LinguistikInternational
Herauscegeben von
Heinrich Weber
Susanne Beckmann
Abraham P. ten Cate
WlLFRIED KURSCHNER
Kazimierz Sroka
Ingo Warnke
Lew Zybatow
Band 9
PETER LANGFrankfurt am Main • Berlin • Bern • Bruxelles • New York • Oxford • Wien
Peter Kosta/Jens Frasek (eds.)
Current Approaches to Formal
Slavic LinguisticsContributions of the
Second European Conference on Formal Description of Slavic Languages FDSL II
held at Potsdam University,November 20-22, 1997
PETER LANGEuropaischer Verlag der Wissenschaften
Bibliographic Information published by Die Deutsche BibliothekDie Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the internet at <http://dnb.ddb.de>.
ISSN 1436-6150 ISBN 3-631-50311-3
US-ISBN 0-8204-6060-5© Peter Lang GmbH
EuropSischer Verlag der Wissenschaften Frankfurt am Main 2002
All rights reserved.All parts of this publication are protected by copyright. Any
utilisation outside the strict limits of the copyright law, without the permission of the publisher, is forbidden and liable to prosecution. This applies in particular to reproductions, translations, microfilming, and storage and processing in
electronic retrieval systems.Printed in Germany 12 3 4 6 7
www.peterlang.de
5
Table of Contents
Preface......................................................................................................................................... 7
Leonard H. Babby, Argument Suppression and Case in Russian Derived Nominals................9
Steven Franks, The Pros and Cons of Clitic Cluster Formation..................... ......................... 23
Jens Norg&rd-Serensen, Individuation in Russian: Gender semantics and Other Issues.........39
Tania Avgustinova, Clustering Clitics in Bulgarian Nominal Constituents............................ 63
Kai Alter & Uwe Junghanns, Topic-Related Prosodic Patterns in Russian............................. 73
Lukas Szucsich, Case and Configuration: The Problem of Nominal Adjuncts.......................89
Olga MiSeska Tomic, The Site and Status of the South Slavic Negative Particle...................99
Andreas Spath, On Definite Reference of Determinerless Nouns in Russian.......................109
Danko Sipka, A Decisionist Model of Slavic Morphology................................................... 119
Nenad KonCar & Danko Sipka, Slavic as a Source and Target Language in NeuroTran®Sentence Translation........................................................................................................129
Bistra Andreeva & William J. Barry, Intonation von checks in der Sofia-Varietat desBulgarischen....................................................................................................................139
Nancy Smith & Bernhard Staudinger, Einige Probleme der Aspektberechnung in einemdeutsch-russischen Ubersetzungssystem........................................................................ 149
Juliane Lagunov, Modaltransfer Russisch-Deutsch: Disambiguierung mittelsThema-Rhema-Gliederung............................................................................................. 159
Ljudmila Geist, Russisch byt'\ zwei Kopulae, zwei Kasus, ein integrierter Ansatz..............169
Assinja Demjjanow, Semantik aspektueller Verben im Russischen........................................179
Hermann Fegert, Die Hierarchie der Verbbetonungsmuster im Russischen.......................... 189
Andrzej Boguslawski, Zur sogenannten .negative transportation*..........................................197
Irina Sekerina, The Late Closure Principle in Processing of AmbiguousRussian Sentences...........................................................................................................205
Richard Zuber, Some Categorially Polyvalent Modifiers in Polish........................................ 219
6
Ivonku P. Schick, Clitic-Doubling Constructions in Modem Bulgarian.................................227
Peter Kostn, Minimalism ond Free Constitutent Order in Russian.........................................253
Barbara Kunzmann-Mtlllcr, Impersonalc Konstruktionen im Slavischen..............................273
Jarmila Pnncvovd & Jan HajiC, The Syntactic Tagging of Corpora: New Issuesfor Explicit Syntactic Description of Czech...................................................................285
Elena Budnitskaya. Long-distance Binding into Russian Infinitives: the Blocking Effectof Overt Complementizers and Dative Subjects............................................................ 291
Petr Sgall, Formalizing u Functional Description................................................................... 299
Gerald Penn, On the Plausibility of Purely Structural Multiple WH-Fronting.......................307
