Organizing Grammar: Linguistic Studies in Honor of Henk van Riemsdijk 9783110892994, 9783110188509

Citation preview

Organizing Grammar

W G DE

Studies in Generative Grammar 86

Editors

Jan Köster Harry van der Hulst Henk van Riemsdijk

Mouton de Gruyter Berlin · New York

Organizing Grammar Studies in Honor of Henk van Riemsdijk

Edited by

Hans Broekhuis Norbert Corver Riny Huybregts Ursula Kleinhenz Jan Köster

Mouton de Gruyter Berlin · New York

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter G m b H & Co. K G , Berlin.

The series Studies in Generative G r a m m a r was formerly published by Foris Publications Holland.

@ Printed on acid-free paper which falls within the guidelines of the A N S I to ensure permanence and durability.

Library of Congress Cataloging-in-Publication

Data

Organizing grammar : linguistic studies in h o n o r of Henk Van Riemsdijk / edited by Hans Broekhuis ... [et al.]. p. cm. - (Studies in generative grammar ; 86) Includes bibliographical references and index. ISBN-13: 978-3-11-018850-9 (cloth : alk. paper) ISBN-10: 3-11-018850-3 (cloth : alk. paper) 1. Grammar, Comparative and general. I. Riemsdijk, Henk C. van. II. Broekhuis, Hans, 1 9 5 9 III. Series. P201.068 2005 415-dc22 2005031175

ISBN-13: 978-3-11-018850-9 ISBN-10: 3-11-018850-3 Bibliographic

information published by Die Deutsche

Bibliothek

Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at < h t t p : / / d n b . d d b . d e > .

© Copyright 2005 by Walter de Gruyter GmbH & Co. KG, 10785 Berlin and by the contributors. All rights reserved, including those of translation into foreign languages. N o part of this book may be reproduced in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing f r o m the publisher. Cover design: Christopher Schneider, Berlin. Printing and binding: Hubert & Co. G m b H & Co. K G , Göttingen Printed in Germany.

Henk van Riemsdijk (Picture courtesy Ben Bergmans)

Table of Contents List of contributors

XIII

Hi Morris, this is Henk!

XVII

An intersubjective note on the notion of 'subjectification' Werner Abraham

1

A note on non-canonical passives: the case of the ^/-passive Artemis Alexiadou

13

Displaced and misplaced genitives Josef Bayer

22

Preposition stranding and locative adverbs in German Dorothee Beermann and Lars Hellan

31

Moving verbal complexes in Spanish Reineke Bok-Bennema

43

Unbearably light verbs versus finite auxiliary drop Anne Breitbarth

53

Extraction from subjects: some remarks on Chomsky's On phases Hans Broekhuis

59

A Chinese relative Lisa Lai-Shen Cheng and Rint Sybesma

69

Approximative of zo as a diagnostic tool Norbert Corver

77

A note on interpretable features and idiosyncratic categorial selection Denis Delfitto

87

VIII Contents Transparent, free... and polarised: the (poli)tics of polarity in transparent free relatives Marcel den Dikken

97

The inverse agreement constraint in Hungarian: a relic of a Uralic-Siberian Sprachbund? Katalin E. Kiss

108

Syntactic conditions on phonetically empty morphemes Joseph Emonds

116

Long-distance reciprocals Martin Everaert

127

The notion of topic and the problem of quantification in Hungarian Zsuzsanna Gecseg and Ferenc Kiefer

137

Questions of complexity Casper de Groot

146

Functional heads, lexical heads and hybrid categories Liliane Haegeman

152

Concatenation and interpretation Martin Haiden

162

As time goes by: a digressive discourse Hubert Haider, Masyuki Oishi and Shigeo Tonoike

171

There's that: unifying existential and list readings Jutta Μ. Hartmann

186

Extended projections - extended analogues: a note on Hungarian PPs Veronika Hegedüs Against the sonority scale: evidence from Frankish tones Ben Hermans and Marc van Oostendorp

197

206

Contents

Classifiers, agreement and honorifics in Japanese

IX

222

Masaru Honda

What stranded adjectives reveal about Split-NP Topicalization

230

Hanneke van Hoof

Past tense interpretations in Dutch

241

Angeliek van Hout

Why phonology is the same

252

Harry van der Hulst

Recursively linked Case-Agreement: from accidents to principles and beyond

263

Riny Huybregts

Enfoldment as Economy

275

Takashi Imai

"GP, I'll have to put your flat feet on the ground"

283

Jonathan Kaye

On parameters and on principles of pronunciation

289

Richard S. Kayne

What to do with those fools of a crew?

300

Evelien Keizer

Why indefinite pronouns are different

310

Istvän Kenesei

Seeing the forest despite the tree

319

Hans-Peter Kolb

When to pied-pipe and when to strand in San Dionicio Octotepec Zapotec

331

Hilda Koopman

Free relatives as light-headed relatives in Turkish Jaklin Kornfilt

340

X

Contents

Is linguistics a natural science? Jan Köster

350

Two asymmetries between Clitic Left and Clitic Right Dislocation in Bulgarian Iliyana Krapova and Guglielmo Cinque

359

On dative subjects in Russian S.-Y. Kuroda

365

On the nature of case in Basque: structural or inherent? Itziar Laka374

374

Examining the scope of Principles-and-Parameters David LeBlanc

383

Theory

Clitics and adjacency in Greek PPs Winfried Lechner and Elena Anagnostopoulou

390

A minimalist program for parametric linguistics? Giuseppe Longobardi

407

A syntactic approach to negated focus questions in Bulgarian Krzysztof Migdalski

415

The case of midpositions Natasa Milicevic

424

Quechua P-soup Pieter Muysken

434

Semantic compositionality of the way-construction Heizo Nakajima

439

Soft mutation at the interface Ad Neeleman

447

Abracadabra, the relation between stress and rhythm Anneke Neijt

458

Contents

What do we learn when we acquire a language? Marina Nespor, Judit Gervain and Jacques Mehler A prosodic contrast between Northern and Southern Dutch: a result of a Flemish-French sprachbund Roland Noske

