their polarity. The first time we did it was in a dissertation on Gothic and Welsh (196S),5 but since then different refined approaches have been published. We propose to come to grips with the intricate pattern through the employment of 2 x 2 new terms, those of diaxy and synaxy, mapping out the categorical decision in the following way: I do not know him.

(÷/r,

synaxy,

logically

negative, semasiosyntactically recusative) Ask him if he doesn't know him.

(+/r,

diaxy,

logically

positive, semasiosyntactically recusative) I know him.

(+/p,

synaxy,

logically

positive, semasiosyntac­ tically propositive)

167

A FINAL WREST Nobody knows him.

(÷/, diaxy, logically negative, semasiosyntactically propositive)

The deep secret of this apparatus is only seemingly a cryptic one: our four sentences show us that we get nowhere just using the words 'positive' and 'negative'. Instead we speak about 'polarity': a sentence is recusative if it has the word 'not' and propositive if it hasn't got that word 'not'. The rest is a question of being able to combine these two simple sentence categories with the logically 'negative' and 'positive' (seman­ tic!) manifestations. That's all there is to it. So if we look at a couple of extremely simple Estonian paradigms, we see in a flash what all the traditional talk about 'negations' is about (or rather what it isn't about): Ia) Ma olen lugenud. b) Ma ei olen lugenud. Ma pole lugenud. IIa) b)

Ma olin lugenud. Ma ei olnud lugenud.

(= 'I have read') (='I haven't read')

(= 'I had read') (='I hadn't read')

Ma polnud lugenud. Ia differs from Ha only in the fact that the nucleus (a form of the verb 'to be'!) changes its investment of time. The form 'olen' is the present form of olla. In Ib the functive 'ei' (which escapes the conjugation in person and number!) takes over and is joined to the finite form olen which we met in Ia. Or its alternative 'p-' (= 'not') is put directly before 'ole', giving the corroded result 'pole (= 'is not') which again is not susceptible to any conjugation in number and person. In IIb the functive 'ei' is placed in front of the participle 'olnud' to which it

168

PAPERS IN THEORETICAL LINGUISTICS

belongs from a semasiosyntactic point of view, or it is again ex­ changed with the alternative 'p-' (= 'not'), and the end result is the most astounding 'polnud lugenud' (with no verbum finitum the remnants of which have been devoured by its closest infinite co-segment!). Recusativization through the verb form alone is known from a great many languages, English is a language developing specific recusative verb forms for quite a considerable number of frequently used nuclei. The intention of language is obviously being counteracted by the technical devices we have invented to make our generation the happiest of all generations. So, a sentence like He ain't very sweet. has become a signal of un-eruditeness rather than the end point of a tranquil and quite natural development. The forms 'don't', 'shan't' and 'won't' have all passed the hurdle and are the alternatives, nothing less, nothing more, of the longer constructions 'do not', 'shall not' and 'will not'. The English state of affairs is quite unique: no other Germanic language has developed specific recusative forms of any nuclei; in English we find what we might describe as the British answer to a universal demand triggered, not least, by the usual Celtic heritage: all Celtic languages have specific recusative forms of the existential nucleus. This forces me to say a lot more about human negation. And so I shall write and say much more about it in due time. I am afraid I must leave the negativity here. I am sure you will forgive me this abrupt finis of my paper if I promise you a whole book on this theme which was once the topic of my prize essay at the University of Copenhagen (1959). Institut for germansk filqlogi Odense

Universitet

DK-S230 Odense M

169

A FINAL WREST Notes 1.

A Reader on the Sanskrit Grammarians, Cambridge, Mass. and London 1972.

edited by J.F. Staal,

2.

Cf. Niels Danielsen, Linguistic Studies, Heidelberg 1980.

3.

Cf. Niels Danielsen, 'Complex Mesonomic Verb Structures', in: Friserstudier II, Odense 1982.

4.

O.B. Anderson, A Concordance to Five Systems of

Transcription

for Standard Chinese, 2nd edition, Lund 1975. 5.

Niels Danielsen, Status und Polarität im Gotischen - im Lichte des Kymrischen dargestellt, Odense 1968.

Acknowledgement: I am extremely grateful to N.M. Knudsens Fond for subsidizing the excellent drawings made by Inger Bjerg Poulsen for this paper.

PLÄDOYER GEGEN DIE GENERATIVEN TIEFENOPERATIONEN Kritik einer Scheinlehre In seinem Buch Language and Mind1 behauptet Noam Chomsky, es sei nützlich, für den Satz a wise man is honest folgende Tiefen­ struktur anzusetzen:

Diese Behauptung Chomskys ist eins von zweien: entweder eine fixe Idee oder ein verrückter Einfall. So viel ist sicher, daß sie im Lichte früherer Ausführungen des Magicus aus Massachusetts zu sehen ist, die alle den Eindruck erwecken, als seien sie der sponta­ nen Intuition entwachsen und nicht konsequent zu Ende geführt. Mehrere urteilsfähige Sprachforscher haben energisch gegen das Chomskysche Verfahren Einspruch erhoben. Selbst haben wir un­ zweideutig die seichte und wie aus der Luft gegriffene Operation als epistemologisch leer bzw. sprachlich falsch abgelehnt 2 , wobei diese Ablehnung prinzipiell sämtlichen mit dem vorliegenden gene­ rativen Vorgang zu vergleichenden Transformationen gilt. Zahl­ reiche irrationale Diskussionen mit ätherischen Modepredigern, für die zwischen den Begriffen „modern" und „generativ" schon kein Unterschied mehr besteht, zwingen mich dazu, noch einmal die Fe­ der zu ergreifen, um ein bißchen ins Detail zu gehen. Und somit wird diese Darlegung letzten Endes nichts als eine Fußnote zu mei­ nem schon längst mit Verdutztheit und Befremdung geäußerten Beitrag zu einer grotesken Debatte 2 . 1 Noam Chomsky, Language and Mind, New York/Chicago/San Francisco/ Atlanta 1968. 2 Niels Danielsen, Das generative Abenteuer, Språkliga Bidrag Vol. VI, Nr. 26 (Sonderdruck), Lund 1971, sowie in: ZPSK, Heft 4—5/72, Berlin 1972.

172

PAPERS IN THEORETICAL LINGUISTICS

Ganz verwickelt wird die Chomskysche Prozedur, wenn wir sie beispielsweise an einem modernen Sprachzustand wie dem Bretoni­ schen durchführen wollen. Betrachten wir den Satz: An dén fur a zo onest (= „der kluge Mann ist ehrlich").

Er ist folgendermaßen strukturiert: an (bestimmter Artikel) dén („Mann") fur („klug") a (Zeichen der propositiven Polarisation eines neutral enuntiativen Status) zo („ist") onest („ehrlich"). Als tiefenstrukturelle Basis dieses Satzes wäre jetzt folgendes Konglo­ merat anzusetzen: an dén / dén a zo fur / a zo onest

Dem Rezept gemäß tilgen wir nun das zweite dén, um es dann durch das Relativmerkmal a zu ersetzen: an dén / a a zo fur / a zo onest

Das neuerdings eingeführte a muß jetzt vor dem zweiten a weichen, da sich beide nebeneinander in keinem bretonischen Satz vertragen, und da das Aufrechterhalten des ersteren den Sinn des Satzes stö­ ren würde: an dén / a zo fur / a zo onest

Der Satz bedeutet jetzt: „Der Mann ist klug, ist ehrlich" und ist ein schlechter Satz. Der Satz: „Der Mann, der klug ist, ist ehrlich" heißt auf gut Bretonisch 1. Onest eo an dén a zo fur (mit dem aktuellen Präsens eo), 2. Onest e vez an dén a zo fur (mit dem iterativen Präsens vez, dem das propositive Status/Polarisationszeichen e vorausgeht).

Wir müssen jetzt zu unserer Enttäuschung feststellen, daß sich die Sätze 1 und 2 unter keinen Umständen als operationelle K S Basen der für die als Ausgangspunkt verwendete Oberflächenstruk­ tur anzusetzenden Tiefenstruktur verwerten lassen: der eingebet­ tete Satz steht am Schluß des Satzgefüges, das Prädikativ steht ganz am Anfang (während es früher ganz am Ende stand), und die Ko­ pula hat sich gänzlich geändert. Der Satzgenerator funktioniert nicht. Das Bretonische kommt glänzend mit attributiven Adjektiven aus, die keine generative T Adj -Transformation durchlaufen kön­ nen. Wozu soll man dann eine solche Transformation mit dem An­ spruch auf Allgemeingültigkeit für das Englische und das Deutsche ansetzen, wenn sich wie oben nachweisen läßt, daß sie nicht in allen Sprachen vernünftig aufgeht? Mit unserem „bretonischen Knoten" wäre an und für sich jegliche weitere Beweisführung gegen das generative Luftgewebe als völlig

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

173

überflüssig zu betrachten. Da wir jedoch jetzt die Möglichkeit ha­ ben, möchten wir noch auf einige wichtige Einzelheiten eingehen. Ein für allemal. Im Georgischen steht das attributive Adjektiv in der Regel vor dem zu bestimmenden Substantiv: Čkwiani kaci patiosania ( = „ein kluger Mann ist

ehrlich").

Das attributive Adjektiv kann auch nach seinem Substantiv ste­ hen. Das vorangestellte Adjektiv hat seine eigene Deklination. Das nachgestellte Adjektiv hat ebenfalls seine eigene Deklination. Das alleinstehende Adjektiv wird wie das nachgestellte flektiert (wobei das Adjektiv als alleinstehend zu betrachten ist, wenn es von dem zu bestimmenden Substantiv durch ein anderes Satzglied getrennt wird). Das vorangestellte attributive Adjektiv unterliegt keinerlei Synkope, das nachgestellte oder alleinstehende attributive Adjek­ tiv wird dagegen vielfach synkopiert. Das vorangestellte attribu­ tive Adjektiv weist im Georgischen keinen Plural auf. Als Ersatz für den fehlenden Plural dient vielfach der verdoppelte Stamm des betreffenden Adjektivs. Das darauffolgende Substantiv steht dann entweder im Plural oder im Singular. Prädikative Adjektive, die sich auf eine Mehrzahl von Dingen beziehen, stehen im Plural. Be­ ziehen sie sich auf eine Mehrzahl von Personen, stehen sie gelegent­ lich im Singular. Zur Beschreibung dieser Erscheinungen hat die generative Gram­ matik ihre eigenen seltsamen Mittel: außer den unseligen Transfor­ mationen eine nichtssagende Lexikonregel (Wenn Q ein komplexes Symbol ist und (PM, SM) ein Lexikoneintrag, und wenn Q nicht verschieden von SM ist, dann wird Q durch die Vereinigungsmenge M aus Q und SM ersetzt und M durch PM symbolisiert) sowie eine genau so leere Anweisung, die strikte Subkategorisierung für ein Symbol A durchzuführen (Einer Kategorie A werden die Teilket­ ten X, A, Y als Merkmale zugewiesen, wenn XAY eine Kette bildet, die unmittelbar von einer Kategorie  dominiert wird): Q→ SM und  → KS 3 . Ich glaube aber nicht an Q und ebensowenig an A, da ich nicht das Vergnügen habe, sie jemals getroffen, geschweige denn festgestellt zu haben. Und wenn ich sie kennen würde, wäre meine erste Frage: Woher kommt dann Q bzw. A? Aus Q bzw.  oder aus Z und C? Und so weiter ad infinitum und ad absurdum. 3 Vgl. Johannes Bechert, Danièle Clément, Wolf Thümmel, Karl Heinz Wagner, Einführung in die generative Transformationsgrammatik, München 1971, S. 72 u. 77.

174

PAPERS IN THEORETICAL LINGUISTICS

Mit einem Q oder einem A kommt man bekanntlich nicht sehr weit in der Medizin oder in der Astronomie. Höchstens in der Astrologie und damit verwandten „Wissenschaften". Nun läßt sich aber mit der Magie in bezug auf die georgischen Adjektive kaum etwas Neues aufdecken. An anderer Stelle fanden wir schon eine passende Gelegenheit, auf das verwickelte Wechselspiel zwischen attributiven und prädi­ kativen Adjektiven im Rätoromanischen einzugehen 4 . Fast so neckisch wie die Adjektiva im Rätoromanischen sind ihre attributiven Kollegen im Kymrischen. In dieser Sprache steht manchmal der attributiven Gemeinform des Mask. Sing. und des gesamten Plurals eine spezifische Form des femininen Singulars gegenüber, die sich durch Lenierung des anlautenden Konsonanten sowie durch einen besonderen Vokalismus auszeichnet: Y m a e ' r d y n cryf y n onest ( = „der s t a r k e M a n n ist ehrlich"). Y m a e ' r w r a i g gref y n onest ( = „die s t a r k e F r a u ist ehrlich").

Ähnliche Adjektivpaare sind besonders zu beobachten: Mask. Sing. + Pluralis brych byehan byr

gwlyb gwyn gwyrdd llyfn llym melyn sych syth brwnt crwm crwn dwfn llwm swrth tlws trwm brith

: : : : : : : : : : : : : : : : : : :

Fem. Sins. frech fechan fer wleb wen werdd lefn lern feien sech seth front grom gron ddofn lom sorth dlos drom fraith

(= (= (= (= (= (= (= (= (= (= (= (= (= (= (= (= (= (= (= (=

„gesprenkelt") „klein") „kurz") „naß") „weiß") „grün") „glatt") „scharf") „gelb") „trocken") „gerade") „schmutzig") „krumm") „rund") „tief") „nackt") „verdrießlich") „hübsch") „schwer") „fleckig")

Die generative Grammatik wird sicherlich eine „adäquate" Be­ schreibung der einschlägigen Erscheinungen geben. Man wird wie­ der tilgen, wo es „notwendig" ist, und inserieren, wo es „erforder­ lich" ist. Man wird permutieren, wo es „praktisch" erscheint. Man 4 Vgl. Niels Danielsen, „Das generative Abenteuer", in ZPSK, H. 72/4, Berlin 1972.