Roland Meyer, M'/i-Qucstions without Overt IfTj-Movement in Russian and Polish.............317
Milan MihnljcvuS, The Structure of Croatian Alternative Questions..................................... 327
Anna Knp.46 A Adam Przcpi6rkowski, Morphological Aspects of Verbal Negationin Polish............................................................................................................................ 337
Eva Ht\ji6ovA, Focalizcrs and their Status in the Topic/Focus Articulationof the Sentence..................................................................................................................347
1 I<Mdne Le Guillou de Penanros, Prefixes-prepositions Revisited.............................................357
Chris Wilder <fc Damir 6avar, Verb Movement, Cliticization and Coordination.................. 365
Joanna Blaszczak, Towards a Binding Analysis of Negative Polarity Items in Polish......... 377
Adam Pr/.cpitSrkowski, Verbal Proforms and the Complement-Adjunct Distinctionin Polish....................................................................................................................... 405
Loren A. Billings, Word Order and Argument Structure of Russian Psych Predicates........ 415
Leonid Birjulin, HTcpimiBiu>ic KoHCTpyiatHH:Ohi.it CoMaimiMCCKoro MoACJinpoBaiuiH..................................................................... 425
FDSL 11,415-424
Loren A. Billings
Word Order and Argument Structure of Russian Psych Predicates
So-called psych verbs - those which subcategorize for an argument with the Experiencer thematic role - have received considerable attention in the scholarly literature because they are apparent exceptions with regard to various syntactic effects. Such verbs have also been argued to project clauses with unusual word orders. This paper investigates psych verbs in Russian to determine both their argument structure and their neutral constituent order.
In neutral discourse, Russian transitive clauses have subject-verb-object order. It has also been observed that dat-case Experiencers in Russian appear first in a clause. This work deals additionally with whether dat Experiencers are clause-initial due specifically to their having an Experiencer role. That is, I factor out other variables - whether the Experiencer (i) is animate, (ii) appears as a pronoun, or (iii) is realized using a particular morphological case. Specifically, this paper makes the following proposal: Among verbs which subcategorize for two arguments, only those with noncanonical argument structure - i.e., those without an external argument - display noncanonical constituent order.
Section 1 assesses psych verbs’ argument structure. Section 2 then lists such verbs in Russian by type. Finally, section 3 discusses their neutral word order.1
1. Argument StructureThis section assesses the morphosyntactic properties of psych verbs in order to determine their argument structure. This is done by assessing first the literature on other languages, and finally research specifically on Russian psych verbs.
1.1. The literature on Italian and BulgarianBelletti & Rizzi (1988), in their seminal generative study of psych verbs, assess three types of predicates in Italian, exemplified in (la-c). They propose that in (la) the nom- case argument is an external argument, while in (lb-c) both arguments are mapped into the syntax VP-intemally. Belletti & Rizzi argue therefore that there is no external argument in either of (lb-c); i.e., these verbs are unaccusative. Moreover, they propose that the dat-case Experiencer in (lb) and the acc-case Experiencer in (lc) are assigned their case inherently.
1 I am indebted to the following linguists for their valued suggestions: Leonard Babby, Stephanie HARVES, Uwe JUNGHANNS, Eric Komar, Catherine RUDJN, Maaike Schorlemmer, Petr SGALL, and Gerhild ZYBATOW. I also thank my two primary informants for enduring literally thousands of queries: Dmitrij FEOKT1STOW and Irina KADUKOVA. Any errors that remain in this paper are my responsibility alone.
416 FDSL II
(I) a. Gianni teme questo.G NOM fearsj.sG this^cc Experiencer
b. Questo piace a Gianni.this^M/ is.pleasingj.5G G-DAT
Type 1. ‘Gianni fears this.’
Theme Type 2.
‘This is pleasing to Gianni.’ThemeQuesto preoccupa this//GA/ worriesj.5G Theme
Experiencer (also attested: A Gianni piace questo.) Gianni. Type 3.G.acc ‘This worries Gianni.’Experiencer
Linguists agree that (la) is a garden-variety transitive clause, with an external-argument subject. Also, there is near consensus that (lb) is unaccusative.2 There is, however, widespread disagreement about the structure of (lc).