XI

466

474

The object of verbs like help and an apparent violation of UTAH Christer Platzack

483

A note on relative pronouns in Standard German Martin Prinzhorn and Viola Schmitt

495

Agreeing to bind Eric Reuland

505

Positive polarity and evaluation Hendrik C. van Riemsdijk

514

Phase theory and the privilege of the root Luigi Rizzi

529

On the role of parameters in Universal Grammar: a reply to Newmeyer Ian Roberts and Anders Holmberg Welsh VP-ellipsis and the representation of aspect Alain Rouveret A new perspective on event participants in psychological states and events Β ozena Rozwadowska

538

554

563

A glimpse of doubly-filled COMPs in Swiss German Manuela Schönenberger

572

Missing prepositions in Dutch free relatives Chris Sijtsma

582

Final sonorant devoicing in early Yokuts field-records Norval Smith

592

XII

Contents

Cyclic NP structure and trace interpretation Dominique Sportiche

599

Appositive and parenthetical relative clauses Tim Stowell

608

Overt infinitival subjects (if that's what they are) Anna Szabolcsi

618

Wanna and the prepositional complementizers of English Tarald Taraldsen

625

A note on asymmetric coordination and subject gaps Craig Thiersch

633

The representation of focus and its implications: towards an alternative account of some 'intervention effects' Jean-Roger Vergnaud and Maria-Luisa Zubizarreta

641

Circumstantial evidence for Dative Shift Edwin Williams

661

Why should diminutives count? Martina Wiltschko

669

Adjacency, PF, and extraposition Susi Wurmbrand and Jonathan David Bobaljik

679

A note on functional adpositions Jan-Wouter Zwart

689

Bibliography of Henk C. van Riemsdijk

696

Index

707

List of contributors

Werner Abraham Universität Wien [email protected]

Guglielmo Cinque Universitä di Venezia [email protected]

Artemis Alexiadou University of Stuttgart [email protected]

Norbert Corver University of Utrecht [email protected]

Eena Anagnostopoulou University of Crete [email protected]

Denis Delfitto University of Verona denis.delfitto@ univr.it

Josef Bayer University of Konstanz [email protected]

Marcel den Dikken CUNY Graduate Center [email protected]

Dorothee Beermann NTNU, Trondheim [email protected]

Katalin E. Kiss Hungarian Academy of Sciences [email protected]

Jonathan David Bobaljik University of Connecticut, Storrs j onathan. bobalj [email protected]

Joseph Emonds Kobe-Shoin Women's University [email protected]

Reineke Bok-Bennema University of Groningen [email protected]

Martin Everaert Utrecht University [email protected]

Anne Breitbarth University of Tilburg [email protected]

Zsuzsanna Gecseg Szeged University [email protected]

Hans Broekhuis University of Tilburg [email protected]

Judit Gervain SISSA [email protected]

Lisa Lai-Shen Cheng Leiden University [email protected]

Casper de Groot University of Amsterdam [email protected]

XIV

Contributors

Liliane Haegeman UMR 8258 Silex du CNRS — Lille III liliane.haegeman@univ-lille3 .fr

Harry van der Hulst University of Connecticut [email protected]

Martin Haiden CNRS, Univ. Lille 3 [email protected]

Riny Huybregts University of Utrecht/ University of Leiden [email protected]

Hubert Haider Universität Salzburg [email protected]

Takashi Imai Tsuru University timaiOO 1 @mac.com

Jutta Μ. Hartmann Tilburg University j. hartmann@uvt. η 1

Jonathan Kaye Girona, Catalunya [email protected]

Veronika Hegedüs University of Tilburg [email protected]

Richard S. Kayne New York University [email protected]

Lars Hellan NTNU, Trondheim [email protected]

Evelien Keizer University of Amsterdam [email protected]

Ben Hermans Meertens Institute, Amsterdam [email protected]

Istvän Kenesei Research Institute for Linguistics, Budapest/University of Szeged [email protected]

Anders Holmberg University of Newcastle upon Tyne [email protected] Masaru Honda Kwansei Gakuin University [email protected] Hanneke van Hoof University of Tilburg [email protected] Angeliek van Hout University of Groningen [email protected]

Ferenc Kiefer Hungarian Academy of Sciences [email protected] Hans-Peter Kolb TNO, Delft, The Netherlands [email protected] Hilda Koopman UCLA [email protected] Jaklin Kornfilt Syracuse University [email protected]

Contributors

Jan Koster University of Groningen [email protected]

Pieter Muysken Radboud University [email protected]

Iliyana Krapova Universitä di Venezia [email protected]

Heizo Nakajima Gakushuin University [email protected]

S.-Y. Kuroda University of California, Diego/International Institute Advanced Studies, Kyoto. [email protected]

San for

Itziar Laka University of the Basque Country EHU/UPV [email protected] David LeBlanc University of Prince Edward Island [email protected] Winfried Lechner University of Stuttgart [email protected] Giuseppe Longobardi Universitä di Trieste longbard@units. it Jacques Mehler SISSA [email protected] Krzysztof Migdalski University of Tilburg [email protected] Natasa Milicevic University of Tilburg [email protected]

XV

Ad Neeleman UCL [email protected] Anneke Neijt Radboud University, Nijmegen [email protected] Marina Nespor University of Ferrara [email protected] Roland Noske Universite Lille 3/CNRS 8528 Silex roland.noske@univ-lille3 .fr

UMR

Masyuki Oishi Tohoku Gakuin University [email protected] Marc van Oostendorp Meertens Institute, Amsterdam [email protected] aw.nl Christer Platzack Lund University [email protected] Martin Prinzhorn University of Vienna martin. prinzhorn@un iv ie. ac. at Eric Reuland University of Utrecht [email protected]

XVI

Contributors

Hendrik C. van Riemsdijk Freelance Linguist [email protected]

Anna Szabolcsi New York University [email protected]

Luigi Rizzi University of Siena [email protected]

Tarald Taraldsen CASTL Tromso [email protected]

Ian Roberts University of Cambridge [email protected]

Craig Thiersch University of Tilburg/University of Potsdam c. 1 .thiersch@uvt. η 1