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

175

wird aber dabei zu guter Letzt nur feststellen können, daß die Waliser Kymrisch sprechen, und das wußten schon mehrere. Im Arabischen stimmen die Adjektiva weitgehend mit ihren sub­ stantivischen Korrelaten überein: in bezug auf Genus, Numerus und Kasus richtet sich das Adjektiv nach seinem Substantiv: Rajulun hakïmun šarīfun Bintun hakīmatun šarīfatun khädimun ghä'ibun khädimäni ghä'ibäni khädimūna ghä'ibūna (= khädimatun ghä'ibatun khädimatäni ghä'ibatäni

(= „ein kluger Mensch ist ehrlich"). (= „ein kluges Mädchen ist ehrlich"). ( = „ein abwesender Diener") (= „zwei abwesende Diener") „abwesende Diener") (= „eine abwesende Dienerin") (= „zwei abwesende Dienerinnen") (= „abwesende Dienerinnen")

Adjektiva können im Plural neben den „ungebrochenen" Formen auch „gebrochene" Formen aufweisen. Die gebrochenen Formen werden in der Regel verwendet, wenn das substantivische Korrelat eine männliche menschliche Person bezeichnet: rijälun tiwālun

(= „hohe Männer")

In der arabischen Poesie (vor allem in der klassischen), in rheto­ rischen und wissenschaftlichen Texten finden sich gebrochene Plu­ ralformen des Adjektivs neben den ungebrochenen auch in Verbin­ dung mit Mehrzahlformen von Substantiven, die keine männlichen Personen bezeichnen: (= „schwierige Wissenschaften") (= „lange Schiffe")

Sonst werden in Verbindung mit gebrochenen Pluralformen von Substantiven vorwiegend Adjektivformen verwendet, die formal mit dem Femininum Sing. zusammenfallen: mudun un kabīratun durüs u n sa'batun al-mudunu kabīrat""

(= „große Städte") (=. „schwierige Lektionen") ( = „die Städte sind groß")

Gebrochene Pluralformen vom Adjektiv werden oft in Verbin­ dung mit ungebrochenen maskulinen Substantiven im Plural ver­ wendet: khädimüna kibärun

(= „alte Diener")

Ein gutes arabisches Wörterbuch gibt übrigens seinem Benutzer, sei er generativ oder nicht, deutliche Auskunft über die schon längst festgestellte Tatsache, daß das Adjektiv bīrun lediglich den gebrochenen Plural kibārun bildet, der folglich im Mask. Plur. als überall alleinherrschend betrachtet werden muß.

176

PAPERS IN THEORETICAL LINGUISTICS

Feminine Substantiva werden im Plural in der Regel mit einer Form des Adjektivs verbunden, die morphematisch mit der des femininen Singulars zusammenfällt:

Im letzteren Falle ließe sich ebenso gut die Mehrzahlform des Adjektivs verwenden: Die Regel ist nämlich, daß in Verbindung mit femininen Sub­ stantiven im Plural, die weibliche Personen bezeichnen, neben der Einzahlform des Adjektivs auch die Mehrzahlform verwendet wer­ den kann. Für eine eingehendere Behandlung des arabischen Ad­ jektivs finde ich hier keinen Anlaß. Es ist überdies ein recht schwie­ riges Kapitel. Fest steht jedoch, daß uns die generative Grammatik in dieser Hinsicht kein Jota zu unserem bereits errungenen Wissen hinzuzufügen vermag 4a . Im Baskischen steht das attributive Adjektiv in der Regel nach dem Substantiv, auf das es sich bezieht: Alo! Ardi konkor bizardun bat eta aker churi gorri handi bat! (= „Nanu! Ein buckliges und spitzbärtiges Schaf und ein großer weiß- und rotfarbiger Bock !")

Der unbestimmte Artikel (bat) steht nach dem Block Substantiv + attrib. Adjektiv(e). Wird der Block dekliniert, nimmt nur das blockabschließende Adjektiv Deklinationsmorpheme an: Ardi konkorraren eta aker grorriaren etchean oro bitchiak dire (= „in dem Haus des buckligen Schafes und des roten Bocks sind alle drollig").

Adjektive, die eine Nationalität oder eine Religionsangehörigkeit bezeichnen, stehen jedoch in attributiver Stellung oft vor ihrem jeweiligen Substantiv: Han zohan handituz eskualdun frailearen omena (= „da verbreitete sich der Ruhm des baskischen Mönchs").

Gewisse baskische Adjektive ändern die Bedeutung je nachdem, ob sie bestimmt oder unbestimmt auftreten: klechu da bizi da ari da eri da gose da

= „er ist verärgert" = „er ist am Leben" = „er arbeitet" = „er ist krank" = „er ist hungrig"

klechua da bizia da aria da eria da posea da

= = = = =

„er ist zornmütig" „er ist lebhaft" „er ist aktiv" „er ist kränklich" „er ist unersättlich" usw

4a Mit an Schamlosigkeit grenzendem Freimut wird mancherorts in der gene­ rativen Literatur die totale Nacktheit, ihre Scheintheorie, dem staunenden Leser offengelegt (siehe z. B. die beiden unbezahlbaren Rezepte einer "angewandten" Satzgeneration auf S. 338—39 in: Norman C. Stageberg: An Introductory English Grammar, New York, Chicago, San Francisco, Toronto, London 1967.

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

177

Eine Reihe von baskischen Adjektiven läßt sich nur prädikativ verwenden. Es handelt sich hierbei vor allem um adjektivische Semanteme, die weitgehend durch indoeuropäische Relativsatz­ umschreibungen genau erfaßt werden müssen: die konkretisie­ rende and aktualisierende Rolle des Prädikativ-Relativsatzes ist offensichtlich nicht zu unterschätzen! Wo die generative Grammatik diese kaum geringzuschätzende Tatsache verachtet, indem' sie gerade beim attributiven Adjektiv eine T A d j -Transformation kri­ tiklos ansetzt, wird erst beim prädikativen Adjektiv im Baskischen eine ähnliche Operation des öfteren aktuell (aber praktisch kaum durchführbar): ahalge da = „der sich schämt" e r ist == „der schwitzt" e r ist izerdi da haiduru da = „der w a r t e t " er ist bali da = „der t a u g t " e r ist konda da --= „der zählt" e r ist u n d viele a n d e r e , die ein Gefühl, eine Empfindung, einen G e d a n ­ k e n z u s t a n d , eine H a n d l u n g , eine Fähigkeit, eine V e r a n l a g u n g , eine Ü b e r e i n s t i m m u n g , ein A b k o m m e n o d e r einen Z u s t a n d a k t u e l l h e r ­ v o r t r e t e n lassen.

Nachdem wir jetzt auch für das prädikative Adjektiv eine T A d j Transformation haben ansetzen müssen, wird das generative Formalisierungsspiel erst richtig interessant! Daß die generative Grammatik keine allgemeine Sprachtheorie ist, wird sogar von generativer Seite bestätigt 5 . Wir warten nunmehr darauf, daß sie auch ihr eigenes epistemologisches Unvermögen wahrhaben will 6 . Im Japanischen gibt es fünf verschiedene Arten von attributiven Adjektivrealisationen: 1) 2) 3) 4) 5)

Attributive Attributive Attributive Attributive Attributive

Adjektive Adjektive Adjektive Adjektive Adjektive

o h n e Begleitpartikel. mit d e r P a r t i k e l no. m i t d e r P a r t i k e l na. m i t oder o h n e die P a r t i k e l no. auf -i oder m i t der P a r t i k e l na.

5 Vgl. Even Hovdhaugen, Transformasjonell generativ grammatikk, Oslo 1969, S. 140: „A anta den samme dypstruktur og systematiske fonemiske representasjon i f. eks. alle indoeuropeiske sprâk ville vaere absurd." 6 Vgl. Karl R. Popper, The Logic of Scientific Discovery, London 1959, S. 121: "Thus I regard the comparison of the empirical content of two statements as equivalent to the comparison of their degrees of falsification. This makes our methodological rule that those theories should be given preference which can be most severely tested equivalent to a rule favouring theories with the highest possible empirical content." — S. 121—122: "There are other methodological demands which may be reduced to the demand for the highest possible empirical content. Two of these are outstanding: the demand for the highest attainable level (or degree) of universality, and the demand for the highest attainable degree of precision." — S. 280: "Those among us who are unwilling to expose their ideas to the hazard of refutation do not take part in the scientific game,"

PAPERS IN THEORETICAL LINGUISTICS

178 Beispiele:

1. Kashikoi hito wa shöjiki desu (= „der weise Mann ist ehrlich"). 2. Shöjiki na hito wa kashikoi (desu) (= „der ehrliche Mann ist weise"). 3. Uso no hanashi wa baka desu (= „die unwahre Geschichte ist töricht"). 4. Ki no shindai wa fuhitsuyö desu (= ,,.das hölzerne Bett ist un­ nötig"). 5 (= „die große Stadt ist schön").

In prädikativer Stellung treten sämtliche Adjektive, die prädi­ kativ verwendet werden können, ohne Begleitpartikel auf. Adjek­ tive, die in attributiver Stellung ohne Partikel auftreten, vertrei­ ben in prädikativer Stellung weitgehend den kopulativen Prädikatsnucleus desu ( = „ist") (vgl. Beispiel 2). Adjektive, die unserer Gruppe 4 angehören, sind sogenannte Stoffadjektive. In der Umgangssprache werden sie meistens mit der Partikel no verwendet, in der wohlgepflegten Literatursprache wird diese Partikel in der Regel ausgelassen: kin no tokei gomo no kutsu buriki no kan kinu no ito

: : : :

kin dokei gomo gutsu buriki kan kinu ito

(=

„eine goldene Uhr") (=„Gummischuhe") (=„Blechdose") (= „ein seidener Faden").

Wird die Konstruktion ohne die Partikel no vorgezogen, muß das vor der Partikel no stehende Adjektiv in gewissen Fällen kon­ sequent vor einem ganz anderen Adjektivlexem weichen: ki no shindai ishi no ie

: :

mokusei shindai (= „ein hölzernes Bett") sekizö kaoku (= „ein steinernes Haus")

Was tut nun ein gewiegter Generativist, um diese nicht gerade einfachen Tatsachen zu beschreiben? Er relativiert, wo kein Japa­ ner davon träumen würde. Zu seinem Zweck verwendet er eng­ lische Tiefenstrukturen, weil er fest daran glaubt, daß sie die uni­ versellsten seien. Er tilgt und inseriert, wo seine von vornherein untaugliche Prozedur auf Hindernisse stößt. Sein Verfahren hat einen unabweisbaren Vorteil: es braucht sich nicht um empirische Gegebenheiten zu kümmern. Billige und altmodische Erwägungen über wissenschaftliche Redlichkeit (im guten, alten Sinne, wohl­ gemerkt) kann man sich in einer neumodischen und freizügigen Sprachtheorie nicht leisten. Ach, käme bloß die hippokratische Heilkunde mit derselben Großzügigkeit aus! Dann wären Krebs und Sklerose nicht mehr. Sie ließen sich jederzeit tilgen. Und wie viel billiger wäre nicht die Reise zum Mond, wenn wir die Schwerkraft

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

179

tilgen könnten? Ein jeder von uns könnte beweisen, daß seine Mut­ ter ein Stein wäre. Rasmus Berg führt tatsächlich diesen Beweis in Ludvig Holbergs Komödie „Erasmus Montanus". Holberg nennt sein Stück eine Komödie. Hat er doch besser als irgend jemand ge­ wußt, was das Leben komisch und zugleich unerträglich langweilig machen würde. In bezug auf die gegen den transformationeilen Generator auf­ sässigen japanischen Adjektiva ist der Tisch jedoch lange noch nicht abgetragen. Japanische Adjektiva benehmen sich bekanntlich wie Verba. Das heißt: sie werden konjugiert. Dies trifft zwar n u r für die „echten" Adjektiva zu. Die „echten" Adjektiva werden von unserer oben angeführten Gruppe 1 umfaßt und gehen auf -ai, -ii, -ui oder -oi aus. Im Präsens wird das Adjektiv takai („teuer") in fünferlei Weise realisiert: takai takai desu takai n' desu (takaku arimasu) takö gozaimasu

(= „[es] ist teuer")

Die vierte Form ist in der propositiven Konjugation außer­ ordentlich selten. Die fünf Formen bedeuten genau dasselbe, unter­ scheiden sich aber in bezug auf Höflichkeitswert, wobei eine höf­ lichere Form immer unter einer unhöflicheren aufgeführt ist. Vor der ehrenvollen Form gozaimasu ( = „ist") ändert das Adjektiv takai seine Struktur und tritt als takö hervor: Das Verbum gozai­ masu verlangte ursprünglich die adverbiale Form des Adjektivs takaku, das  der Endung -ku fiel weg, und aus au wurde ö. Auf ähnliche Weise wird beispielsweise kuroku (aus kuroi, „schwarz") zu kurö, und aus samuku (zu samui, „kalt") wird samū. Diese Eigen­ tümlichkeit trifft für sämtliche Adjektiva auf -ai, -oi und -ui zu. Neben der propositiven Konjugation gibt es nun eine entspre­ chende rekusative (bzw. „negative"). Sie weist im Präsens die fol­ genden Formen auf: takaku nai takaku nai desu takaku nai n' desu takaku arimasen takö gozaimasen

(= „[es] ist nicht teuer")

In der rekusativen Konjugation tritt n u n die adverbielle Form des Adjektivs auf. Sie wird dabei an die konjugierten Formen des

180

PAPERS IN THEORETICAL LINGUISTICS

negativen Existenzadjektivs nai gerückt. Wir bekommen demnach folgende Formen des japanischen Adjektivs: propositiv: rekusativ: Präsens: takai takaku nai Präteritum: takakatta takaku nakatta 1. Futurum takakarö takaku nakarō 2. Futurum takakattarō takaku nakattarō 1. Konditionalis takakereba takaku nakereba 2. Konditionalis takakattara takaku nakattara Gerundiv takakute, takui de takaku nakute / takaku nai de Alternativ: takakattari tataku nakattari / tatakunandari. Das Futurum weist noch die folgenden Wahrscheinlichkeitsfor­ men auf: takai takai darö takakarö

deshō

takaku nai deshō takaku nai darö takaku nakarö.