Grimshaw (1990) proposes instead that what distinguishes type-1 verbs from types 2 and 3 is the latter’s inability to project an external argument, which she defines as the argument with both causative semantics and the thematically highest role. Whereas all external arguments are mapped into the syntax as D-structure subjects, not all D- structure subjects are external arguments. The distinction is pre-syntactic. Grimshaw adds that most morpholexical operations affect only (underived) external arguments; notable among the verb types which lack an external argument (by her definition) are type-3 psych verbs, as in (lc), because the thematically highest argument (the Experiencer) does not have causative semantics. By this reasoning, even though this predicate selects a D-structure subject (the NOM-case Theme), there is no external argument. The other verb type without an external argument is unaccusatives (e.g., go), because the predicate’s only argument lacks causative semantics.
Otherwise ignoring type-2 verbs, Grimshaw (1990:29-30) admits that this type is unaccusative. That is, these verbs lack even a D-structure subject.
Another study, Pesetsky (1995), agrees with Belletti & Rizzi (1988) and Grimshaw (1990) in the relevant respects about the structures of (la-b). But Pesetsky crafts a compromise with regard to (lc). He proposes an elaborate model that base-generates the Theme in (lc) below the Experiencer (inside VP), followed by raising of this causative Theme argument (to SpecVP).
Before turning to the Russian data, a brief discussion of another Slavic language provides a hint for determining the structure of one Russian verb type. Slabakova (1996:250,269) reports that the distribution of cases and thematic roles of Bulgarian psych verbsis identical to the Italian pattern in (la-c) above. She concludes (p. 260) that in Bulgarian the NOM-case argument in each of types 1 through 3 is the D-structure subject. I disagree with regard to type 2. Few of the syntactic tests which Slabakova applies to types 1 and 3 are tested on type-2 verbs. Indeed, one of her diagnostics, that all three types can passivize (p. 259), is inconclusive: Scatton (1984:344, 348) reports that even
2 SLABAKOVA (1996) argues for the deep-subjecthood of the NOM-case argument in each of types 1-3 in Bulgarian. LEGENDRE & AKIMOVA (1994) propose that the Experiencer in each of types 1-3 in Russian is an underlying subject (“initial 1” in RG terms). KfNG (1994:119-20) proposes that with Russian type-2 verbs the NOM Theme is an external argument. (See my unaccusative analysis of type-2 verbs below.)
L. A. BILUNGS, Word Order and Argument Structure of Russian Psych Predicates 417
,walk’, an intransitive (and most likely unaccusative) verb, can passivize: xodeno. More evidence for the unaccusativity of Bulgarian type-2 verbs comes from the order in which multiply fronted wh phrases appear. Billings & Rudin (1996:40) report that, animacy being equal, any external-argument wh phrase must precede any other WH phrase. The only structure in which two WH arguments can be freely ordered is if both are internal arguments.3 Billings & Rudin use (2a-b) to show that if there is no external argument, then either argument can be clause-initial. That is, either of (2a-b) is acceptable.
(2) a. Koj na kogo mu xaresva? (Bulgarian)who^A/ to whom^r CLdat.3.SG is.pleasing5.5c
b. Na kogo koj mu xaresva? ‘Who is pleasing to whom?’
Unfortunately for the current study’s purposes, Stepanov (to appear) argues that Russian is unlike Bulgarian in an important respect: WH-movement is absent in Russian; other factors trigger the fronting of Russian WH phrases. Still, although the two languages differ with respect to WH-movement, the Bulgarian data suggest that both of the arguments of a type-2 verb are internal.4
1.2. Previous work on Russian psych verbsKing (1994) investigates psych verbs in Russian, which pattern the same as the Italian types in (la-c) with regard to the mapping of cases to thematic roles.
Ja Ijublju starye knigi.hoM lo ve/.5c [old booksbcc./v.
Type 1.
Experiencer Theme ‘I love old books.’Mne nravjatsja starye knigi.me^j- are.pleasingj ^ [old books]WOM/>i
Type 2.
Experiencer Theme ‘I like old books.Menja interesujut starye knigi.me^cc interest^ [old booksJ^MPZ.
Type 3.
Experiencer Theme‘Old books interest me.’
The PL agreement in (3b-c) identifies the Theme as subject If starye knigi ‘old books’ in (3) were in the SG, then there would be a morphological distinction between acc.SG staruju knigu in (3a) and NOM.SG staraja kniga in (3b-c).The GEN of negation is one phenomenon that specifically affects an argument only if it is internal; cf. the bold-faced arguments in (4a-c): the Theme argument of types 1 and 2, and the Experiencer argument of type-3 verbs.