Alain Rouveret Universite Paris-7 arobur@club- internet. fir Bozena Rozwadowska University of Wroclaw [email protected] Manuela Schönenberger University of Stuttgart [email protected] Chris Sijtsma [email protected] Viola Schmitt University of Vienna [email protected] Nerval Smith University of Amsterdam [email protected] Dominique Sportiche UCLA/ENS Paris [email protected] Tim Stowell UCLA [email protected] Rint Sybesma Leiden University [email protected]

Shigeo Tonoike Aoyama Gakuin University [email protected] Jean-Roger Vergnaud U.S.C vergnaud@usc. edu Edwin Williams Princeton University [email protected] Martina Wiltschko UBC [email protected] Susi Wurmbrand University of Connecticut, Storrs [email protected] Maria-Luisa Zubizarreta U.S.C zub izarr@usc. edu Jan-Wouter Zwart University of Groningen [email protected]

689

Hi Morris, this is Henk!

According to an apocryphal story, in 1970, a group of students at the University of Amsterdam was discussing a problem related to phonology. They could not solve the problem, so, one of the students, Henk van Riemsdijk, said: "OK, let me call Morris Halle at MIT". So, Henk grabbed the nearest telephone, dialed a number, followed by a conversation that started as follows: "Hi Morris, this is Henk! We've got a problem here and we thought you might know the answer, etc., etc." We have not been able to ascertain the authenticity of the story, but we all agree that it could have been true and that it is typical of a kind of flair that did not exist in Dutch linguistics before Henk's arrival in 1970. Just imagine, Morris and MIT were world-famous institutions and quite intimidating to a group of relatively provincial and diffident students. Henk was only 22, already a complete linguist and talking to the big shots in the field as if he were their equal and daily pal. It is this exceptional flair and initiative that would become the basis of Henk's decisive role in organizing the field of generative linguistics in Europe. Born in 1948 in The Hague, The Netherlands, Henk's Dutch family soon moved to Zurich, Switzerland, where he grew up. That was a lucky accident because it would make Henk the internationally-minded polyglot that helped him so much to play a European role in his later life. In Zurich, he spoke Dutch within the family, Swiss German in the streets and High German in school. Moreover, since not only German but also French and Italian are official languages of Switzerland, he learned these languages better than most of us. In fact, he became absolutely fluent in French when he became a student of linguistics in the revolutionary Paris of the late 1960s. Going to Paris at that time was another stroke of good luck, because said revolutionary climate had created the experimental curriculum in Vincennes, which included much interest in generative grammar. This attracted various MIT linguists, next to local pioneers of the field, like Nicolas Ruwet and the psycholinguist Jacques Mehler. The MIT linguists in Paris included Richard Kayne, Yuki Kuroda and Joseph Emonds.

XVIII

Preface

Especially Joe Emonds would become a lifelong friend of Henk's and a major influence on early Dutch generative syntax. When Henk arrived at the University of Amsterdam in 1970, not only was he a mature linguist but he also brought along the international network he had built up in his Paris years. With his social and intellectual talents, his initiative and, above all, with his international background, Henk turned the aforementioned somewhat diffident and provincial group of Amsterdam students into a force that would make Amsterdam the leading European center of generative linguistics for many years to come. Luckily, the older generation in Holland, with figures like Frits Staal, Albert Kraak, Pieter Seuren, Wim Klooster and Henk Verkuyl, had already created a favorable climate for generative linguistics, with various connections to MIT and other international centers. The typical orientation of generative linguistics in Amsterdam in those years was generative semantics and that created a short-lived consensus for a while across generations. It even included the head of the Amsterdam linguistics department, the late Simon Dik, who had serious misgivings about MIT linguistics but who -it must be said to his credit- was always supportive of the initiatives of Henk and his fellow students, especially in the early years. After a while, the consensus about generative semantics broke down because the students of Henk's generation became more and more skeptical about the ever wilder extensions of grammar, especially in the hands of George Lakoff. They started looking for something new and, lo and behold, in flew Joe Emonds in 1972, on Henk's invitation and precisely at the right time. Curiously, thanks to a lecture given by Joe in Utrecht, the Amsterdam students had their first real contacts with the like-minded Utrecht linguists Arnold Evers and Riny Huijbregts. Joe was a representative of the then new EST model, with a much more down-to-earth, surface-oriented kind of syntax, with an emphasis on solid argumentation. The minor paradigm shift was already initiated by Chomsky's visit to Amsterdam the year before, when he presented a version of his article Conditions on Transformations. The new orientation of Henk and his fellow students created some minor tensions with the older generation, who stuck to the generative semantics framework, but on the whole things remained friendly and there never was anything like 'the linguistic wars' that allegedly were waged in the United States. Occasionally, the breaking down of the consensus was blamed on Henk's pernicious influence, but it really was a collective process and Henk was not even the first to make the transition to the EST model. Apart from

Preface

XIX

considerations of substance, this transition also was a development defining the identity of a new generation, for which Chomsky's Conditions of Transformations would become the leading program. This development was consolidated by Henk's sabbatical at MIT in 1974 (together with Frans Zwarts), which was preceded by an influential anti-generative semantics seminar given by Mike Brame in Utrecht. The MIT visit also led to a considerable expansion of Henk's already impressive network, with new friends, like Mary-Louise Kean, Edwin Williams and many others. The cooperation with Edwin would lead to a popular textbook that was translated into several languages. The association with MIT became rather close at the time, especially when Henk's friends and fellow students Hans den Besten and Jan Koster also visited MIT in 1976. By that time, we had already seen the birth of what would become GLOW. In the spring of 1976, there was a student reading and discussion club led by Henk under the name Green Ideas. Henk took the initiative for a concluding conference in Amsterdam called Green Ideas Blown Up, which was an unforgettable experience. The Amsterdam linguists discovered that they were not alone in their enthusiasm for Conditions on Transformations and were delighted to share ideas and impressions with Jacqueline Gueron, Jean-Yves Pollock, Luigi Rizzi, Alain Rouveret, and several others. GLOW, in other words, really goes back to this AmsterdamParis association, all of it started by Henk, who had had the good fortune to have lived in both cities. Henk's friend Jean-Roger Vergnaud could not come in 1976, but the next year he was one of the three founders of GLOW, together with Henk and Jan Köster, but clearly, as always, on Henk's initiative. The GLOW manifesto - i n retrospect a document of touching orthodoxy- was written by Jan Köster and corrected and expanded by Henk and Jean-Roger. The rest is history. Henk chaired GLOW for many years and made it the leading platform of European generative linguistics, which it has remained until the present day. Last but not least, Henk has been co-editor of the renowned book series Studies in Generative Grammar since its inception (1978) and also was co-founder of The Linguistic Review in 1981. Apart from his leading role in organizing our field, Henk also played a major role in building the quality of Dutch linguistics. His dissertation of 1978 about PPs is a classical text on the topic. Henk made major contributions to X-bar theory, theories of sentence Pied Piping, and particularly to non-tree-representable structures, as, according to Henk, found in coordination, parasitic gaps, parentheticals, and many other