Diesen Formen entsprechen im Präteritum die folgenden sprach­ lichen Realisationen: takakatta deshö takaku nakatta deshö takakatta darö takaku nakatta darö takakattarö takaku nakattarö. Von den überall im System auftretenden „progressiven" Verbal­ konstruktionen, auf die wir in diesem "Zusammenhang nicht näher eingehen können, sind besonders die futurischen von Belang: takute imashö takute inai deshö takute iru deshö takute inai darö takute iru darö takute imasumai. Der Konditionalis der Vergangenheit tritt in folgenden Varian­ ten auf: takai deshö (ni) takakunai deshö (ni) takai darö (ni) takakunai darö (ni) o-takai deshö (ni) o-takakunai deshö (ni) (Höflichkeitsform). Hierzu kommt ein Konditionalis der abgeschlossenen Vergan­ genheit: takakatta deshö (ni) takaku nakatta deshö (ni) takakatta darö (ni) takaku nakatta darö (ni) takakattarö (ni) takaku nakattarö (ni) o-takö gozaimashita deshö (ni) o-takö gozaimasen deshita deshö (ni) o-takakatta deshö (ni) o-takaku nakatta deshō (ni). Die beiden letzten P a a r e sind die Höflichkeitsformen. Wer jetzt transformieren will, der rücke dem soeben dargelegten konjugierten Adjektiv des Japanischen mit seinen Tilgungen und

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

181

Insertionen zu Leibe. Es kommt dabei nichts heraus, was jeder einigermaßen vernünftige Mensch sich nicht an den fünf Fingern hätte abzählen können. Er würde dabei einfach das vornehmste Instrument des homo sapiens verwenden: das Gehirn. Und keine Transformationen7. Das Gehirn kann tausendfach mehr leisten, auch wenn man seiner unbewußt ist, als Tausende verkrampfter Transformationen, auf die man sein ganzes Bewußtsein richtet. Dem Menschen gegenüber kommt der transformationelle Genera­ tor ständig zu kurz. Sie verhalten sich zueinander wie die Früh­ lingsnachtigall im grünen Wald zu einem künstlichen Singvogel in einem Spielzeugladen. Dies zeigt sich vor allem in der entmutigen­ den Tatsache, daß sich die generative Grammatik nur um die Ana­ lyse von zumeist kurzen und einfachen Sätzen kümmert, die dann immer zu ihrem eigenen Zwecke zusammengebastelt sind. Vor der lebenden Sprache weicht sie zurück. Das scheint leider ihre einzige Möglichkeit zu sein. Streng wissenschaftlich gesehen hat sie kaum irgendwelche Berechtigung, es sei denn, die Psychologen oder die Geisteswissenschaftler stellten sich die wissenschaftlich relevante Frage: Inwieweit läßt sich die gesunde Vernunft über den Haufen werfen? Betrachten wir jetzt den Satz „Der große Laden war nicht teuer", wie er in japanischer Ausformung zutage tritt: sono sono sono sono sono sono

ökii ökii ökii ökii ökii ökii

mise mise mise mise mise mise

wa wa wa wa wa wa

takaku nakatta t a k a k u n a i deshita t a k a k u n a i n' deshita t a k a i ( n ' ) d e w a a r i m a s e n deshita t a k a k u a r i m a s e n deshita t a k ö gozaimasen deshita (deshita = „war").

Ein und derselbe Inhalt erfordert sechs verschiedeneTiefenstrukturoperationen, und zwar genau so viele, wie es Oberflächenstruk­ turen gibt! Dies allein scheint äußerst verdächtig zu sein. Hierzu kommt das gewöhnliche: die Tilgungen — was muß nicht hier an völlig zufälligen und sinnlosen Tilgungen unternommen werden, um TAdj-Transformationen durchzuführen? — und die Insertio­ nen. Überhaupt das normative Generationsgetändel, das nur von einem Sprachgeist wissen will, und zwar dem englischen aus den 7 Vgl. Niels Danielsen, „Das generative Abenteuer", in: ZPSK, H. 1972/4, Berlin 1972; Niels Danielsen, Die Frage. Eine sprachwissenschaftliche Untersuchung. Det Kongelige Danske Videnskabernes Selskab, Hist.-Filos. Skrifter 7,1. Kopenhagen 1972; Niels Danielsen, „A First Constitutional Step Towards a Universal Syntax", in ZPSK 1974.

182

PAPERS IN THEORETICAL LINGUISTICS

sechziger Jahren des 20. Jahrhunderts — und das auch nur annähe­ rungsweise und mit Mühe kaum. Werfen wir jetzt einen Blick auf einige Beispiele aus dem Russi­ schen: 1. a) b) c)

(= (= (=

„der Marienkäfer ist fleckig"). „der Gottestempel ist verfallen"). „die Gottesfurcht ist im Abnehmen").

2. )  —  (= „die Erbsenschote ist grindig").                — b)   —              (= „die Vogelscheuche ist wirkungs­ los"). 3. )  (= „der Herrensitz ist berüchtigt").   — —   b)   (= „das Herrenhaus ist ausstaffiert").   — —   4. )  —                 —            (= „ein idealer Arbeiter ist ar­ beitsam"). b)   —  ( — „eine ideale Welt ist nicht wirklich"). 5. )              —  (= „eine reife Frau ist bedacht­ sam"). b)   —  (= „der dichte Rauch ist undurchdring­ lich"). c)   ,   ,  . (I. S. Nikitin). 6. )    —  (= „der wohlriechende Duft der Linden ist berauschend"). b)   —  ( = „die Butterblume ist unge­ schminkt"). c)           . (S. A. Jesjenin). 7. ) b) c) d)

  —    —  (= „die oberste Gewalt ist   —  korrumpiert").   — 

8. )    (= „ein erfahrener Mann ist vor­ zuziehen"). b)   —   (— „[es ist] eine triviale Angele­ genheit [, es] ist sterbenslangweilig"). 9. )   — (= „das Himmelsgewölbe ist gestirnt"). b)   —  ( = „der himmlische Vater ist allmäch­ tig"). 10. )   —  ( — „der Zitronensaft ist säuerlich"). b)   —  (= „das Zitronat ist gesüßt"). 11. )    -  ( = „das deutsche Substantiv hat vier Kasus").

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

183

b)    —  ( = „das deutsche Adjektiv hat vier Kasus"). c)    —  ( = „das deutsche Ordinalzahlwort hat vier Kasus"). 12.

 ,       ( = „ein knochiger, kahlköpfiger und krummer Mann lag fast bewußt­ los").

13.

- ,  -, ,      ( = „die blutrote Beere, die kleinen festen dunkelgrü­ nen Blätter vergilben nicht einmal unter dem Schnee").

14.

  —   ( = „die ganze Blase ist kanonenvoll"). ! (= „Junge, Junge; nein, so was!").

15.

               !

(Puschkin)

,  ,   ! ,   ,   !   —  ,    .     ,     .        .    ,     .

(Bunin)

(Nasjimowjitsch)

16.    —   .    ,   . . .   ,  .   . 17. )    .     . b)  .

Die Frage ist nun: inwieweit ist die Stellung und damit die Funk­ tion des russischen Adjektivs von einem etwaigen generativen Satzgenerator bestimmt? Einen jeden, der unser soeben aufgestell­ tes russisches Satzmaterial durchgelesen hat, muß eine solche Frage schon von vornherein als unglaublich befremden. Manchmal steht das attributive Adjektiv im Russischen nach dem finiten Satzverbal, wobei es in Kopulasätzen dort auftritt, wo man in der Regel ein prädikatives Adjektiv erwarten würde (16).

184

PAPERS IN THEORETICAL LINGUISTICS

Manchmal steht umgekehrt das prädikative Adjektiv unmittel­ bar vor seinem substantivischen Korrelat (17). In solchen Fällen, wo das Adjektiv unmittelbar vor seinem sub­ stantivischen Korrelat steht, entscheidet manchmal die Sprech­ situation allein, inwieweit es als attributiv oder als prädikativ zu werten ist (17 b). Eine noch so oberflächliche Durchsicht des vorgeführten russi­ schen Satzmaterials muß zur Feststellung einer kaum widerleg­ baren Tatsache führen: unerforschlich sind die syntaktischen Wege des attributiven Adjektivs im Russischen (und sonstwo)! In der Regel stehen im Russischen attributive Adjektive vor dem substantivischen Beziehungswort (1a, b; 2a; 3a; 4a; 5a, b; 6a, b; 7a, c;8 a ; 9 a ; 10a). In einigen Fällen können sie nur nach dem substantivischen Be­ ziehungswort auftreten (11 a, b, c). Attributive Adjektive stehen im Russischen manchmal nach dem substantivischen Beziehungswort dann, wenn sie besonders her­ vorgehoben werden sollen (14). Die attributiven Adjektive können aber auch nach dem substanti­ vischen Beziehungswort stehen, ohne daß dadurch ein besonderer Nachdruck beabsichtigt ist (5 ,  , 7 b, d). Nachstellung des attributiven Adjektivs erfolgt auch oft aus stili­ stischen Gründen, z. B. in der Poesie (15). Mehrere attributive Adjektive stehen gern nach dem substantivi­ schen Beziehungswort (12). Mehrere attributive Adjektive stehen außerordentlich oft vor dem substantivischen Beziehungswort. Darin unterscheiden sie sich nicht von anderen Adjektiven. Mehrere attributive Adjektive können sowohl vor als auch nach dem substantivischen Beziehungswort stehen (13). Nachgestellte attributive Adjektive können den Sprachwert einer nachdrücklichen bzw. präzisierenden Aussage haben (3b). Durch Nachstellung des attributiven Adjektivs können dem ent­ sprechenden vorgestellten Attributiv gegenüber besondere seman­ tische Merkmale hervorgehoben werden (2 b, 4 b). Durch die Nachstellung eines attributiven Adjektivs kann eine ganz besondere Souveränität oder eine spezifische semantische Prägnanz eines bestimmten Begriffes erzielt werden (9 b, 10 b). Durch Nachstellung des attributiven Adjektivs kommt in gewis­ sen Fällen eine innere Agens → Patiens-Konzeption eindeutig zum Ausdruck (lc).

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

185

Ein attributives Adjektiv hat manchmal eine mehr oder weniger, in gewissen Fällen sogar eine völlig unterschiedliche Bedeutung, je nachdem ob es seinem substantivischen Beziehungswort vor- oder nachgestellt ist (8 a und b). Der generativen Grammatik bereitet das bunte Bild angeblich keinerlei Schwierigkeiten. Mit Hilfe der Transformationen leitet man aus der „Tiefenstruktur", die mit den formalen Mitteln der KS-Grammatik beschrieben wird, Strukturen ab, die zu verschie­ denen Oberflächenstrukturen führen. Man müsse, so wird behaup­ tet, verlangen, daß Sätze, die die gleichen terminalen Konstituenten enthalten sowie die gleiche Bedeutung haben, auch eine gemein­ same Konstituentenstruktur besitzen. Diese sei dann die „Tiefen­ struktur" der gegebenen Sätze. Die angenommene „Tiefenstruk­ tur" findet ausschließlich darin ihre Rechtfertigung, daß sie auf einer ganz bestimmten Ebene der Grammatik für die operationelle Überführung einer Tiefenstruktur in eine Oberflächenstruktur „adäquater", „notwendig" bzw. „praktisch" zu sein scheint. Um die Oberflächenstrukturen aus der „Tiefenstruktur" abzuleiten, seien Operationen notwendig, die Transformationen genannt werden. Im wesentlichen handelt es sich in diesem Zusammenhang um zwei Operationen: (1) eine Operation, die einen eingebetteten Satz (/man is wise/) auf Adj reduziert, (II) eine Operation, die, wenn erforder­ lich, Adj vor das substantivische Beziehungswort stellt. Wenn erforderlich, wohlgemerkt. Man fragt sich unwillkürlich: Wovon ist die etwaige Erforder­ lichkeit bestimmt? Von der Grammatik als Satzgenerator? Von einer auf logisch-axiomatischen Kalkülen aufbauenden rein operationellen Prozedur? Gewiß nicht. Die etwaige Erforderlichkeit ist allein von der Intuition bestimmt. Noam Chomsky gesteht es offen: „The structural descriptions assigned to sentences by the grammar, the distinctions that it makes between well-formed and deviant, and so on, must, for descriptive adequacy, correspond to the lin­ guistic intuition of the native speaker (whether or not he may be immediately aware of this) in a substantial and significant class of crucial cases" (Aspects [1965], S. 24). Dank der Intuition ist es mög­ lich, die Stellung des attributiven Adjektivs im Russischen (und anderswo) zu „beschreiben" (denn das ist es, was das Wort „gene­ rieren" bedeutet!). Die Transformationen entlarven sich selbst als reine Taschenspielerei. Sie sind unwissenschaftlich, weil sie außer­ stande sind, etwas Generelles, geschweige denn irgend etwas Neues über ihren Gegenstand auszusagen. Wenn jemand glaubt, die Stel-

186

PAPERS IN THEORETICAL LINGUISTICS

lung des attributiven Adjektivs sei im Russischen oder in irgend einer anderen menschlichen Sprache von futilen Transformationen bestimmt, dann irrt er sich. Denn die Intuition kann nichts „be­ stimmen" — das liegt außerhalb ihrer Kompetenz! Die transformationelle generative Prozedur wird womöglich noch sinnloser, wenn es um die Analyse von Sätzen geht, in denen die Bedeutung eines attributiven Adjektivs davon bestimmt ist, in­ wieweit es vor oder nach dem dazugehörigen substantivischen Be­ ziehungswort steht. Man betrachte ein französisches Satzpaar: 1. Les pauvres gens sont malades. 2. Les gens pauvres sont malades.