3 Billings & RUDIN (1996) present detailed arguments to show that either wh internal argument of a ditransitive verb (with a non-WH external argument!) can be clause-initial.
4 Slabakova (1996:269) further argues that with type-2 verbs the DAT case is quirky. This does not bear on BILLINGS & RUDlN’s analysis. See, however, section 1.2 below.
418 FDSL II
(4) a. Ja ne ljublju knig.\N0M not love/.*; booksc£MW. ‘I don’t love books.’
b. MaSe ne ponravilos’ niMaSa/xir not liked#.*; nary‘MaSa didn’t like a single movie.’
c. Ni odin student nenary [one student] (m)som not
Type 1.
[King (1994:119)] odnogo fil’ma. Type 2.[one movie] (m)gen.sg
[Legendre & Akimova (1994:299)] udivil ucitel’nicy. Type 3.surprised*,.*; teacher^#.*;
‘Not one student surprised the teacher.’ [King (1994:120)]
Moreover, the ACC-case argument of types 1 and 3 appears as the NOM-case subject of a passivized predicate; see King (1994:120). But type-2 predicates cannot passivize; this follows directly from the unaccusativity of type-2 verbs.5
(5) Summary so far Type 1 Type 2 Type 3a. Thematic role Exper Theme Exper Theme Exper Themeb. Case without negation NOM ACC DAT NOM ACC NOMc. With gen of negation NOM GEN DAT GEN GEN NOMd. In the passive INST NOM - - NOM INST
The italicized columns (type-1 Experiences and type-3 Themes) are conclusively shown to be external arguments, while the bold-faced columns (type-1 and -2 Themes and type-3 Experiences) are conclusively internal arguments. This leaves only the status of type 2’s DAT-case Experiencer undecided.King (1994) fails to consider another verb with the same respective cases and thematic roles as type-2 verbs, exemplified in (6b). But unlike (other) type-2 verbs this verb form has an apparent type-1 counterpart, shown in (6a):
5 Linguists agree about (4a), but there is disagreement about (4b-c). LEGENDRE & AKIMOVA report that a type-2 Theme does, but a type-3 Experiencer doesn’t, undergo GEN of negation. But KING (1994:119-20) reports that while a type-3 Experiencer can exhibit this phenomenon, a type-2 Theme cannot. My informants confirm the data as shown in (4a-c). This disagreement in judgments is not surprising. considering the increasing limitation on the use of the GEN of negation in Russian, documented in TlMBERLAKE (1975/1986). LEGENDRE & AKIMOVA base some of their proposal (that type-3 Themes are underlying subjects) on the unacceptability of (4c). Similarly, much of KING’S analysis (namely, that a type-2 verb’s NOM-case Theme is a D-structure subject) relies on the impossibility of (4b). Also, because she dismisses unaccusativity as an option for type 2, KING (1994:120) conjectures that lack of passivization may be due to inherent DAT case assigned to the Experiencer, which is incompatible with the NOM-case subject position. KING’S suggestion is supported by the following contrast between Russian and Czech (respectively): *Ivanu bylo pomozeno uiiteljami; Honzovi bylo pomozeno udteli ‘{Ivan/Honza }(M)dat wzsn.SG helped,*,*; (by) teachers/#^/*,.’ These examples’ source, La VINE (1997), proposes that in Russian an arbitrary parameter prohibits the quirky-DAT-case argument from being able to appear in the NOM. FOWLER (1996) confirms that Russian verbs which assign DAT (as opposed to ACC, GEN or INST) case, including some type-2 psych verbs, cannot passivize.
L. A. BlLUNGS, Word Order and Argument Structure of Russian Psych Predicates 419
(6) a. My vspomnili staruju pesnju. Resembles type 1.I.nom remembered/.^, [old song](f)acc.sg
b. Nam vspomnilas’ staraja pesnja. Resembles type 2.mQDAT remembered/rsc+REFL [old song](f)nom.sg ‘We remembered the old song.’6 [modified slightly from Babby (1998:12)]
Legendre & Akimova (1994) propose that (6b) is derived from (6a) via the inversion construction (the underlying subject’s demotion to indirect object, with concomitant promotion of direct object to subject). Babby (1998:11-14) disputes such a derivation based primarily on the inability of (6a) to passivize. Instead, Babby proposes that the predicate in (6) subcategorizes for two internal arguments: a direct internal Theme and an indirect internal Experiencer. More controversially, Babby argues that this predicate requires an external categorial position without subcategorizing for an external argument; one of the two internal arguments must therefore “externalize” to fill this position. Similarly to middles of transitive verbs, in (6b) the direct internal argument is externalized and realized with NOM case, with the concomitant appearance of the reflexive element (-sja ~ -s ■). In (6a) it is the indirect internal argument that is externalized, shedding its (semantic) DAT case and appearing in the NOM; this then allows the direct internal argument to revert to its (structural) ACC case.