XX

Preface

constructions. He also has been influential as a teacher and thesis supervisor, first in Amsterdam and since 1981 in Tilburg. He supervised many dissertations, including those of Hilda Koopman, Reineke BokBennema, Hans den Besten, Hans Bennis, Norbert Corver, Marco Haverkort, Angeliek van Hout, Paola Monachesi, Elena Anagnostopoulou and Marc van Oostendorp. Next to his major achievement of founding GLOW, Henk made several other important contributions to the infra-structure of our field in Europe. The major European summer schools, like the ones in Salzburg (Austria) in the 1980s, and in Girona (Spain) and other places later on, were inconceivable without Henk's contributions and recommendations. Also in his Tilburg years, Henk, as a true social animal, organized numerous conferences, which often led to published volumes with major contributions to the field. During the last ten years, when Henk became more and more skeptical about certain developments in the field and their lack of solid empirical foundation, he doubled his consolidating efforts, which produced, among other things, the monumental and truly encyclopedic SynCom project (The Syntax Companion), which promises to become the best resource for information on what our field has achieved during its first 50 years. As everyone knows, there are many other aspects of Henk's personality that we have hardly touched upon. He is, for instance, a connoisseur of opera and an impressive singer himself, with a beautiful, deep bass voice. Henk could not have succeeded as the architect of the European infrastructure of our field if he had not been a warm and loyal friend to many, always organizing well-catered social events, always sharing his enthusiasm about good food and wine. Many of us will never forget Henk's descents into his well-supplied wine cellar, returning after a while with some superb grand cru and a happy smile of anticipatory satisfaction on his face. With Henk, it's not just a matter of initiative and organizational talents, but also the right kind of personality and the right kind of charisma that have given 'glow' and cohesion to the social structure of our field. It is hard to believe that Henk will leave The Netherlands after 35 most successful years as a friend, inspiring colleague, as educator and as the founder of GLOW and of so many other institutionalized activities that give our field social cohesion and a sense of common enterprise. Since Henk is moving to Tuscany, he will once more have the advantage of a Swiss education and its attention for things Italian. He does not plan to go into full retirement, but still has many plans to organize linguistic activities in

Preface

XXI

his new environment. Needless to say, we are all very much looking forward to that and wish him and Elisabeth many happy years under the Tuscan sun! Most important of all, we would like to express our gratitude to Henk, not only on behalf of ourselves but on behalf of the entire generative community, for more than 35 years of inspiration and initiative. It is in friendship and great appreciation of what he achieved that we offer this collection of articles to him. Without Henk's unforgettable contributions, European generative linguistics would most certainly not be where it is today. Molte grazie, Henk!

October, 2005 Hans Broekhuis Norbert Corver Riny Huijbregts Ursula Kleinhenz Jan Köster

An intersubjective note on the notion of 'subjectification' Werner Abraham

1. The problem On explaining the apparent gap of epistemic readings of modal verbs in Old English and early Middle English as well as a number of other, independent diachronic phenomena and their historical emergence, Traugott (1989, 1995, 1997) speculated that the notion of rising Subjectification and, more recently (2003), also Intersubjectification has played a determining role.1 Many others have adopted this notion as a driving and controlling force behind a host of grammaticalizing developments (Krug 2000, Verhagen 2005). The present paper presents arguments to the effect that such a notion may methodologically unhelpful or even undesirable and empirically unsatisfactory in explaining diachronic facts in general. In particular, it will demonstrate how a different line of explanation can be pursued 2 making use of a purely linguistic reasoning in terms of marking (redundancy as well as underspecification) and optimizing the relation between form and function. This is how the discussion will be organized. First, the term of Subjectification will be highlighted in all due brevity quoting Traugott herself and other authors who have followed her in employing the notion of (Inter) Subjectification to explain diachronic phenomena. Second, it will be argued for Old and Middle High German that modal verbs show evidence of the polyfunctional uses of modal verbs in diachronic texts and that an account of such historical material in terms of Subjectification, though not impossible, can be replaced by two scenarios in terms of concrete semantic and syntactic accounts: the Modal Aspect Hypothesis and the Marking and Optimizing Hypothesis. A simple

1 I have profited greatly from discussions with Elisabeth Leiss, Munich and exposition to audiences in Florianopolis/Brazil (Semana de Eventos 2005), Santiago de Compostela (NRG 3 2005), Paris (IVG 2005) and Lund, Sweden in Sept. 2005. 2 There is, at least indirectly so, Müller's (2001) account drawing a parallel between deontic readings of modal verbs in the Middle High German prose Lancelot and the rise of coherently embedded infinitives, on the one hand, and epistemic readings and biclausality, on the other hand.