Für beide Sätze gilt die folgende Tiefenstruktur: les gens / gens sont pauvres / sont malades.

Die gegebene Tiefenstruktur kann nur den zweiten Satz unseres Satzpaares generieren. In der prädikativen Stellung hat das Adjek­ tiv pauvre bekanntlich nur die Bedeutung „arm an irdischem Gut". Diese Bedeutung hat das Adjektiv auch, wenn es attributiv nach­ gestellt ist. Sie trifft also ohne weiteres für unseren zweiten Satz zu. Für unseren Satz Nr. 1 muß aber augenscheinlich ein generativer Kniff vorausgesehen werden, der den schon allzubekannten und endlosen Schlupfmanövern der generativen Grammatik kaum etwas zu wünschen übrig lassen wird. Das freie Spiel des Manövrierens scheint das großzügige Privilegium der generativen Gramma­ tik zu sein: alles dient der Theorie, nichts dient der Sprache; daraus entspringt die ausgeprägt normative Haltung, die für die transformationalen Operationen so symptomatisch ist. In einigen Sprachen können Substantive für attributive Adjek­ tive eintreten:        (= „der starke Fischer und der schneidige Flieger waren Nord­ länder")-

Einem Generativisten muß jetzt der Gedanke kommen, diese Substantive beruhten auf Transformationen von eingebetteten Prädikativsätzen, die durch ein Relativum eingeleitet sind. Daß eine solche Annahme überflüssig, unsinnig und geradezu irrefüh­ rend ist, beweist zur Genüge das folgende polnische Beispiel: Czy wolno zarznac kurę, matkę piskląt? (= „Darf man eine Henne schlachten, die Mutter von Kücken ist?")

Die Apposition matkę piskląt hat ungefähr den Wert eines prä­ dikativischen Relativsatzes. Die Intuition sagt einem das. Die Spra-

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

187

che sagt uns etwas ganz anderes. Die Sprache, die unseres Erachtens der eigentliche Gegenstand der Sprachwissenschaft ist, sagt uns, daß matkę piskląt als reine Apposition aufzufassen ist, die sich in bezug auf Kasus, Numerus und Genus nach dem Objekt kurę richtet. Nichts mehr und nichts weniger. Mit der Sprache läßt sich nicht feilschen. Oder will man wirklich ernsthaft den Akkusativ der Oberflächen­ struktur in einen prädikativen Nominativ der Tiefenstruktur trans­ formieren, um ihn dann später wieder in einen Akkusativ zurückzugenerieren? In dem Fall muß ich fragen: Wozu bezahle ich Steu­ ern? Schon die alten Inder haben weitaus unterhaltendere Zeitver­ schwendungen erfunden. Daß die attributiv verwendeten Substantive im Russischen kei­ neswegs als zutiefst prädikativ zu werten sind, zeigen mit aller Deutlichkeit Sätze wie die folgenden, wo das Attributivum durch einen Bindestrich an seinen Träger gelötet ist: -    (= „die ruckhafte Libelle sang den schönen Sommer"). -       ( = „das kleine Bäuerchen führte mit wichtiger Miene sein Pferd am Zügel daher").

Eine generative Analyse des Eskimoischen führt zu bedrücken­ den Ergebnissen. Auf diese Tatsache ist schon anderswo mit lin­ guistischer Kompetenz und wissenschaftlichem Verantwortlich­ keitsgefühl hingewiesen worden 8 . Hier sei das schon Vorgeführte bloß durch ein paar Beispiele komplettiert, die indoeuropäischen Sätzen entsprechen, die nach dem Muster Subjekt mit attributivem Adjektiv (+ Kopula) + adjektivisches Prädikativ aufgebaut sind: Angut silatôq ilumôrpoq (= „Mann klug-seiend ehrlich-ist"). Arnaq pínersoq píkorigpoq (= „die schöne Frau ist tüchtig"). Frau / schön seiend / tüchtig ist Nunarssuaq íssigpoq (= „das große Land ist kalt"). großes Land / (es) ist kalt Niuvertoq náparsimassoq píkorigsorssüngíkaluarpoq Kolonienvorsteher I krank geworden seiend I ist freilich nicht be­ sonders tüchtig („der kranke Kolonienvorsteher ist freilich nicht besonders tüchtig").

Alle vier eskimoischen Sätze haben die folgende Tiefenstruktur: S / \ NP VP 8 Jørgen Rischel, „Some Characteristics of Noun Phrases in West Greenlandic", in: Acta Linguistica Hafniensia, Vol. XIII, No. 2, Kopenhagen 1971.

188

PAPERS IN THEORETICAL LINGUISTICS

Damit ist aber nichts Neues oder ein Mindestmaß an irgendwie aufschlußreicher Information über die in Frage kommenden Sätze im Eskimoischen gegeben worden. Eine generative Analyse über­ sieht völlig die semantisch gesehen außerordentlich wichtige Tat­ sache, daß, was in indoeuropäischen Sprachen als innerhalb der Designations- und Qualitätszonen realisierte adjektivische Deter­ minationen hervortritt, sich im Eskimoischen vielfach in der Ein­ kleidung von verbalen Determinationen manifestiert. Das Eskimo­ ische muß in prädikativen Verbalformen bzw. in Partizipien von solchen das ausdrücken, was zum Beispiel indoeuropäische Spra­ chen durch Adjektiva auszudrücken vermögen. So ist pínersoq in­ transitives Partizipium zum Verbum pínerpoq („ist schön"). Dem­ entsprechend ist píkorigpoq ein prädikatives Verbum in der 3. Pers. Sing. mit der Bedeutung „ist tüchtig". Die beiden Elemente píner­ poq und pikorigpoq liegen somit innerhalb der von uns anderswo erörterten WE?SE:KEW¿-Zone9. Sie bestehen aus einer eigentli­ chen Determination, die mit einem interpolierten Nucleus in der 3. Pers. Sing. Indikativ (-poq) versehen ist. Eine generative Analyse übersieht, was mit ganz schlichten und menschlichen Worten gesagt werden kann, und zwar die nicht zu unterschätzende Tatsache, daß ein ieur. Substantiv mit attributivem Adjektiv außerordentlich oft im Eskimoischen in einem Substantiv sein semantisches Gegenstück findet (sikuaq = „dünnes Eis", „neues Eis"; tissat = „gespaltener Fisch"; maneringneq = „ruhiges Wasser" usw.). Was in indoeuro­ päischen und vielen anderen Sprachen in adjektivischen Determi­ nationen zutage tritt, wird im Eskimoischen in weitem Ausmaße in Suffixen realisiert (nuna = Land, nunarssuaq = „ein großes Land"). Was in europäischen und zahlreichen anderen Sprachen als Satzmodifikatoren hervortritt, wird im Eskimoischen weitgehend mit­ tels prägnanter Suffixe ausgedrückt (pikorigpoq = „er ist tüchtig"; píkorigsorssungíkaluarpoq = „er ist freilich nicht besonders tüch­ tig"). Dasselbe trifft für diejenigen Inhaltselemente zu, die sich in anderen Sprachen als Adverbia manifestieren. Wie steht es überhaupt mit dem universellen Wert der generati­ ven Operationen? Welche Bedeutung kann man den diesbezügli­ chen Visionen ihrer Ausüber beimessen? Versuchen wir, anhand 9 Niels Danielsen, „Die Relativa im Neuhochdeutschen — und anderswo. Eine Vorstudie zu einer konstitutionellen Sprachtheorie", in: Språkliga Bidrag, Vol. 6, Nr. 27, Lund 1972; Niels Danielsen, „Zu den Nucleuskonstruktionen in der mensch­ lichen Sprache. Vorbemerkungen zu einer Semasiosyntax", in: ZPSK 1973.

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

189

von drei oder vier illustrativen Beispielen der höchst fraglichen Sache ein ruhiges Ende zu bereiten. Wir wollen dabei von zwei ganz trivialen deutschen Nominalsätzen ausgehen: I. Die Züge sind geheizt. II. Der Himmel ist klar.

Für die beiden deutschen Sätze gilt natürlich ein und dasselbe Satzstemma:

Im Ungarischen heißt Satz I: A vonatok fűtve vannak (= „die Züge sind geheizt").

Das ungarische Satzstemma sieht folgendermaßen aus: S /

\ NP VP / \ / \ det N Adv V

Der Satz hat nämlich im Ungarischen die folgende Struktur: Be­ stimmter Artikel + Substantiv (NP) + Verbaladverb + Existenz­ bzw. Situationsverb (VP). Die generative Grammatik kann das nicht aus dem Felde schlagen: sie will bloß eine neue Gesetzmäßig­ keit gefunden haben, eine unter Zehntausenden oder Hunderttau­ senden, die uns nichts über das hinaus sagen, was wir jeweils schon wußten. Die generative Grammatik irrt sich: zum ersten ist hier durchaus nicht von einer Gesetzmäßigkeit die Rede, zum andern hat sie die generative Grammatik nicht gefunden. Man kann schon in jeder guten ungarischen Grammatik eine vernünftige Darstel­ lung dieser ungarischen Sonderkonstruktion lesen. Im Japanischen heißt Satz II: Sora wa harete imasu (= „der Himmel ist klar").

Wir stellen fest, daß der japanische Satz folgendermaßen struk­ turiert ist: Substantiv + Partikel (NP) + konnektives Verbum + terminaler Nucleus. Die generative Grammatik hat nur eine Mög­ lichkeit: sie muß uns nochmals die schon längst bekannte Wahrheit erzählen, denn die kann sich — groteskerweise — keineswegs von der gegebenen Oberflächenstruktur losreißen. Das kümmerliche

190

PAPERS IN THEORETICAL LINGUISTICS

Resultat ist eine neue Gesetzmäßigkeit unter Zehn- oder Hundert­ tausenden:

Jetzt fängt es wirklich an, erregend zu werden! Betrachten wir noch einige wenige völlig triviale japanische Prädikativsatzstruk­ turen, die zur Genüge die Bedingtheit der sinnlosen Tiefenstruktur durch die frei waltende Oberflächenstruktur illustrieren: Furukawa san wa hijō ni kimae ga yoi desu (-= „Herr Furukawa ist sehr großmütig").

wa („Subjekt 1") ga („Subjekt 2")

Nakamura san no okusama wa inu ga taihen o-suki de gozaimasu (= „Frau Nakamura, Hunde sind gefällig" : „Frau Nakamura hat Hunde sehr gern").

Tokugawa san wa yoku irasshaimashita (= „Herr Tokugawa ist willkommen").

GEGEN DIE GENERATIVEN TIEFENOPERATIONEN

191

Minasan go-jöbu de irasshaimashita (= „alle bei dir sind wohl­ auf").

(Man beachte die Vielfalt der Kopulanuclei; vgl. im übrigen oben S. 248—251)

In dieser Weise könnte man fortfahren. Und so heißt es ja auch immer wieder von generativer Seite, daß die Möglichkeiten der transformationalen generativen Theorie unendlich, unerschöpflich, allumfassend, allversprechend, grenzenlos verheißungsvoll seien. Man rüste sich bloß einige Jahrzehnte oder Jahrhunderte mit Ge­ duld! Das Tausendjährige Reich der Linguistik steht vor der Tür. Dieses Reich wird sehr lange auf sich warten lassen. Dank ihrer logisch-axiomatischen Prozedur hat sich die generative Grammatik für immer der Möglichkeit beraubt, die menschliche Sprache wis­ senschaftlich belangreich zu beschreiben. Dank diesem metalingui­ stischen Verfahren treibt sie hilflos ins Uferlose und in das Atomistisch-Triviale. Dank ihrer rein operationellen Beschreibungsweise ist sie dazu verurteilt, eine unendliche Reihe von trostlosen Natio­ nalgrammatiken aufzustellen, die unser schon längst erworbenes Wissen von der Sprache höchstens verpfuschen. Mit allgemeiner Linguistik hat sie nichts zu tun. Das Märchen vom generativen Würdenträger hat schon im vori­ gen Jahrhundert mein berühmter Landsmann Hans Christian An­ dersen treffend geschrieben. Es heißt „Des Kaisers neue Kleider" und endet bekanntlich folgendermaßen: „Aber er hat ja gar nichts an!" sagte endlich ein kleines Kind. „Hört die Stimme der Unschuld!" sagte der Vater; und der eine zi­ schelte dem andern zu, was das Kind gesagt hatte. „Aber er hat ja gar nichts an!" rief zuletzt das ganze Volk. Das ergriff den Kaiser, denn das Volk schien ihm recht zu haben, aber er dachte bei sich: „Nun muß ich aushalten". Und die Kammerherren gingen und trugen die Schleppe, die gar nicht da war. Inwiefern ist die generative Scheinlehre überhaupt imstande, eine linguistisch adäquate Beschreibung menschlicher Sätze zu geben? Wie beschreibt man meinetwegen menschliche Sätze, die

192

PAPERS IN THEORETICAL LINGUISTICS

keine Rücksicht auf die generativen Absolute NP und VP nehmen? Betrachten wir fünf subjektlose Sätze: 1. Ist schon weg! 2. Abest (lat. „er ist weg") 3. Gorąco mi (poln. „mir ist heiß") bzw. Vrlo je toplo (serbokr. „es ist sehr warm") 4. Pluit (lat. „es regnet") 5. Mir ist nicht rechthaberisch und streitsüchtig zu Mute.