Babby (1998:14 fn. 17) suggests that (po)nravit'sja ‘be pleasing’, exemplified above in (3b) and (4b), is similar to the predicate in (6), “but the affix -sja is permanently affixed to the verb [lexically], which forces externalization of the initial direct internal theta role, thus effectively blocking the external ization of the experiencer, which must therefore always be realized internally as the semantic dative case.” This suggestion is problematic, considering that three other type-2 verbs, which Babby does not discuss, do not bear reflexive morphology. (Section 2 lists an inventory for each type of psych verb.) But, as Babby points out elsewhere (not in connection with psych verbs), “The status of unaccusative verbs in Russian is a problem. We see [...] that external ization of the direct internal argument is accompanied by affixation of the suffix -sja. It is not clear why -sja is not affixed to many verbs that are classified as unaccusative in other languages” (1998:7 fn. 11). Without solving the -sja puzzle, it’s clear from the preceding circumstantial evidence that type-2 psych verbs are unaccusative; both arguments are base-generated within VP.
To conclude this section on argument structure, Russian psych verbs pattern with those of other languages. Type 1 takes a NOM Experiencer as external argument and an ACC Theme as direct object. Type 2 has no external argument or D-structure subject; both the DAT Experiencer and the NOM Theme are base-generated inside VP. Finally, a type-3 verb’s ACC Experiencer is mapped into the syntax VP-intemally, while its nom Theme is the D-structure subject (although perhaps not the external argument by Grim shaw’s definition).7
6 Legendre & Akimova (1994:314 n. 1) report that the structures in (6a-b) correspond to agentive and non-agentive interpretations, respectively.
7 PESETSKY (1995:98) discusses one facet of psych verbs specifically in Russian. A link exists between type-3 verbs (NOM Theme, ACC Experiencer), none of which take -sja, and verbs with -sja in which the
420 FDSL II
2. Inventory of Psych VerbsUsing the three types discussed above, I list the psych verbs in Russian according to the mapping of roles to cases (or PPs) in each; some verbs have morphologically related forms in other verb types; e.g. zalet' [1] and sozalet' [9]. I further group the verbs in each type according to whether the etymologically reflexive clitic -sja appears. This is especially noteworthy with type 1, because quite a few of the dozen or so -sja verbs listed by Israeli (1997:43-44) which take ACC case are psych verbs; all of these are of type 1. By comparison, no -sja verbs are attested in type 3 (which also assigns ACC case to an argument).
Types 4 through 10 each have a NOM-case Experiencer. The Theme is in a particular oblique case or prepositional phrase. For types 5 and 10, the entire set of verbs consists of (reflexiva tantum) counterparts of type-3 verbs.
Type 1. (Experiencer: nom; Theme: ACC) bogotvorit' (imperf.) ‘adore’(po)ljubit' ‘love’ obozat' (imperf.) ‘adore’ predpotttat' / predpodest ’ ‘prefer’ uvazat' (imperf.) ‘esteem’ bojat'sja (imperf.) ‘fear’(po)stesnjat'sja ‘feel shy toward’
cenit' (imperf.) ‘appreciate’(voz)nenavidet' ‘hate’ oplak(jv)at' ‘lament’ prezirat' /prezret' ‘despise’ zalet' (imperf.) ‘feel sorry for’ opasat'sja (imperf.) ‘be apprehensive of stydit'sja (imperf.) ‘be ashamed in front of
Type 2. (Experiencer: DAT; Theme: NOM)dosazdat' / dosadit’ ‘vex’ nadoedat' / nadoest' ‘bother’naskud (iv)at' ‘bore’ (po)nravit'sja ‘appeal to’
Experiencer is NOM and the Theme is expressed with an oblique case or a prepositional phrase. Compare, for example, the following sentences:
(i) Ee postupok udivljaet Ivana.[her action]**,.*; surprises*.*; Ivan^cc
(ii) Ivan udivljaetsja efi postupku.Ivanmm surprisesj.sc+REFL [her action]DxrsG
Class 3.‘Her action surprises Ivan.’