2

Werner Abraham

concrete discussion on the synchronic polyfunctionality of the modal reading of aspectual werden 'become/be' will serve the purpose of showing that hard-core linguistic reasoning suffices to make the required predictions. 2. The pragmatic notion of (Inter)Subjectification and its alleged force Consider the following definition Traugott (1996:5): 'Subjectification is a phenomenon affecting large domains of the lexicon. If the meaning of a lexical item or construction is grounded in the world of reference, it is likely that over time speakers will develop polysemies grounded in their world, whether reasoning, belief, or metatextual attitude to the discourse. In other words, Subjectification is a semasiological development of meanings associated with a meaning-form pair such that the latter comes to mark subjectivity explicitly.' The general line of diachronic emergence (for epistemic='subjective' modals) is as in (1) below (Traugott 2003: 128; '>' standing for 'grammaticalizes to'). (1)

non-subjective > subjective > intersubjective

Intersubjectification, 'in the sense of the development of meanings that encode speaker/writers' attention to the cognitive stances and social identities of addressees, arises out of and depends crucially on Subjectification.' (Traugott 2003: 124). (1) schematizes the complete grammaticalizing cycle (Traugott 2003:134). Among the concepts the author discusses as instances of this diachronic control mechanism are Japanese politeness markers (emerging from spatial lexemes), modal particles in German, epistemic adverbs in English, promise and threaten in their heavily bleached epistemic raising meaning, epistemic well, the development of the perfect aspect (tendency toward analytics), the numeral one eventually yielding only, and others. Non-subjective lexemes such as the root modals in Old High German darf, seal, mag, kan, toug, muoz, Old English dearf, sceal, mceg, con, deaz, mot are thus thought to yield epistemic (= subjective) readings as their subjective tint emerges in due course. Root/deontic readings of modal verbs (MV) are thus taken to be 'non-subjective'. Notice that this way of speaking about diachronic development is akin to Meillet's jargon about 'greater expressivity' — a notion which is also often appealed to as a driving force in diachrony.

Subjectification

3

An attempt at bringing to a close the notion of Subjectivity and Subjectification with other notions in the literature does not appear to yield much. •

•

•

One can consider a version of Theory of Mind. 'Subjective' then could mean something like seeing things through a crucially different mind, like a child's below the age of two, or a lion's. If indeed (1) with 'non-subjective > subjective > intersubjective' completes and closes the full grammaticalizing cycle, it would appear that the two notions of Non-subjectivity and Intersubjectivity would be identical developmental states. Notice that, if the entire world of addressees is implied, the referential state of Non-subjectivity must be reached. This renders the state of Intersubjectivity as a separate developmental diachronic state superfluous, since non-distinct from the source state. Finally, if a purely linguistic line of argument leads to an identical result as under the controlling factor of (inter)subjectification, the vaguer notion not belonging to any calibrated linguistic layering should be given up to the benefit of the latter.

We will indeed claim that such a an explanation under calibrated linguistic layering is possible in accounting for what has been seen as the 'emergence of epistemic modal readings'. Two empirically plausible grammaticalizing paths will be sketched: first, the Aspect Hypothesis according to which aspectual sensitivities of DMV vs. EMV will be exploited for an explanation; and, second, the Marking and Optimality Hypothesis, which makes use of modal over- and under specification. Let us first look at some diachronic material. 3. Terminativity/root vs. interminativity/epistemics It is held that root modals, but not epistemic modals have a event-semantic structure as perfective full predicates. See (2a) for the full perfective verbs and (2b) for DMV. The event-semantic structure of epistemic MVreadings, on the other hand, is akin to that of imperfective verbs as in (3a,b).

4

Werner Abraham

(2) a. terminative (in)transitive verbs: sterben 'die' and töten 'kill' respectively: ®ext' ®int tj

® int

E,

tm

E2

tn

b. DMV: cf. mögen, müssen, sollen corresponding etymologically to 'may, must, shall': w J Θext' Θint Θ int |-»-»->·-»-»I 1 tj

Ε,

tm

E2

tn

(3) a. non-terminative (in)transitive verbs: leben 'live' and stoßen 'push', respectively: ®ext> ®int

b.

EMV/raising-V: ®ext' ®int tn

The quasi-auxiliaries (raising verbs, according to their syntactic derivational properties) scheinen 'seem' and pflegen 'be used to' are monophasic, as are EMV. They have no similarity to terminatives and their event structure, respectively. This is in line with the picture that we receive about EMVs as well as the epistemic readings of drohen 'threaten' and versprechen 'promise' in their use as semantically bleached raising verbs. On the other hand, their control readings are bi-phasic as shown in (2a,b). Notice that perfectives as bi-phasics as well as root MVs, DMV, project futural meaning (viz. The manager promises to help and its fiitural reading). This is in line with the empirical fact that MVs such as will and shall have substituted for the expression of future tense auxiliaries throughout the Germania. According to the linear distributions (Abraham 2001), the following operator relations hold for DMV vs. EMV. See (4) and (5).

Subjectification

5

(4) a. Mood scoping over the whole proposition is open for DMV and EMV: MOD](TEMP (p)) b. Tense scoping over the whole proposition excludes EMV: TEMP(MOD2 (P)) soll/muß[MoD] ...DMV, EMV that he in A. lived have shall/must 'that he had to/was supposed to have lived in A.' b. daß es hat ein Fehler gewesen sein sollen/müssen ... DMV, *EMV that it has a mistake been be shall/must 'that it must have been a mistake'

(5) a. daß

er

in A. gewohntpart habenTEMP

The underscored predicate components carry finiteness. The representations in (6a) and (6b) correspond to (5a) and (5b), respectively. (6) a.

SUPPOSE MOD, sollshallb. ANTERIOR TEMP hab- (ge-t) have been shall

(ANTERIOR TEMP ge-t habbeen have (OBLIG MOD2 soil live

(V)) Ρ wohnlive (V)) Ρ wohn-

.. DMV, EMV

DMV, *EMV

The scope relations in (5) are in line with the syntactic relations as sketched in (6). Notice that XP* can merge only in English non-V2. The examples under {XP*} exemplify what can be inserted and what not. Modern English full verbs do not raise out of VP, while MVs occur only in their epistemic or tense readings (Abraham 2001). Modern German DMVs merge in VP, while EMV merge late above VP, way-up beyond IP - a conclusion which is corroborated by the independent findings in (8) and (9).