Das Fehlen des Subjekts muß in jedem der fünf Sätze anders erklärt werden. Im ersten Satz könnte es schon verantwortet wer­ den, von einer Tilgung desselben zu sprechen. Im zweiten Satz ist dieser Ausweg kaum so plausibel, da das Subjekt in der Regel ohnehin unausgedrückt bleibt. Im dritten Satz klingt das „Wetter" als logisches Subjekt mit, wird aber niemals syntaktisch vertreten. Im vierten Satz regnet „es" — in der Verdeutschung. Im Kymrischen regnet „sie" (y mae hi'n bwrw glaw). Im Lateinischen reg­ net „?". Der Jupiter? Angeblich ja. Im Isländischen besteht kein Zweifel: hier regnet „er" (kann rignir). (Was ist nun gestrichen, „er", „sie" oder „es", im finnischen sataa, im rumänischen plouä oder meinetwegen im russischen  ?) — In unserem Satz Nr. 5 ist kein Subjektsbegriff denkbar. In allen fünf Fällen werden die Generativisten von einer Tilgung des Subjekts schlichtweg re­ den. Die einen nennen das eine zweckmäßige Großzügigkeit, denn kleinlich ist man ja nicht. Wir anderen nennen das eine himmel­ schreiende Fahrlässigkeit. Ich erinnere mich an die polnische Version einer alten biblischen Legende: Pewien mędrzec rozpoczął: naukę po ukończeniu czterdziestego roku zycia. Gdy zapytano go, czy nie jest to zbyt pózno, przytoczyl on przyklad, ze gdy woda przez lata spada malymi kroplami na kamień w jedno i to samo miejsce, to wyzlobi wglębienie. Wiedza natomiast jest twardsza niz woda, zas moja głowa bardziej chłonna od kamienia, odpowiedział mędrzec. Moral: na naukę nie jest nigdy za pózno.

Credo quia absurdum scheint die Erkenntnishaltung der genera­ tiven Schule zu sein. Das ist an und für sich auch ganz legitim. Kaum zu rechtfertigen ist aber der generative Theoretiker, der sich Linguist nennt, ohne den tieferen Sinn dieser lateinischen Sentenz jemals voll erkannt zu haben. Die Sprachwissenschaft muß aber eine Theorie haben, werden die Generativisten erwidern. Dazu kann ich nur antworten: In die­ sem Punkte sind wir uns völlig einig. — Aber lassen Sie die Ihre fallen!

KINETIC ANALYSIS OF SENTENCES It seems that we have now come to a point where we can start analysing sent­ ences. This has always been difficult. The task has not become easier since the rise of various transformational schools of syntax which have deprived the great majority of linguists of independent thought. We consider any theory which speaks of learning models 1 and of language having no objective existence apart from its mental repres­ entation 1 as false, and what is even worse: totalitarian. Totalitarian because it does not allow for spontaneous, unguided activity which might produce unforeseen results. We believe the depth of human language is as unfathomable as man's own self of which it is the auditive means of communication. This is our firm belief which is still waiting for its refutation. What we can do is to consider any human sentence as expressive of a possibility rather than of a necessity. Nothing is necessary in the strict sense of this word as long as man is free to be born in the jungles of Cameroun, in the Isles of the Aleuts, or in an English manor-house, and as long as the woman giving birth to the child in the manor-house may be a native Aleutian, a girl from Yaounde, or a Russian actress. Whether the mother returns to her home country with the child or not does not change the fact that the child's brain is equally open to any language which might be imposed on it by its mother's tongue, by its birth­ place, or by its mother's dispositions or lack of dispositions. One thing is sure: you cannot say a thing of universal value about human language by analysing English syntagms and claiming that the arrangement of English sentences is, in some way or another, the clue, given a priori, to all other languages. It is incredible that American linguistics has succeeded in making hordes of linguists accept the validity of what you feel tempted to call the noble intellectual outrage upon the needy intellectualism of the 60's and 70's. 2 No member of our species has ever pointed out so clearly as Søren Kierkegaard why this totalitarian muddle is utterly wrong. Actually, he was 1 Cf. Noam Chomsky: Language and Mind (Enlarged Edition), New York, Chicago, San Francisko, Atlanta 1972 (p. 169). 2 Cf. F. A. Hayek: The Road to Serfdom, London 1944 (p. 122: "Probably it is true that the great majority are rarely capable of thinking independently, that on most questions they accept views which theyfindready-made, and that they will be equally content if born or coaxed into one set of beliefs or another. In any society freedom of thought will probably be of direct significance only for a small minority: But this does not mean that anyone is competent, or ought to have power, to select those to whom this freedom is to be reserved. It certainly does not justify the presumption of any group of people to claim the right to determine what people ought to think or believe.").

194

PAPERS IN THEORETICAL LINGUISTICS

the thinker who once and for all anticipated the generative disaster. 1 "All coming into being comes about through possibility, and not through necessity" is one of his most weighty arguments for human freedom, including the freedom of human language. No knowledge of language has been internalized in some manner by the language user, as Chomsky will have it. A language user knows that he has a language, and he has to work hard in order to obtain more or less the knowledge of the particular language which he has been brought up with, but often his knowledge of foreign tongues will be deeper and more comprehensive than the knowledge he has of his mother tongue. A human being has no more internalized knowledge of his language than he has of his own self. So our only choice is to deal with our language as we deal with other natural things. We can only describe it when it has come into being, because only there start the symptoms of lawfulness, only there begins the necessity.2 In order to describe the genesis of a sentence in an exhaustive manner, we shall operate on three planes, 1) a semantic plane, 2) a semasiosyntactical plane, and 3) a syntactic plane. The semasiosyntactical plane is split up into the upper semantic level and a lower syntactic level: 1 Cf. Seren Kierkegaard's "Mellemspil" in the "Philosophiske Smuler" (Seren Kierkegaard's Samlede Værker, udgivne af A. B. Drachmann, J. L. Heiberg og H. O. Lange, vol. IV, Copenhagen 1923) (p. 265: "Hvorledes forandres det, som bliver til; eller hvilken er Tilblivelsens Forandring Al anden Forandring forudsætter, at Det, med hvilket Forandringen foregaaer, er til, selv om Forandringen er den at ophøre at være til. Således ikke med Tilblivelsen ; thi dersom det Tilblivende ikke i sig selv forbliver uforandret i Tilblivelsens Forandring, saa er det Tilblivende ikke d e t t e tilblivende, men et andet, og Spørgsmaalet forskylder en idet den spørgende i det givne Tilfælde enten med Tilblivelsens Forandring seer en anden, som forstyrrer ham Spørgsmaalet, eller han seer feil af det Tilblivende og bliver derfor ikke istand til at spørge. Dersom en Plan, idet den bliver til, forandres i sig selv, saa er det ikke denne Plan der bliver til; bliver den derimod uforandret til, hvilken er da Tilblivelsens Forandring? Denne For­ andring er da ikke i Vaesen men i Vasren, og er fra ikke at vaere til, til at være til. Men denne Ikke-Væren, hvilken det Tilblivende forlader, maa jo ogsaa vaere til, thi ellers "forblev det Tilblivende ikke uforandret i Tilblivelsen", uden forsaavidt det slet ikke havde vasret, hvorved da Tilblivelsens Forandring atter af en anden Grund vilde blive absolut forskjellig fra enhver anden Forandring, da det slet ingen Forandring Var; thi enhver Forandring har altid forudsat et Noget. Men en saadan Vaeren, der dog er Ikke-Vaeren, det er jo Muligheden; og en Vaeren, der er Vaeren, det er jo den virkelige Vaeren, eller Virkeligheden ; og Tilblivelsens Forandring er Overgangen fra Mulighed til Virkelighed. Kan det Nødvendige blive til? Tilblivelse er en Forandring, men det Nødvendige kan slet ikke for­ andres, da det altid forholder sig til sig selv, og forholder sig til sig selv paa den samme Maade. At Tilblivelse er en L i d e n , og det Nødvendige kan ikke lide, ikke lide Virkelighedens Lidelse, hvilken er denne, at det Mulige (ikke blot det Mulige, der bliver udelukket, men selv det Mulige, der bliver antaget) viser sig Intet i det 0ieblik det bliver virkeligt; thi ved Virkeligheden er Muligheden t i l i n t e t g j o r t . Alt, hvad der bliver til, beviser netop ved Tilblivelsen, at det ikke er nødvendigt; thi det Eneste, der ikke kan blive til, er det Nødvendige, fordi det Nødvendige er. Er da Nødvendigheden ikke Eenhed af Mulighed og Virkelighed? - Hvad skulde dette sige? Mulighed og Virkelighed ere ikke forskjellige i Væsen, men i Vaeren; hvorledes skulde der af denne Forskjellighed dannes en Eenhed, der var Nødvendighed, hvilken ikke er Vaerens Bestemmelse, men Væsens Bestemmelse, da det Nødvendiges Vaesen er at vaere. I saa Fald vilde jo Mulighed og Virkelighed, ved at blive til Nødvendighed, blive til et absolut andet Vaesen, hvilket ingen Forandring er, og vilde, ved at bUve til Nødvendighed eller det Nødvendige, blive til det Eneste, der udelukker Tilblivelsen, hvilket er ligesaa umuligt som selvmodsigende. (Den aristoteliske Sætning "det er muligt", "det er muligt ikke", "det er ikke muligt". - Leeren om falske og sande Sætninger (Epicur) griber 'forstyrrende herind, da der reflekteres paa Vaesen, ikke paa Vaeren, og der følgelig ad den Vei Intet naaes med Hensyn til Bestemmelsen af det Tilkommende)." 2 That linguistic freedom:linguistic necessity cannot possibly be described in terms of deep structure operations has been shown by several scholars, for example by Hans-Peder Kroman in his excellent article "Wortstellung in der Transformationsgrammatik des Deutschen", in: Sprache der Gegenwart (Schriften des Instituts für deutsche Sprache), 24, Linguistische Studien IV (Festgabe für Paul Grebe zum 65. Geburts­ tag), Teil 2, Düsseldorf 1973.

KINETIC ANALYSIS OF SENTENCES

195

upper syntactic level nomothetic line lower syntactic level _ _ upper semantic level

index line .,

nomogenetic line

lower semantic level kinetic limit sub quo nihil linguisticum 1 A constitutional analysis of a sentence starts with a description of the lower semantic level. Our only guarantee of the furnishing of this level is the sentence we have in front of us on the upper syntactic level. We have no reason whatsoever to believe in deep structures or transformations, because all we know about a sentence is ultimately to be deduced from the arrangement of its surface structure. This does not prevent us, however, from describing the genesis of a sentence starting with the lower semantic level. Only this procedure presupposes a broad abstract basis of generalization. It requires that you have dealt with languages of many different families and syntactic structures, and a profound research in the field of basic semantic cate­ gories must be foreseen. Any deduction inferred from the outcome of such invest­ igations must have as its lodestar the absolute demand for the highest possible empirical content (including the demand for the highest attainable degree of univers­ ality, and the demand for the highest attainable degree of precision). This must be so because having a language is primary to being able to discuss a language. Or : the language we describe is a language we have already got and without which we cannot describe anything whatsoever, including language itself. Let us now consider the lower semantic level of a given English sentence. The means we have for this purpose are still very green and untested, and they have to a great extent been drowned by preachers of antisemanticism or by people who took a greater interest in futile discussions about personal competence or about "more adeq­ uate problems". We cannot see that any linguistic analysis may have the faintest chance of becoming exhaustive and "to the point" if it does not comprise semantics. 2 1 A German dative or a Russian Imperfect is no more psychological than an Icelandic definite article or a synthetic transversal relation in Eskimo. (Nevertheless Noan Chomsky and his school of syntactic deep-sea explorers claim that mental representations (i.e. psychology) are the only objective existence of language (!) (cf. Noam Chomsky: Language and Mind (Enlarged Edition), New York, Chicago, San Fran­ cisco, Atlanta 1972 (p. 169)). 2 Cf. Niels Danielsen: Status und Polarität im Gotischen, im Lichte des Kymrischen dargestellt, Odense University Studies in Linguistics, Vol. 2, Odense 1968; Niels Danielsen: Die Frage. Eine sprachwissenschaft­ liche Untersuchung, Det kongelige Danske Videnskabernes Selskabs Historisk-Filosofiske Skrifter 7,1, København 1972; Niels Danielsen: Die Relativa im Neuhochdeutschen - und anderen Sprachen. Eine Vorstudie zu einer konstitutionellen Sprachtheorie, in: Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikations­ forschung, Vol. 28, l/75, Berlin 1975. We shall have to use some abbreviations from this article in our attempts to describe the lower semantic level in terms of semantic unities (Z design = zone of designation, Z pers = zone of person, D = determination, ID = demonstrative identifier, IZ = personal identifier, etc.)