Class 4.‘Ivan is surprised at her action.’
(Classes 4 through 10 are discussed below in section 2.) By the same reasoning as in (6a-b) above, Babby (1998) would characterize (i) and (ii) as having an externalized indirect and direct internal arguments of the same predicate. One problem with such an analysis is that (i) can passivize (with an iNST-case ‘by’-phrase to prove it), as in (iii):
(iii) Ivan byl udivlen efi postupkom. Passive of (i).Ivanww [was surprised]MSC [her action]/^.*; ‘Ivan was surprised by her action.’
LEGENDRE & Akimova (1994:288, 302-03) identify this problem. Recall that Babby disputes the derivation of (6b) from (6a) on the grounds that (6a) does not passivize. PESETSKY reports that the Theme is causative in (i), but not in (ii). It appears that the causative semantics of the subject in (i) allows it to function as an initial external argument, with its own passive counterpart in (iii). Because of space limitations, I must forego discussion of this issue here. I plan, however, to revisit this issue in future work.
L. A. BlLLtNGS, Word Order and Argument Structure of Russian Psych Predicates 421
Type 3. (Experiencer: acc; Theme: nom)
Nearly every verb in type 3 has a rejlexiva tantum counterpart in another type (indicated with brackets following the gloss). In order to minimize space, these verbs are not repeated in the inventories of these types (4 through 7, 9 and 10).
(vz)besit' ‘enrage’ [7] interesovat' {imperf.) ‘interest’ [6] izumljat' / izumit' ‘astonish’ [4] nervirovat' (imperf.) ‘make nervous’ ocarov(yv)at' ‘enchant’ [6] potrjasat' lpotrjasti ‘shake’({po!ob})radovat' ‘gladden’ [4] razdrazat'/ razdrazit' ‘irritate’ [6] razo6arov(yv)at' ‘disappoint’ [10] (ras)serdit' ‘annoy’ [7] stesnjat' / stesnit' ‘make shy’ [4] udivljat' / udivit' ‘amaze’ [4] utesat' / utesit' ‘comfort’ [6] uzasat' l uzasnut' ‘terrify’ [6] zabavljat' / zabavit’ ‘amuse’ [6]
(<o)bespokoit' ‘trouble’ [9](.za)intrigovat' ‘fascinate’ naslazdat'/ nasladit' ‘delight’ [4] obizat7 obidet' ‘offend’ [7] ogorcat'/ ogorcit' ‘grieve’ [6] {{is/na})pugat' ‘frighten’ [5] razdosadovat' (perf.) ‘annoy’ [7] raz'jarjat'/ raz'jarit' ‘infuriate’ [6] razvlekat’/ razvled’ ‘entertain’ [6] sokirovat’ (imperf.) ‘shock’ strasit’ (imperf.) ‘frighten’ ustrasat'/ ustrasit’ ‘intimidate’ [5] uvlekat'/ uvlec' ‘captivate’ [6] vozmuscat' / vozmutit' ‘exasperate’ [6] zanimat'lzanjat' ‘interest’
Type 4. (Experiencer: NOM; Theme: DAT) See also type 3.soboleznovat' (imperf.) ‘commiserate with’
Type 5. (Experiencer: nom; Theme: gen) See type 3.
Type 6. (Experiencer: NOM; Theme: INST) See also type 3.prenebregat7prenebred' ‘scorn’ (po)ljubovat'sja ‘feast one’s eyes on’ (cf. 7)
Type 7. (Experiencer: nom; Theme: na ‘at’ + ACC) See also type 3.dosadovat' (imperf.) ‘be vexed at’ {po)ljubovat'sja ‘feast one’s eyes on’ (cf. 6)
Type 8. (Experiencer: NOM; Theme: pered ‘before’ + INST) preklonjat’sja /preklonit'sja ‘admire’
Type 9. (Experiencer: NOM; Theme: o(b(o)) ‘about’ + prep) See also type 3.sozalet' (imperf.) ‘feel sorry about’
Type 10. (Experiencer: NOM; Theme: v ‘in’ + prep) See type 3.