6

Werner Abraham

(V)

MV/Aux V,E n g l i s h V (XP*} = Neg =? = emph. do = adverb = Quant (8)

German

V/DMV/Aux *He hopes not to go/*hopen't *Hope we to go? I do hope to come *We hope always to go *They hope all to come

EpistP > TP > (VP >) DeontP > AspP > (Aktionsart)V 0

Typological adverb and agglutinal affix hierarchy (Cinque 1999) The conclusion drawn from the distributions in Modern German appears to have solid cross-linguistic support in the LI-research by Hyams (2005).

Subjectification

4.

7

The first purely linguistic diachronic account for missing EMV: the Modal Aspect Hypothesis

Can the aspect-based concept in (2)-(5) be divided up into components such that clearer, purely linguistic factors are yielded in order to justify the emergence of something that is absolutely implausible to not have existed before 'Subjectivization' has taken possession of mankind (no earlier than Early Modern English)? Is it not a lot more plausible to assume that Epistemicity, also as the opposition correlate of deontic/root MVs, was at the disposal all along in the diachrony of English (and earlier). How could such a modal status be confirmed given the absence of EMV in Old English and Old High German (and Gothic and Modern Russian and Modern Armenian)? Assume that (strong) Aspect renders the opposition and distribution with the triggering momentum provided by Imperfectivity for Epistemicity. This - if empirically confirmed - would render Traugott's concept of Subjectification epiphenomenal. Notice that such an assumption is critically dependent upon the links observed to hold between root readings and perfectivity and, vice versa, imperfectivity and Ε-readings in Modern German (but not English). It is assumed that Modern German aspectual sensitivities of DMV vs. EMV can be exploited for an explanation of why Ε-readings are scarce to non-existing in the early stages of English and German. Let us assume that we can carry over the findings for Modern German - i.e., the finding that both D- and Ε-readings are yielded for embedded perfective infinitives, but that Ε-readings are out for embedded imperfective infinitivals (cf. (2)-(3) and (5a,b) above). In Abraham (2002) it was speculated that the demise of root readings of MVs in English (specifically American English) is due to the demise of Aktionsart and aspect in early Middle English (Denison 1993). What are left over of English modals, with some lexical islands (such as You may) and with newer lexical substitutes (such as have to for must or be allowed/permitted to for may; see Krug 2000) are epistemic or tense readings (Abraham 2002). Since Old English had aspect and Aktionsart like Modern German, it could have expressed modal epistemicity on the basis of aspect distributions. Research on strong aspect languages such as Gothic and Old High German (Leiss 2002) as well as Modern Russian and Modern Armenian (Gevorgyan-Ninness 2005) appears to support this assumption. I repeat the respective distributional relations from (3)-(4) above.

8 Werner Abraham (10)

a.

DMV:

| ->->->->->-> | tj

b.

5.

~Έ|

tIT

EM V/raising-V:

The second purely linguistic diachronic account for missing EMV: the Marking and Optimizing Hypothesis.

Let us assume that epistemic readings arise from mismatches between tense marking and aspectual tense reference such that if tense agreement, TP, cannot be fully satisfied, the next categorial node, EpistP, must be realized. Recall (7)-(8), which open up such a optimizing mechanism. See the following examples restricting the use of werden as an optional epistemic verb. [±ASS] = modal assertiveness; EM=epistemic modality; [-pert] = Aktionsart-imperfectivity; Spr=speech act epistemicity, Prop=proposition epistemicity]. (11) a.

Was geschieht gerade? Wo ist Anna? (Leiss 2002: 95) what is happening right now? where is Anna? Tense reference of Ρ = Present; Predicate tense = Future; collision of time reference and Tense form results in modal-epistemic reading. Sie wird in der Badewanne liegen she becomes/will in the bath tub lie ... liegen = [-perf]; EM; [-ASS] Spr = Sie liegt wahrscheinlich in der Badewanne she will probably in the bath tub lie Sie wird auf der Terrasse eine Zigarette rauchen she will on the terrace a cigarette smoke (Leiss 2002: 95)

Subj edification

9

b. Tense reference of Ρ = Future; Predicate tense = Future; verb perfectivity yields collision of seemingly compatible forms: PerfV has future reference in the present tense form. This results in epistemic modalization due to non-optimal tense form. Er wird zuhause eintreffen, he will at home arrive ... eintreffen = [+perf]; EM; [+ASS] Spr = Er trifft mit Sicherheit zuhause ein it is beyond doubt that he will be home c. Tense reference of Ρ = Future; Predicate tense = Present; Present tense is the optimal tense form for perfective verbs to express future time reference; therefore, future reading without epistemic modalization; only seeming collision of time reference and Tense form - as soon as aspect/Aktionsart is considered the Present tense form is the optimal tense form to yield optimal time reference. Er trifft zuhause ein. ... assertive Present/Future; [+ASS] Plop he is at home arriving Despite featuring identical perfectivity as ( l i b ) , (11c) reflects speech act assertiveness under a purely temporal reading. By contrast, werden both in (11a) and ( l i b ) triggers the epistemic meaning ofnon-assertiveness despite the fact that aspectual Aktionsart is mutually exclusive ([-pert] Φ [+perf]). The argument derivable from these illustration is this: Like other modal verbs, werden 'become' is inchoative, i.e. perfective. Neither by embedding a perfective nor an imperfective infinitive (eintreffen 'arrive' vs. in der Badewanne liegen 'lie in the bath tub') is werden 'become' able to trigger a pure assertion. This is appears to be due to its perfectivity which points to the future - which is what perfectives usually do. Think of the fact that Russian budu 'become' forms future expressions only by embedding imperfective verbs. Infinitival perfective embeddings are excluded in Russian. Think also of the diachronic evidence, respectively the lack thereof, that Middle High German (MHG) modals (usually prQ-werden sculan 'shall') expressing future temporality sided with imperfective verbs only. MHG perfective predicates expressed temporal futurity when used in the present in themselves in the first place. Consider again (10a) above: Besides the epistemic reading assumed above, Er wird zuhause sein 'he will be at home' or Er wird zuhause eintreffen 'he will arrive at home', respectively, may mean "he is going to be at home" alright. However, its stylistic function is an epistemically highly marked statement to the extent that its assertiveness is far from ascertained. The very fact that its

10 Werner A braham assertiveness needs this assertive expression by form, makes its epistemicity shaky - as if its truth value were hyper-ascertained against the odds. Old and MHGerman modal verbs, it is assumed as an option to claims extended by Traugott and others, behaved the same way as modern werden being likewise perfectives. In short, what we appear to have is a sort of competition of forms: If formal tense is not employed optimally, what is yielded is epistemic modality. In other words, the LF-reading check raises a node higher: from TP to EpistP (cf. Cartography by Cinque 1999; cf. (7)-(8) above. Needless to say that it remains to be seen whether such reasoning can be applied to MV-occurrences in earlier stages of German and English. However, given the weak notion of (inter)subjectification this diachronic research line is worth-while pursuing. 6.