196

PAPERS IN THEORETICAL LINGUISTICS Let the following lower semantic level be given: positive NE

On this lowest level of our description of the sentence genesis only the semantic ingredients as such are taken into consideration. Nothing is yet organized linearly. We find ourselves in a realm of unformed entelechy. We can see that the sentence is positive in meaning ("logically"). We can see that it has two adverbs. It is evident, too, that it has a Π. But we cannot see whether it is a transversal S- or R-relation. Or perhaps a Q-relation? Any semantic content may be expressed in the form of any of these three relations. This is exactly what justifies the introduction of a lower semantic level in our analysis. It is the prior-to-TR-level. The [A] and the [O] have not yet assumed their transversal functions. The notation m used with the 77 reveals modality. When we cross the nomogenetic line we enter the upper semantic level. The nomogenetic line is marked with the notations (S, R, or Q) of the transversal relation in question. On the upper semantic level the semantic ingredients are organized linearly according to their function in the given transversal relation. The order of the elements is completely optional on this level of our description : upper semantic level S: + [A] ~ [] ~ [O] ~ [adv] ~ [adv] nomogenetic line lower semantic line positive [O]

[adv] [A]

[Π]

[adv]

Alloyings take place on the upper semantic level. As we shall see clearly from the following examples, alloyages must be considered as prior to the final choice of stasis. The choice of stasis may be a rather complicated affair. In our sentence the stasis

KINETIC ANALYSIS OF SENTENCES

197

is already given: the nomogenetic line tells us that our sentence is to be materialized in the form of a transversal S-relation. So the [A] automatically takes the place of the stasis on the upper semantic level. As this [A] is not alloyed in our example, nothing can prevent it from being the indisputed stasis on the lower syntactic level:1 lower syntactic level S ~ fV ~ oV ~ O ~ adv ~ adv ...

index Une

upper semantic level +[A] ~ [Π m ] ~ [O] ~ [adv] ~ [adv] S:

nomogenetic Une

lower semantic level positive NE [O]

[adv] [A]

[Π m ]

[adv]

Apart from the final settUng of the stasis, the most important thing happening on the lower syntactic level is the basic allotment of the 77. It may remain undivided, or it may, as in this case, be spUt up into a nucleus (or more nuclei) + an operative verb (or more operative verbs, of which some may be nuclei of second, third, or fourth order). The former [O] is described as O1 (omikron one) on this level: the epistasis is taken over from the upper semantic level. On the upper syntactic level the elements are adjusted according to a particular nomos. The nomogenetic line carries the sentence notation (cf. 126). The nucleus and the operative verb are distributed according to the lawfulness of the given language. The grammatical information given on the upper syntactic level differs in amplitude from one language to the other. Our language gives very little grammatical information in the sentence we operate with: upper syntactic level You ought to put that money there tomorrow. S fV oV O adv adv ßZ propositive

nomothetic line

lower syntactic level S ~ fV ~ oV ~ O1 ~ adv ~ adv index line 1

In the Hungarian alloyage szeretlek the [A] and the [O] have both become alloyed with the II. They may both be reintroduced, though, on the lower syntactic level (én szeretlek téged. In this case, the stasis finally settles where it was supposed to from a broader syntactical view upon "abundant" Hungarian syntagms).

198

PAPERS IN THEORETICAL LINGUISTICS

upper semantic level +[A] ~ [Π m ] ~ [O] ~ [adv] ~ [adv] S: lower semantic level

nomogenetic line

positive NE [adv]

[O] [A] [Π m ]

[adv]

The stasis may become elided from the lower to the upper syntactic level (in our example this is irrelevant). The grammar of a given language is extra information from the nomogenetic line to the upper syntactic level. Some languages are completely devoid of grammar (e.g. Pidgin English), others are more or less rich in grammar (e.g. Latin and Danish) and some are abundant (e.g. Basque, or Georgian). All grammar is universal property, but particular languages differ as to informativity. Grammatical extra information given in one language may be concealed in another : the grammatical form is individual for every particular language, and so is its syntactic arrangement. What is universal is to be found below the nomothetic line. The generative grammarians seem to over­ look this simple fact. They try, in earnest, to analyse human language universally starting with the accidental syntax of a language which they happened to learn at school. You need not be ingenious to realize what the outcome of such operations is going to be: unlimited multiplications of truisms walking on an everlasting safetynet of endless cramped English (or, at the most, Indoeuropean) deep structures. He who wants to make linguistics a proper science of orderly habits must obviously set to work in exactly the opposite way. In order to say something of universal value about human language he must study human languages, and not constructed English sentence patterns. What he must find, first of all, is all that is relevant, linguistically relevant, below the nomothetic line. Then, a d only then, will he be able to describe the syntactic filigree above this Une in terms of what he found. A man who wants to make a map of France does not start by making a digging machine meant to analyse those grains of sand which fit into it. On the contrary, he must be aware of the far-reaching perspectives which only journeys through the land and hard work with the ruler and the theodolite can give him. Linguistic perspective is hardly attainable without a semasiosyntactic wide-angle lens. With the accent no less on semasio than on syntactic. What we need, most of all, is a limited amount of concise semantic notations (or universally adequate basic semiotic categories). We already have a very limited stock of sentence semantic manifestations. 1 We have an exact classification of the basic 1 Cf. Niels Danielsen : Status und Polarität im Gotischen, im Lichte des Kymrischen dargestellt, Odense University Studies in Linguistic vol. 2, Odense 1968 ; Niels Danielsen : Zum Wesen des Konditionalsatzes — nicht zuletzt im Indoeuropäischen: Odense University Studies in Linguistics vol. 1, Odense 1968; Niels Danielsen: Die Frage. Eine sprachwissenschaftliche Untersuchung, Copenhagen 1972.

KINETIC ANALYSIS OF SENTENCES

199

constitutional elements in human language. 1 And we have our nuclei, our transversal relations, and our limited number of sentence notations (cf. pp. 127—28) by means of which the component semantic entities of a sentence are realized. So what on earth are we waiting for? We can now analyse sentences either by feeding a given semantic lower level with adequate notations, or by looking at the sentence as we meet it auditively/visually. Let us take any German (or Turkish, or Quechua, or for that matter English) sentence and see what it tells us : Was für eine Dummheit du da begangen hast! The sentence is characterized by an -marker. This implies that it is a con­ stitutional sentence. It is exclamative. It is a βZ-sentence (the subject comes before the verbum finitum). It is propositive (the word not does not occur in it). And it is positive (in meaning). It is a plain transversal S-relation. The  of this transversal relation is begangen hast built on the presential nucleus hast of the nucleus haben (the emanent reflected image 2 of sein (= "to be")). The operative verb is begangen. The stasis of the TR-S is du(lZ 2nd person singular), and the epistasis is WAS für eine of the sentence. This is the much used circum­ Dummheit which contains the locution WAS für ein (  ­ + indicator für + the DK (the existential dispositioner) of the zone of designation) for the WELCH/- of the zone of designation. It is attributed to the determination Dummheit (from the zone of species). The sentence contains one adverb, the ID 2/3 pers. da (of the locus zone). The element da, though, is not used here in its strict local sense : the locus zone shows, once again, its strongly expanding nature (→zone of time, zone of species, etc.). The element da thrusts itself in between the S du and the oV + fV, thus cutting the mesonomic part of the sentence into halves. The endonomic part of the sentence is hast and the ectonomic part is was für eine Dummheit . . . da . . . . Of these two only was für eine Dummheit has an ectostatic relation to the verb, whereas da is purely adverbial (or circum­ stantial). No syntactic permutations are allowable in this sentence : the sentence follows a particular German nomos which allows the eight words of the given theme to be materialized only in the actual order. The sentence follows a schedule which you will observe in legions of German exclamative sentences starting with an + a determination in the O1 position. In German, the sentence could only be formed syntactically in this very manner, because a German brain is brought up with this special model valid for -sentences. But no transformation! What for? Can't we 1 Niels Danielsen: Die Relativa im Neuhochdeutschen - und anderen Sprachen, in: Zeitschrift für Phone­ tik, Sprachwissenschaft und Kommunikationsforschung Heft 28/1 '75, Berlin 1975. 2 Cf. Niels Danielsen: Zu den Nucleuskonstruktionen in der menschlichen Sprache. Vorbemerkungen zu einer Semasiosyntax, in: Sprachwissenschaft, Vol. 3, Heidelberg 1976 (-77); Niels Danielsen: Zur Univer­ salität der Sprache, in: Sprachwissenschaft, Vol. 1, Heidelberg 1976; Niels Danielsen: Semasiosyntaktische Universalien im Finnischen, in: Ural-Altaische Jahrbücher 66, 1974.

Hist. Filos. Skr. Dan.Vid. Selsk. 7, no. 4.

200

PAPERS IN THEORETICAL LINGUISTICS

say more than enough of universal value without transformations? And, by the way, what would a transformation tell us? But we have not yet quite finished our analysis of the sentence. We have not looked at its grammaticality, or in other words : we have not investigated which gram­ matical universals are registrable on the upper syntactic level. The signals it gives us from the nomogenetic Hne are the following:

<

Was für eine Dummheit du da begangen hast. -e du angen.. .st>

(the grammar of the sentence)

The signal -e stands for attributivity (nominative/accusative). The accusative (O 1 ) is chosen because its counter-stasis turns out to be du (a clear-cut stasis because nominative). The signals -angen + -st indicate perfect tense/ 2nd person singular. The nuclear arrangement of the  is of universal interest, too, but it occurs on another level of the sentence genesis (nuclei are mainly syntactic tools which are introduced immediately on the lower syntactic level). This is about all that we can say about our sentence; and should it turn out to be all there is to be said about it, no matter-we see no reason to say more. Well, we forgot about the phonology (or phonetics, as some have it) and the intonation. They are simply the physically measurable means by which everything below the nomothetic line is expressed. For many years they were considered the only disciplines worthy of linguistic research by too many theoreticians. We can now describe the genesis of our sentence starting with the upper syntactic level or, the other way round, starting with the lower semantic level. We have the means to feed that level with the necessary notations. It must be one of the most important tasks of the linguists of coming generations to ameliorate our semantic apparatus towards complete precision: Was für eine Dummheit du da begangen hast! -e du -angen -st upper syntactic level

upper semantic level

KINETIC ANALYSIS OF SENTENCES

201

lower semantic level p o s i t i v e / G (onstitutional)

[O =

design/qual

(: D spec = Dummheit]

This is to show, especially, the role of the nomothetic line. The nomothetic line is the syntactic filter between universal nomos and particular nomos. The nomothetic line shows us what is universal and what is not. It is the line which puts, so to say (cf. nomothetic), or transfers the given universale into the sphere of a particular syntactic legislation. In our example the legislation turns out to be very strict. It allows the elements of the lower syntactic level to enter its upper counterpart only in a given order. Other particular legislations might be more permissive, but hardly any would be quite unhampered. Several legislations would be as strict as the German one we have just seen, but the strictness might be based on other demands for lawfulness. Some models of our theme in question might not allow for the nucleus allotment on the lower syntactic level. Latin wouldn't, for example. Neither would Russian, nor Eskimo. The latter is practically devoid of nuclei. This is a feature which character­ izes a lot of languages. Nucleus constructions are grammatical mediators between the nomogenetic Une (which dominates, as a grammatical rafter, the lower and the upper semantic levels) on one hand and the upper syntactic level on the other. This mediator may or may not be utilized by human language. Often languages with a highly complicated grammar do not profit from this syntactic possibility. We can now allow elements on the lower syntactic level to pass on to the upper syntactic level (if we consider it from the upper syntactic level), or we can transfer elements from the lower syntactic level to the upper syntactic level (if we look upon it from the lower syntactic level). We should always allow or transfer our genetic elements in accordance with the consistency of a given language, i.e. adjust its lowest levels to its uppermost level as two images of the same thing. Now we may change our sentence a little: we may place the nucleus hast (the finite verb, or the endonomic part of the sentence) between the O1 (was für eine Dummheit) and the S (du): Was für eine Dummheit hast du da begangen? If we do this we alter the semasiosyntactic theme completely. This is a beautiful proof of the absolute coherence between semantics and syntax. Our permutation has

202

PAPERS IN THEORETICAL LINGUISTICS

turned our sentence into an interrogative one with a completely different communicative value. We know of modern language schools which call this a transformation. What we would like to know, though, is whether A is a transformation of B, or vice versa. As long as this question cannot be answered, the word transformation is as meaning­ less as it is uninteresting. By introducing into our model a limited set of significant abbreviations for the basic functions of simple and extended transversal relations, we can now analyse any human sentence in an exhaustive, meaningful, and reasonable way. (All representa­ tions are thereby introduced on the upper syntactic level). Thus, what we owe the reader who has followed us truly so far, are some informative analysis models. First of all, let us look at some transversal relations containing a diastasis, a metastasis, or/and an apostasis. As English is very poor in cases we shall choose our examples from more revealing languages. I.