422 FDSLII
3. Word OrderThe project8 of which this study is a part seeks to establish three claims: First, only one constituent order is felicitous in discourse-neutral contexts. Next, only this neutral order allows focus ambiguity to obtain. A sentence with neutral order (and intonation) can have maximal, intermediary, or minimal focus; see Junghanns & Zybatow (1997). Finally, with this neutral order it is most likely that the leftmost constituent becomes interpreted as topic. This section attempts to deal with the first of these three claims: A neutral word order (with unmarked intonation) exists for each predicate; this order allows the linguist to infer the argument structure of that predicate as stored in the lexicon.
Section 1 repeats examples as they appear in the literature. Many of these are not characteristic of neutral discourse. (See, e.g., the use of pronouns.) In order to confirm whether the NOM-case arguments of types 1 and 3 are D-structure subjects, I devised the following test: I composed three-word sentences, with the verb and one noun for each of the two arguments of verbs from types 1, 2, and 3 (the main focus of the study) and 4 (the control group for type 2), and a sampling of types 5, 6, 7 and 9. I also included a few non-psych (non-y) verbs to confirm the experiment’s premise (that subject-verb- object order is neutral).
The elicitation begins with the minimal-discourse scenario in (7):
(7) — 6to delaes’? ‘What are you doing?’ A- Muftfil'm smotrju o Zivotnyx. ‘Watching a cartoon about animals.’ B- A 6to tarn proisxodit? ‘And what’s going on in it?’ A
Following the setup in (7), I gave a choice of two answers by speaker B. Both have neutral intonation and a single intonational phrase with no break after the first word (to avoid a strong-topic interpretation). An example, for the verb ljubit ‘loves’ is shown in (8). Based on the minimal context in (7), the intended gloss in (8) is with both nominals indefinite: ,A hyena loves a koala.’
(8) a. Giena ljubit koalu. b. Koalu ljubit giena.hyenayvoAf.se lovesj^ koala^ccso
The pair in (8a-b) is just the NOM-verb-ACC and the ACC-verb-NOM permutations. I repeated this for every word-order combination (except verb-initial), in either sequence. This comes out to 12 elicitations per animal-verb-animal set. For each verb of types 1 through 4, five animal-verb-animal sets were elicited, for a total of 60 tokens per verb. The choice of animal names controls for prosody, familiarity, and gender/declension. For example, in (8) both animal names are trisyllabic with second-syllable stress. The two animals in (8) are equally exotic, disallowing a favored topic in either one. All ani
8 The project, funded by the Deutsche Forschungsgemeinschaft and supervised by Professor Gerhild ZYBATOW at the University of Leipzig, is titled Argumentstruktur und Wortstellung als Mittel der In- formationsstrukturierung im Russischen (Argument structure and word order as a means of determining information structure in Russian).
L. A. BILLINGS, Word Order and Argument Structure of Russian Psych Predicates 423
mal names were of feminine gender and of the /-a/ declension. All verbs were 3.SG im- perfective.
Studies of word order in Russian, such as Bivon (1971), report that the SVO order is neutral. Not surprisingly, all type-1 verbs for both speakers had a clear preference for NOM-verb-other order. (I use the notation “other” to refer to the non-NOM argument.) In fact, all verbs except type 2 showed an unwavering preference for NOM-verb-other order. The table in (9) summarizes the results:
(9) Overall results9
Type Preferred order No. of verbs tested No. of Dairs1. NOM£xper-verb-ACC 77wne 3 1802. (Mixed; see discussion below.) 4 2403. NOM7/jgme"Verb-ACC£xper 4 2404. NOM£x/?er“v6rb-dfaf Theme 2 1205. NOM£xper-verb-GEN7y,e;Wg 1 126. NOM^p^-verb-INST Theme 1 127. NOM£fper-verb-[wa+ACC] Theme 2 248. (Not tested.) 0 09. NOMfixpgr-verb-to+PREP] Theme 1 12
10. (Not tested.) 0 0non-y NOM-verb-other 5 96
The results from the four type-2 verbs diverge significantly between the two informants, both of whom had great difficulty choosing a preferred order. At this point, I offer no explanation for this lack of a clear preference for type-2 verbs. Still, all the verb types with a NOM-case Experiencer resulted in clear NOM-verb-other order. These results alone confirm the hypothesis in this paper’s introduction: “... only those [verbs] with noncan- onical argument structure - i.e., those without an external argument - display noncan- onical constituent order.” In fact, these results go on to suggest that type-3 verbs have D- structure subjects. In this sense, I was not able to detect a distinction between Grim- shaw’s more restrictive definition of external argument (those D-structure subjects that are highest both thematically and aspectually) and the definition used by other linguists (namely, any D-structure subject). The lack of any dominant word order with type-2 verbs is not inconsistent with my proposals in section 1 above that type-2 predicates have no external argument.