In conclusion, with respect to covert modal distinctions: Do EMV really arise historically out of nothing due to profounder cognitive alertness?

It was speculated in this paper that it is not plausible to assume that speakers of Old High German, Gothic, and Old English were cognitively less distinctive or poorer, and socially less attentive, than modern speakers - let alone speakers of Modern Russian, and Armenian the last two certainly near-devoid of epistemic (and deontic) modals. We suggested that there is reason to believe that solid linguistic distributive, though modally covert, phenomena were there to substitute for epistemic modals. It is to be noticed in this context that epistemic adverbs are unlikely to take over the function of the versatile EMV. Another argument for the distinct behavior of EMV and DMV may rest in the fact that they extend different syntactic and semantic scopes in the clausal predication - cf. (4)-(6) above: DMVs have a very narrow, probably VP or V° scope, no wider than the verbal complex due to their aspect/Aktionsart property. EMV, on the other hand, has scope including AGRsP and therefore TP at least. What are possible verification paths? The demise of DMV and full verb properties of MV must be linked to the near simultaneous rise of Τ in ME and Early Modern English (van Gelderen 1993). In other words: EMV uses should have exploded pursuant to the rise of TP. However, this assumption is valid only in the case that EMV needed AGRsP (obligatory finite occurrence) all along from OE onwards. While Denison (1993:310) has shown that, much like in Modern German, (D)MV occur non-finitely in ME, there is no evidence for the occurrence of EMV

Subj edification

11

in non-finite f o r m (which is in line with the E M V non-finiteness gap decidedly observable in German, to all appearances in all diachronic stages ( A b r a h a m 2002). See (5b) and (6b) above.

References Abraham, Werner 1991. Modalverben in der Germania. In Ε. Iwasaki (ed.) Begegnung mit dem 'Fremden': Grenzen-Traditionen-Vergleiche. Akten des VIII. Internationalen Germanisten-Kongresses in Tokio 1990. Vol. 4, 109-118. München: ludicium. Abraham, Werner 2002. Modal verbs: Epistemics in German and English.' In S. Barbiers, F. Beukema and W.v.d. Wurff (eds.). Modality and its interaction with the verbal system, 19-50. Amsterdam-Philadelphia: John Benjamins. [Linguistik Aktuell/Linguistics Today 47]. Cinque, Guglielmo 1999. Adverbs and functional heads: a cross-linguistic perspective. New York: OUP. Denison, David 1993. English historical syntax. London: Longman. Gelderen, Elly van 1993. The rise of functional categories. Tübingen: Μ. Niemeyer. Gevorgyan-Ninness 2005. Die Herausbildung des epistemischen Ausdrucks im Deutschen, Russischen und Armenischen. Frankfurt: Peter Lang. Hyams, Nina 2005. Child non-finite clauses and the mood-aspect connection: Evidence from child Greek. In P.M Kempchinsky and R. Slabakova (eds.) Aspectual inquiries. Dordrecht: Springer. Krug, Manfred G. 2000. Emerging English modals. A corpus-based study on grammaticalization. [Topics in English Linguistics 32], Berlin/New York: Mouton de Gruyter. Leiss, Elisabeth 2002. Explizite und implizite Kodierung von Deontizität und Epistemizität: über die grammatische Musterbildung vor der Entstehung von Modalverben. Jezikoslovlje 3.1-2: 69-98. Meillet, Antoine 1912. L'evolution des formes grammaticales. Scientia (Rivista di Scienzaj 12.6, 384-400. Müller, Reimar 2001. Modalverben, Infinitheit und Negation im Prosa-Lancelot. In R. Müller and M. Reis (eds.) Modalität und Modalverben im Deutschen, 239262. [Linguistische Berichte Sonderheft 9]. Hamburg: Buske. Traugott, Elizabeth 1989. On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 65/1: 31-55. Traugott, Elizabeth 1995. Subjectification in grammaticalization. In D. Stein and S. Wright (eds.) Subjectivity and subjectivization: Linguistic perspectives, 31-54. Cambridge: CUP. Traugott, Elizabeth 1997. Subjectification and the development of epistemic meaning: The case of promise and threaten. In T. Swan and O. Westvik (eds.) Modality in Germanic languages: Historical and comparative perspectives. Berlin/New York: Mouton de Gruyter.

12 Werner A braham Traugott, Elizabeth 2003. From Subjectification to intersubjectification, 124-142. In: R. Hickey (ed.) Motives for language change. Cambridge: CUP. Verhagen, Arie 2005. Constructions of intersubjectivity. Discourse, syntax, and cognition. Oxford: OUP.

A note on non-canonical passives: the case of the ^/-passive Artemis Alexiadou 1. Introduction1 In many languages, a passive-like meaning may be obtained through a noncanonical passive construction. The get passive ( l b ) in English, the se faire passive (2b) in French and the kriegen passive (3b) in German represent typical manifestations: 2 (1) a. John was killed in the war. b. John got killed in an accident.

(Haegeman 1985:53)

(2) a.