Diastasis: Ich bot ihm ein gutes Gehalt. ich -o- ihm ein -es

upper syntactic level

lower syntactic level

upper semantic level

lower semantic level positive NE

KINETIC ANALYSIS OF SENTENCES

203

The corresponding French theme is subjected to another legislation : Je lui offris un bon salaire. So if from the very start you choose the same order of basic elements on the French upper semantic level as we just did in our German example, you will have to take the trouble with the syntactic authorities at the nomothetic line. And vice versa; if by chance you had chosen the same order of basic elements on the German upper semantic level as the one represented immediately by the French upper syntactic level, you would have had to adjust the syntactic sequence of these elements to a particular German legislation. No syntactic legislation is in any way primary to any other, much less more distinguished or more preferable. This is a thing which many so-called modernists seem unable to realize. Both our German and our French sentences have a marker of indefiniteness in front of the O1. Articles (indefinite/definite) are introduced in some languages on the upper syntactic level. They are greetings of grammaticality from below the nomothetic line. They are universal entities, but not all languages express them. 1 They ought to be indicated already on the lower semantic level. Thousands of interesting facts could be said about this simple sentence, but this is not the place. This is supposed to be a book, and a book must come to an end. Thus we shall go on with our basic description of the genesis of human sentences. Let us consider the German sentence : Er antwortete mir. er -te mir upper syntactic level

lower syntactic level S ~ fV ~ 0 2

upper semantic level

1 Cf. Niels Danielsen: Bestimmtheit: Semantische Blätter, Heidelberg).

Unbestimmtheit.

Kontrastive

Beobachtungen (forthcoming in:

204

PAPERS IN THEORETICAL LINGUISTICS

lower semantic level positive NE [ = π]

This model shows how  π is internalized into a IT on the upper semantic level. All internalizations of n into  belong here: they are prior to the failing O1 on the lower syntactic level, and they are posterior to the kinetic (or lower) semantic level. All that we have said about genuine diastasis constructions (without or with an internalized π) is valid, too, for metastasis and apostasis constructions, only that you introduce the notations [ M ] ( < [ . . 0 ] ) and respectively, instead of the diastasis notation [A] on the upper semantic level: Man überführte ihn des Betruges. -te ihn des -es man upper syntactic level

lower syntactic level S ~ fV ~ O1 ~ Ω

upper semantic level [A] ~ [ Π ] ~ [O] ~ [M]

KINETIC ANALYSIS OF SENTENCES

205

lower semantic level positive NE

The element man in this sentence is expressive of grammaticality inasfar as it is indicating in a decisive way the stasis of the S-relation. The element des is abundant in grammaticality: it indicates definiteness, singular, not-feminine, and metastasis. And now for a metastasis construction with an internalized π: Der Umfang der Verwüstungen spottede jeder Beschreibung. -en -te -er der der upper syntactic level

lower syntactic level

upper semantic level

lower semantic level positive NE

206

PAPERS IN THEORETICAL LINGUISTICS

The relation of the two elements of the complex stasis is an inherence. All gram­ matical entities of contingence and integration, even pro-destination, must be indicated on the lower semantic level: they are prior to their manifestations on the upper semantic level. (The notation IE (individual identifier) : (I)P signalizes that an individ­ ual identifier like every, German jeder, French chaque, and so on, plays the role of an IP (a universal identifier) such as all, French tout, German all. The two categories IE and IP show an outspoken tendency to overlap in semantic value, cf. Heute ist nicht alle Tage : today is not every day). Up to now we have only considered plain and peaceful S-relations. Let us now take a look at a Q-relation with an apostasis, to kill two birds with one stone: Bello ūtendum est nobis. -5 -ndum -st -bis upper syntactic level

lower syntactic level

upper semantic level

lower semantic level

( = brug in the corresponding Danish construction: der bør gores brug af rig). This Latin instance is an interesting example of how an expression of modality is being materialized through a genetic model with an A -nucleus and a specific operative verb. The epistasis 2 is an example of O 2 :Ω. And bellō is the 'ablative object' (the ablative being the only case of transaction in Latin). That bello virtually

KINETIC ANALYSIS OF SENTENCES

207

is an ablative in this sentence is evident for two reasons: 1) the syntactic structure characterizing a Q-relation points decisively to bellō as the ectostatic implicatum governed directly by the verb ; 2) a substitution of bellō by a feminine like terrā will unveil the ingenious interaction between the categories gender and case: syncretisms of one case paradigm will turn out to be independent categories in a corresponding paradigm of a different gender. It is only through this interaction that the case markers of final points in a transversal relation achieve their unequivocal value as indicators of given roles in given syntactic situations. Such are, at any rate, the conditions in the grammatical system of Latin. Let us now take a look at different syntactic situations which seem to be of cardinal importance in a semasiosyntactic discussion. First of all, let us consider some simple instances of alloyings and elisions. Such sentences have never before been analysed thoroughly, and their similar appearances, at first glance, have lead to many misconceptions in traditional grammar. In other words: what is an unexpressed subject (or object)? (In some languages you will hear about implied subjects, or objects, although traditional grammar for some mystical reason rarely cares for the objects). On the first page of the Latin grammars we used at school we find alloyages They used to startle us in the beginning, and it took us some time to get acquainted with them: Amö patriam. -ō -am upper syntactic level fV 

lower syntactic level

upper semantic level

lower semantic level positive NE

208

PAPERS IN THEORETICAL LINGUISTICS

Our model shows us how the is becoming alloyed to the  on the upper semantic level. A physical sign of this process is the ending -ö on the upper syntactic level. If the stasis is reintroduced on the lower syntactic level, you get the sentence : ego amō patria. This construction may be used if you want to stress the subject. As an example of the opposite phenomenon, that the object is alloyed on the upper semantic level, we can take the English sentence I know. I

= Én tudom. (Hungarian) én -om

upper syntactic level S

fV

lower syntactic level

upper semantic level

lower semantic level positive NE

This model shows how an [O] is alloyed to the  on the upper semantic level. This alloying leaves no trace in the English syntagm, but it does in the Hungarian one: the ending - (of the so-called definite conjugation) indicates a implied O1. And now we shall see an example of how both things happen simultaneously:

KINETIC ANALYSIS OF SENTENCES

209

lower syntactic level

upper semantic level

lower semantic level positive NE

Here both " I " (subject) and "it" (object) have become alloyed to the 77 on the upper semantic level. Instead of speaking of subjects and objects on this level, though, you ought to stick to the strict terminology of stasis and epistasis. So, what actually happens in this case is that the stasis and the epistasis of the sentence are alloyed to the Π on the upper semantic level. Let us now see what happens if you reintroduce them on the lower syntactic level, and let us thereby use the illustrative Hungarian examples szeretlek (= "I love you") and en szeretlek téged (= "I love you"). Szeretlek. -lek.

Én szeretlek téged. -én -lek téged

upper syntactic level fV

lower syntactic level

upper semantic level

S

fV

O

210

PAPERS IN THEORETICAL LINGUISTICS

lower semantic level positive NE

In the left row the [A] and the [O] are alloyed to the  on the upper semantic level. No stasis or epistasis is reintroduced on the lower syntactic level. The result is a synthetic transversal relation (cf. sciō above). In the right row the stasis is reintroduced on the lower syntactic level, and so is the epistasis. You now get a sentence with the alloyment szeretlek (with the ending -lek indicating " I " (stasis) and "you" (epistasis) in one) and the stressed subject én (= " I " ) and the stressed object téged (= "you"). As opposed to the case of an alloying the loss of the materialized stasis in an elision takes place on the nomothetic line : Weiss nicht. -eiss upper syntactic level fV

lower syntactic level S ~ fV

upper semantic level [A]~ [Π]<[0]

KINETIC ANALYSIS OF SENTENCES

211

lower semantic level negative NE

This is the first time we see a model of a recusative sentence. The notation r of the nomothetic line indicates that the sentence is expressive of 'not-ness'. - The epistasis of the sentence is subject to an alloying on the upper semantic level. In this respect our example differs in no way from the English I know and the Latin sciō above. But the subject procedes undisturbed to the lower syntactic level: it does not leave any explicit traces of an alloying (weiss might as well be the 3rd person singular), and stasis alloyings are, on the whole, unsymptomatic of German sentence genesis. On the upper syntactic level, then, the stasis is elided. We have just been looking at a sentence which was negative/recusative. A sentence which is negative (in meaning) is not always recusative (containing the word 'not'), and vice versa: a propositive syntagm need not always be positive. Let us look at a simple model for recusativity/positivity (or a recusative diaxy): Wie entzückend er nicht singt. upper syntactic level

-t fV

er S

adv

Adv

adv

Adv ~ S ~

lower syntactic level r

fV

upper semantic level [A adv]

[Adv]] ~ [A] ~ [] ~ []

212 lower semantic level

PAPERS IN THEORETICAL LINGUISTICS positive Constitutional

[=π]

In some sentences the word 'not' may or may not be there without any impact on the semantic value of the sentence (cf. German : ich komme nicht, bevor du {nicht) antwortest). In other cases the word 'not' must always be there in certain constructions with a positive semantic value (cf. French:je crains qu'il ne vienne). Let us take a look at a sentence without the word 'not' but with a negative meaning (a so-called propositive diaxy): He knows nobody. upper syntactic level he -s

lower syntactic level

upper semantic level

negative NE

KINETIC ANALYSIS OF SENTENCES

213

Our English sentence shows abundancy in respect to its grammatical arrange­ ment: The notion 3rd person singular is expressed twofold: 1) in the subject he, and 2) in the verb morpheme -s. The element 'he' is there to express 'masculine'. (A lot of languages, however, do not distinguish 'he' and 'she'. Some of them do not even distinguish the difference between 'he' and 'she' and 'it'.) This is an example of how we can use our model to show what is abundant in a syntagm and in which sense it is abundant. We have now reached a point where we simply have to set a limit to our in­ vestigations on the semasiosyntactic arrangement of human language, if we do not want to get lost in an endless series of analysis instances. But before we bring our considerations on this captivating topic to a culmination, we shall have to consider three other cardinal issues in the field of semasiosyntactic analysis: 1) representations, 2) the role of the adjective, and 3) the subordinated clause (i.e. the problems in con­ nection with the joining elements of finite superordinates and their finite subordinates). Representations are of three sorts: 1) stasis or epistasis representations, 2) re­ presentations of transactions, and 3) expanding transaction representations. Let us consider a stasis or epistasis representation: Watakushi wa suri ni tokei o suraremashita. (= "I have been robbed of my -aremashita watch by a pick-pocket") (Japanese) upper syntactic level repr repr repr fV nomothetic Une p ßZ lower syntactic level S ~ Ω ~ O1 ~ fV upper semantic level

lower semantic level positive NE

214

PAPERS IN THEORETICAL LINGUISTICS

The transversal relation above is an RQ-relation. The stasis of the R-relation is being represented by the indicator wa. The scopos of the Q-relation is being represented by the indicator o. And the epistasis 2 of the Q-relation is being represent­ ed by the indicator ni (corresponding to the English indicator 'by'). Arabic may use a marvellous stasis representation: 'Inna fattasū dhalika 1-jiwāra. (= "the men searched that neighbourhood") repr: S fV O1 The stasis representation is 'inna r-rijāla ('inna is the indicator and governs the accusative!). In the Spanish sentence yo conozco al señor Martínez you have an O 1 -represen­ tation (a/ señor Martínez). The German sentence er wird von allen bewundert there is an Ω-representation (von allen). On the existential level some languages have representations for the existential epistasis (= the predicative): Y mae dyn doeth yn onest. (= "a wise man is honest")

(Welsh)

yn onest is the predicative representation of the adjective gonest (= "honest" (notice the leniating mutation of gonest > onest after the indicator of predicativity yn)). Indications of transaction are indications of transactional cases (cf. parasta on olla hiiskumatta sanaakaan koko asiasta (Finnish): am besten ist es, kein Wort von der ganzen Angelegenheit verlauten zu lassen (with a representation in the German sentence)), and expanding transaction representations are representations for convergent cases (genitive or dative): Er gab dem Jungen ein Buch: il donna un livre au garçon; Er erinnerte sich seiner Mutter: il se souvenait de sa mère. All this can be easily in­ dicated in a genetic sentence model. The genetic model of a predicative sentence looks like this: Der frühere Weltmeister war .-. der -e war upper syntactic level

lower syntactic level

upper semantic level

215

KINETIC ANALYSIS OF SENTENCES lower semantic level positive NE

(In this model ' :)' indicates an antecedent attributive adjective; equally, ' ( : ' indi­ cates a postcedent attributive adjective. On the lower semantic level we cannot say anything of universal value about the syntactic place of a given attributive adjective meaning. So here we must be content with the universal notation ' : ) ( : ' = "attributive adjective meaning", (:) indicates predicative meaning. P stands for predicative, and O e for existential O). It is symptomatic of the adjective that a given adj meaning selects either a predicative or an attributive place in the sentence, or both. On the upper syntactic level it may have different materializations according to its syntactic place or function (cf. Italian buon (antecedent): buono (postcedent); German: die guten Männer: Die Männer sind gut, Welsh: dyn gonest (= "an honest m a n " ) : Y mae'r dyn yn onest (= "the man is honest"), and so on). Some languages are extremely poor in adjectives (cf. p. 102). Some have to attribute adjectival meanings in the form of participles of (:)-adjectives, for example Eskimo : Angut silatôq ilumôrpoq. (= "a wise man is honest") man having honest-is much spirit We shall now have to consider a more complex syntagm consisting of more finite segments : The man whom you told that she was going to leave was John's uncle. upper syntactic level

S

[77] Ol

S

[]

02 NS

S

[]

(NS stands for neutral/subordinated status). After the segmentation of the complex syntagm you go on describing every single element all the way down to the kinetic limit. On the lower semantic level you will have to indicate that we have got two different sorts of subordinate clauses in front of us: 1) a subordinate constitutional clause signalized by the constitutional element whom (an IB (relative identifier) of the personal zone), and a subordinate neutral 10

216

PAPERS IN THEORETICAL LINGUISTICS

status signalized by the status marker that)1. Subordinate clauses are hierarchized on the lower semantic level as specific units and analysed in detail on the upper semantic level. Differently materialized syntagms may follow identical genetic rules. And vice versa : identical semantic information may follow different genetic models. For instance, the English sentence I have a headache is a nice transversal S-relation with an S, an fV, and an O1 Its Eskimo counterpart is niarqordlugpoq which is an obtuse trans­ versal relation and has to be genetically depicted as such. No less than two important things, though, connect our two sentences, the English one, and the Eskimo one: 1) they have got exactly the same communicative value, and 2) they both have a critical field of distribution.1 This is the highest syntactic universal without which no sentence order would be imaginable. 1 Cf. Niels Danielsen: Semasiosyntaktische Universalien im Finnischen, in: Ural-Altaische Jahrbücher 66, 1974; Niels Danielsen: Status und Polarität im Gotischen, im Lichte des Kymrischen dargestellt, Odense University Studies in Linguistics, vol. 2, Odense 1968; Niels Danielsen: Die Relativa im Neuhochdeutschen — und anderswo. Eine Vorstudie zu einer konstitutionellen Sprachtheorie, in: Spràkliga Bidrag vol. 6, n : r 27, Lund 1972, sowie in: Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung, Heft 28/1 '75, Berlin 1975; Niels Danielsen: Zum Wesen des Konditionalsatzes — nicht zuletzt im Indoeuropäischen, Odense University Studies in Linguistics, vol. 1, Odense 1968; Niels Danielsen: Zur Universalität der Sprache, in: Sprachwissenschaft, Vol. 1, Heidelberg 1976.