9 The type-6 verb tested is interesovat'sja ‘be interested (in)’, which is the middle of type-3 interesovat ‘interest’. This suggests at least that the NOM-case argument here is a D-structure subject. However, due to the small number of tokens, the data in types 5 though 7 and 9, and for non-psych (non-y) verbs, should be treated as preliminary.
424 FDSL II
ReferencesBabby, Leonard H. (1998), Voice and Diathesis in Slavic. Presented at Workshop on Comparative Slavic
Morphosyntax; currently downloadable from <http://www.indiana.edu/~slavconf/linguistics/in- dex.html>. To appear in a proceedings volume (George Fowler, ed. Bloomington, in: Slavica).
BELLETO, Adriana & Luigi RiZZI (1988), Psych-Verbs and q-Theory. in: Natural Language and Linguistic Theory 6: 291-352.
BiLUNGS, Loren & Catherine Rudin (1996), Optimality and Superiority: A New Approach to Overt Multiple WH-Ordering. in: Jindfiich Tomann (ed.). Formal Approaches to Slavic Linguistics: The College Park Meeting, 1994. Ann Arbor: Michigan Slavic Publications, 35-60.
BiVON, R. (1971), Element Order. (= Studies in the Modem Russian Language, 7.) Cambridge, England: Cambridge University Press.
FOWLER, George (1996), Oblique Passivization in Russian. Slavic and East European Journal 40: 519-45.GRIMSHAW, Jane (1990), Argument Structure. Cambridge, MA: MIT Press.JUNGHANNS, Uwe & Gerhild ZYBATOW (1997), Syntax and Information Structure in Russian Clauses, in:
Wayles Browne (eds. et al.), Formal Approaches to Slavic Linguistics: The Cornell Meeting, 1995. Ann Arbor: Michigan Slavic Publications, 289-319.
Israeli, Alina (1997), Semantics and Pragmatics of the “Reflexive” Verbs in Russian. MQnchen: Otto Sagner.
KING, Tracy Holloway (1994), The UTAH and Causation in Russian Psych Verbs, in: Peter Ackema & Maaike Schoorlemmer (eds.). Proceedings of the First Conference of the Student Organization of Linguistics in Europe. (= SOLE Publications, 1.) The Hague: Holland Academic Graphics, 115-30.-
LA VINE, James (1997), A Lexicalist Perspective of Valency Changing Operations in Russian and Czech, in: Prague Bulletin of Mathematical Linguistics 68: 5-34.
LEGENDRE, Geraldine & Tanya Akjmova (1994), Inversion and Antipassive in Russian, in: Sergey AVRUTIN (eds. et al.), Formal Approaches to Slavic Linguistics: The MIT Meeting, 1993. Ann Arbor: Michigan Slavic Publications, 286-318.
PESETSKY, David (1995), Zero Syntax. Cambridge, MA: MIT Press.SCATTON, Ernest A. (1984), A Reference Grammar of Modem Bulgarian. Columbus, OH: Slavica.SLABAKOVA, Roumyana (1996), Bulgarian Psych Verbs, in: Jindfiich TOMANN, (ed.), Formal Approaches
to Slavic Linguistics: The College Park Meeting, 1994. Ann Arbor: Michigan Slavic Publications, 249-72.
STEPANOV, Arthur (to appear), On wh-Fronting in Russian. In Proceedings of the North Eastern Linguistic Society 28. Amherst, MA: Graduate Linguistic Student Association, University of Massachusetts.
TlMBERLAKE, Alan (1975/1986), Hierarchies in the Genitive of Negation, in: Richard D. BRECHT & LEVINE, James (eds.) Case in Slavic. Columbus: Slavica, 338-60. [Originally in Slavic and East European Journal 19: 123-38.]