Jean etait ecrase par une voiture. Jean was run-over by a car b. Jean s'est fait ecraser (par une voiture). Jean refl aux made run-over (by a car) 'Jean was run over by a car.' (Labelle 2002:1)

1

1 am really happy that I can offer this squib to Henk with gratitude for his support and guidance over the years. Thanks are due to David Embick, Winfried Lechner and Florian Schäfer for comments and discussion. 2 For English, see Siewierska (1984), Haegeman (1985), Arce-Arenales et al. (1994), Givon and Yang (1994), Fox and Grodzinsky (1998), Huang (1999), Taranto (2004), and Mclntyre (2005) among others and references therein; for French, see Labelle (2002) and references therein; for German, there is a lengthy dispute as to whether or not this is a passive structure, see Haider (1984), (1986), (2001), Fanselow (1987), Reis (1985), Cook (2004) among others. In addition to kriegen, erhalten and bekommen 'receive/get' may be used as passive auxiliaries in German. As Siewierska (1984) and references therein notes, the three auxiliaries are not interchangeable. Note that in German the kriegen construction promotes only a dative argument which, for some speakers, must cooccur together with an accusative argument, as in (3c).

14 Artemis

Alexiadou

(3) a.

Hans wurde getötet. Hans was killed b. Er kriegte seine Miete von der Firma He got his rent by the firm

bezahlt. paid (Siewierska 1984:132)

This squib focuses on the behavior of the ge/-passive in English and discusses a number of restrictions associated with it as well as the status of get. The gei-construction and its cognates in the other languages are particularly interesting as they raise a number of questions for both the analysis of passives as well as for the standard distinction between major lexical heads and functional heads, and the potential existence of semilexical heads (see Riemsdijk 1998: 1-48, Haider 2001). Standardly, functional heads are assumed to lack a lexical-conceptual argument structure, they select their complement solely in terms of morphosyntactic and category features of the head of the complement. Van Riemsdijk (1998) provided convincing arguments for the existence of semi-lexical heads in the nominal domain, and a number of the contributions to Corver and Van Riemsdijk (2001) suggest the existence of these heads in the verbal domain. With respect to this partition, this squib will suggest that get is a further instantiation of a semi-lexical head in the verbal domain. In particular get will be seen here as the semi-lexical variant of a major lexical head, since it lacks argument selection properties (Emonds 1999, Van Riemsdijk 1998, Haider 2001). This becomes clear if we contrast (lb) with the examples in (4) and (5). While in (1) get does not seem to license the thematic role of the subject, it permits constructions where it does seem to license argument structure. These include (a) cases in which it functions as a 'lexical' verb (4) and (b) constructions which are active/causative, where the subject is interpreted as the causer of the action described by the participial clause (5). The behavior shown by get is similar to that of have and need in English which have been discussed in the literature: (4) (5)

Susan got a book. 3 John got Mary blamed for the accident.

3

Mclntyre (2005) discusses a further reading of get, which he labels 'hindrance' reading, illustrated in (i). (i) means that the result is hard to attain: (i)

I didn't get the key in the lock.

Non-canonical passives 15 The semi-lexical nature of get is manifested by a series of criteria that disambiguate pure functional heads, e.g. auxiliaries and lexical verbs. As demonstrated in examples of the type in (lb), get patterns more like a lexical verb than like an auxiliary, in that it shows an atypical behavior for auxiliaries in a number of contexts, including negation contraction and question formation (6): (6) a. Did he get killed?/Was he killed/*Got he killed. b. He didn't get killed/He wasn't killed/*He gotn't killed. In what follows, I briefly summarize the properties of the ge/-passive as these have been described in the literature. 2. The properties of the gitf-passive Leaving aside the issue of the appropriate register (spoken vs. written language), a certain consensus seems to exist as far as the properties of the ge/-passive is concerned. First of all, unlike their be counterparts, the get passives lack an implicit external argument, since they are unable to control into purpose clauses and cannot license volitional adverbials (Huang 1999, Fox and Grodzinsky 1998, Taranto 2004 and others): (7) a. The ship was sunk [PRO to collect insurance money]. b. *The ship got sunk [PRO to collect insurance money], c. The ship got sunk [for John to collect insurance money]. (8) a. The book was torn on purpose. b. *The book got torn on purpose.

(Fox and Grodzinsky 1998:327)

As discussed in Fox and Grodzinsky (1998), these same conclusion: in ge?-passives, unlike in be argument of the VP has no implicit realization. Second, as pointed out by Arce-Arenales et al. compatible with reflexive action, while be-passives

contrasts all lead to the passives, the external (1994), ge/-passives are are not:

(9) a. I got dressed (by my mother or by myself), b. I was dressed (only by my mother). This is reminiscent of Kratzer's discussion on German participles, where she shows that the following distinction holds:

16 Artemis

Alexiadou

(10) a.

Das Kind war gekämmt. The child was combed Stative: compatible with reflexive action (no agent) b. Das Kind wurde gekämmt. The child was combed Eventive: incompatible with reflexive action (necessarily an agent)

The above two properties seem to suggest that the participle involved is actually an adjectival passive. In fact, this is the conclusion drawn by both Fox and Grodzinsky and Taranto. A further argument in favor of this view is provided by the thematic restrictions that have been observed for adjectival participles. As noted by Levin and Rappaport (1986) among others, adjectival passives of some (in fact the majority of) double object verbs do not allow goal externalization, while both arguments can be externalized in the case of the verbal passive: (11) a. The salesman sold the customer a car. b. The recently sold car c. *The recently sold customer (12) a. The car was sold to the customer, b. The customer was sold a car. If the gei-passive is based on the adjectival participle, one might expect similar restrictions on what passives are possible there. (13) confirms this: goal externalization is not possible: (13)a. The car got sold to the customer. b. The customer got sold a car.

(Siewierska 1984: 132)

The third property characterizing the ge?-passive is that, as opposed to the &e-passive, it does not seem to be fully productive: 4

4

Note, however, that verbs that do not normally have transitive counterparts can form the gei-passive (see Arce-Arenales 1994: 15): (i)

a. He got fogged in. b. *The weather fogged him in. c. The cotton balls got decayed. d. *The bad weather decayed the cotton balls.

Non-canonical passives 17 (14) a. b. c. d. e.

*The truth got known. *Mary got feared. *Mary got followed by a little lamb. *Mary got seen. *The electricity light got invented.

It has been noted that the g