A CONCISE LIST OF THE MOST IMPORTANT SYMBOLS AND ABBREVIATIONS USED IN A NOMIC ANALYSIS + = positive (semantically/logically); ÷ = negative (semantically/logically) p = propositive; r = recusative + /q 1) He came are the two alternatives of a sunaxu. ÷/r 2) He didn't come ' are the two alternatives of a diaxy. + /r 2) Ain't s h e sweet! S = 1 ) subject 2) stasis  = t h e sentence verb ( t h e mediator of a transversal relation) O1 = epistasis ("accusative object") 2 = diastasis ( " d a t i v e object") Û = metastasis ("genitive object") Ω1 = partitive i n t h e stasis field of t h e m e s o n o m i c p a r t of the sentence Ω2 = partitive in the ectonomic part of the sentence 0 3 = apostasis (any other object in a transversal relation) = 1) the epistasis in an R-relation 2) the epistasis 2 in a Q-relation Y = hyperstasi s fV = finite verb (e.g. he has come) oV = operative verb (e.g. he has come) vP = verbal protone (= the fV of a critical field of distribution)  = terminal nucleus (e.g. you shouldn't have done that) 0 2 = nucleus of second order (e.g. you shouldn't have come) 0 3 = nucleus of third order (e.g. you shouldn't have come) 0 = unmaterialized nucleus [O] = the basis of transversal relation (the terminal of content) [A] = the stasis of an S-relation, the epistasis of an R-relation (the principle of content) Oe = existential epistasis (e.g. he is Peter) O P = epistasis of progress (e.g. he grew old) Or = remanential epistasis (e.g. he remained quiet) = member of the constitutional axis (WHO?, WHAT?, WHICH?, WHEN?, HOW?, etc.) i = indicator (on, at, with, before, as, than, etc.) ind = 1 ) indication (on the table, before that time, underneath the carpet, as a whole, than me, etc.), 2) indicatum repr = representation (quiero a Antonio = I love Antonio) repr : O1 [. O] = the semantic basis of a diastasis [A] [. .0] = the semantic basis of a metastasis [M] [. . . O] = he semantic basis of an apostasis TR = transversal relation

INDEX OF LANGUAGES This index lists the languages mentioned in the articles of this volume. Mentions in (foot)notes are indicated by an n added to the page, without the actual footnote number. A. Agul: 23 Albanian: xvii, xxi, 62, 63, 135137, 147, 152, 158 Arabic: xvi, 16, 72, 100, 102, 103, 118, 124, 125, 130, 139, 148, 152, 162, 163, 175, 176, 214. Quaday-Arabic: 15, 17, 56, 82, 145 Armenian: xvii, xxi, 62, 64, 135, 139 Avestan: xvii, xxi B. Basque: 22, 54, 63, 64, 69, 70, 110, 111, 176, 177, 198 Bats: 23 Breton: 71, 80, 81, 103, 104, 172 Bulgarian: 65, 137, 153, 154. Old Bulgarian: xvii  Celtic: 6, 168 Czech: 63, 65, 72, 103, 134, 155 Chinese: xvii, xviii, 27, 51, 52, 64, 158, 164-166 D. Danish: 7, 14, 18, 19, 22, 41, 43, 51, 52, 59, 62, 63, 65, 68, 75, 79, 83, 93, 94, 99, 101, 103, 114, 119, 120, 131, 133,

147, 156, 158, 198 Dutch: 18, 65, 151, 158 E. Early Middle Welsh see Welsh English: 7, 8, 12-15, 18-20, 22, 27, 28, 41, 49, 51, 52, 59, 60, 62, 63, 65, 68, 71, 73, 75, 79, 81, 82, 84, 88, 90, 92, 99-101, 103, 111, 112, 114, 118, 119, 124, 131, 147, 158-160, 168, 172, 193, 195, 199, 208, 209, 211-213, 215, 216. Pidgin Eng­ lish: 198 Eskimo: 10, 31, 56, 100, 103, 123, 187, 188, 195n, 201, 215, 216 Estonian: xxi, 110, 150, 155, 156, 167, 168 Finnish: xvii, 11, 19, 24, 31, 32, 94, 95, 110, 113, 149, 157, 159, 162, 192, 214 French: vii, xvi, xviii, 7, 12-14, 16, 18, 19, 27, 29, 33, 54-60, 62, 65, 67, 79, 82-84, 88, 9193, 99-101, 103, 118, 132-134, 148, 156, 158, 186, 203, 206, 212, 214. Old French: 18 Frisian: xvi, xx. Old Frisian: xvii

220

INDEX OF LANGUAGES

G. Gaelic: xxi, 71, 103. ScottishGaelic: 23, 73, 81, 100, 101, 103, 110, 111, 125 Georgian: 54, 55, 71, 110, 122, 123, 173, 174, 198 German: vii, xvi, xviii, 6-138 passim, 151, 156-160, 164, 165, 172, 189, 192, 195n, 199206, 210-212, 214, 215. Low German xvi. Middle High German: xvii. Old High Ger­ man: xvii, 60, 107-109 Gothic: xvii, xviii, 14, 55, 69, 70, 71, 88, 166 Greek: xvi, xvii, 18, 22, 33, 43, 55, 65, 69, 70, 99, 100, 103, 104, 106, 118-121, 149, 158. Modern Greek: 62, 63, 65, 152 Greenlandic: xxi, 54, 78, 113, 115, 123, 145 H. Hausa: 99 Hebrew: 100, 102 Hindi: 23 Hungarian: xxi, 31, 42n, 54, 55, 57, 62-64, 70-72, 75, 84, 91, 99, 101-103, 133, 135, 136, 157, 163, 189, 197n, 208-210

Old Irish: xvii Ishkashmi: 56, 57, 87, 145 Italian: xvi, 9, 18, 19, 65, 88, 90-93, 99, 102-104, 132, 133, 148, 215 J. Japanese: 10, 25, 54, 84, 101, 119, 128-130, 163, 164, 177181, 189-191, 213, 214 K. Kannada (Kanarese): 7, 50n, 62, 112 Kashubian: 134 Khalkha-Mongolian see Mongo­ lian L. Latin: vii, xvi, xvii, 15, 22, 29, 33, 41, 50n, 52, 54-58, 65, 6971, 75, 99-101, 103, 117, 128, 148, 192, 198, 201, 206-208, 211 Latvian: 125, 137, 147-150, 154, 156, 159 Lezghian: xxi Lithuanian: xvii, xx, 63, 72, 101, 113, 121-123, 149, 150, 154, 155, 159, 161 Low German see German

I. Ibo 101 Icelandic: xix, 19, 23, 26, 31, 90, 99, 101-103, 106, 107, 110, 125-129, 132, 137, 138, 146, 149, 155, 156, 158, 192, 195n Indo-European: xvii, 88 Iranian: xvi Irish: 28, 31, 71, 72, 81, 102, 107-109, 125, 149, 160, 161.

M. Macedonian: 137 Malagasy: xxi, 102 Malayalam: 63 Manx: 71, 72, 99, 125 Menomini: xix, 56, 57, 69, 145 Middle High German see Ger­ man Middle Welsh see Welsh

INDEX OF LANGUAGES

Modern Greek see Greek Mongolian: 14, 19, 49, 56, 57, 91, 145. Khalkha-Mongolian: 88, 92 N. Narranyeri: xxi, North Frisian: 163 North-Samojede: xxi O. Old Old Old Old Old Old Old

Bulgarian see Bulgarian French see French Frisian see Frisian High German see German Irish see Irish Norse: xvii Saxon xvii

P. Persian: 114, 156 Pidgin English see English Polish: xv, 16, 24, 39, 49, 62, 65, 74, 85, 99, 100, 103, 110, 134, 149, 153, 158, 186, 187, 192 Portuguese: 29, 65, 99, 104, 133, 157 Q. Quaday-Arabic see Arabic Quechea: 199 R. Rhaeto-Romance: 65, 174 Rumanian: 65, 130, 136, 137, 150, 152, 192 Russian: xvi, xviii, 14, 16, 18, 29, 39, 54, 57, 59, 62-66, 70, 72, 74, 78, 80, 85, 90, 99-103, 112, 118, 134, 147-150, 153,

221

155, 158, 182-187, 192, 195n, 201 S. Sanskrit: xvii, 9, 15, 16, 18, 27, 31, 102, 154 Scottish-Gaelic see Gaelic Serbo-Croatian: 9, 65, 83, 136, 151, 152, 158, 159, 192 Singhalese: 147, 161, 162 Spanish: xvi, 18, 19, 63, 65, 68, 80, 82, 83, 89, 101, 103, 104, 130, 132, 133, 158, 160, 214 Subiya: 56, 62, 145 Swahili: 9, 10, 19 Swedish: 43, 83, 84, 99, 101, 155, 156 T. Tagalog: 92n Telugu: 7, 10 Turkish: xvi, xviii, 31, 50-52, 59, 65, 83, 85, 99, 103, 114, 199 U. Ukrainian: xxi, 103, 110 Urdu: xxi, 23, 29, 90, 91, 100, 110, 114, 149, 158, V. Vietnamese: 22 W. Welsh: xviii, 7, 10, 16, 50n, 51, 52, 58, 59, 62, 63, 66, 71, 84, 101, 102, 125, 166, 174, 192, 214, 215. Early Middle Welsh: 37, 54, 56, 57, 64. Middle Welsh: 145 Z. Zulu: 56, 63, 145

INDEX OF AUTHORS This index lists the names of the scholars mentioned in the articles of this volume. Mentions in the respective bibliographies are indicated by a postponed b. Mentions in (foot)notes are indicated by an n added to the page, without the actual footnote number. A. Anderson, J.M.: 26, 27, 35b, 110, 140b Anderson, O.B.: xix, xxi, 11, 35b, 164, 169b Aristotle: 24, 86, 105, 142, 143, 144

E. Einstein, A.: 2 Eldjárn, Kr.: xix

B. Barr, : xvi Bechert, J.: 61n, 173n Berdyayev, N.: 134, 140b Bhadkamkar, H.M.: 142 Bloomfield, L.: 57, 69n Bowen, J.T.: 7n, 35b Bruno, G.: 4

G. Galileo: 2 Gardner, Th.: 3, 35b Giuliani, M.V.: In, 35b Gr0nbech, K: xvi

 Carbou, H.: 56, 82n Chomsky, N.: 1, 2, 3, 11, 35b, 49, 52n, 53n, 73n, 105, 171, 171n, 172, 185, 193n, 194, 195n Clément, D.: 61n, 173n Copernicus: 2 Coseriu, E.: 73n D. Doke, ..: 56, 63n Drachmann, A.B.: 194n

F. Fillmore, Ch.: 26, 35b, 110, 140b Fowler, R.: 2, 35b

H. Hammerich, L.L.: xvi, xix Harris, Z.S.: 11 Hayek, F.A.: 193n Heger, : 110, 140b Heiberg, J.L.: 194n Helbig, G.: 110, 140b Hjelmslev, L.: xvi, 27, 131, 140b Hovdhaugen, E.: 2, 35b, 177n J. Jacottet, E.: 56, 62n Jakobson, R.: 110, 140b Jones, J.M.: 37n, 64n Jones, T.J.R.: 7n, 35b Jørgensen, P.: xvi

224

INDEX OF AUTHORS

 Kierkegaard, S.: 193, 194n Kromann, H.-P: 194n Kurylowicz, J.: 110, 140b L. Lange, H.O.: 194n N. Noss, R.B.: 53n P. Pānini: 2 Poppe, N.: 57, 88n, 92n Popper, KR.: 4n, 35b, 177n Pulgram, E.: 52n R. Rask, R.: xv, xviii, xix Rischel, J.: 187n Robinson, J.: 110, 140b

S. Saussure, F. de: 11 Schmidt, W.: 125, 140b Spore, P.: 15, 35b Staal, J.F.: 169b Stageberg, N.C.: 176n Steinthal, H.: 143 Strachan, J.: 37n, 64n Strawson, P.F.: 21, 35b T. Thümmel, W.: 61n, 173n Tricot, J.: 144 W. Wagner, K.-H.: 61n, 173n Y. Yāska: 142