Case Checking versus Case Assignment and the Case of Adverbial NPs Author(s): eljko Bokovi Reviewed work(s):

Source: Linguistic Inquiry, Vol. 37, No. 3 (Summer, 2006), pp. 522-533 Published by: The MIT Press Stable URL: http://www.jstor.org/stable/4179380 . Accessed: 30/07/2012 14:12
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Linguistic Inquiry.

http://www.jstor.org

522
CASE CHECKING VERSUS CASE ASSIGNMENT AND THE CASE OF ADVERBIAL NPs ZeljkoBoskovic Universityof Connecticut

SQUIBS AND DISCUSSION

On the basis of genitive of quantification(GQ) in Slavic, I argue in this squib that the Case-checkingtheoryis empiricallysuperiorto the Case assignmenttheory.I also arguethatSlavic GQ providesevidence that adverbialNPs (Larson's (1985) bareNP adverbs)are not subject to the traditionalCase Filter, and I suggest that the Case of such Duringthe discussion, I also examine how CaseNPs is interpretable. markedmodifying adjectives and noninitial NP conjuncts are Caselicensed. The squib is organizedas follows. In section 1, I summarizethe basic propertiesof Slavic GQ. In section 2, I show that GQ can tease apartCase checking and Case assignment.In section 3, I discuss Case propertiesof adverbialNPs. 1 Introduction: Genitive of Quantification Examples (1)-(2) show what happenswhen a numeralNP occurs in a structural Case context in Russian.' masinu. (1) a. Ivan kupil odnu Ivan bought one(ACC) car(ACC.sG) b. *Ivan kupil odnu masiny. Ivan bought one(ACC) car(GEN.sG) (2) a. Ivan kupil pjat' masin. Ivan bought five cars(GEN.PL) b. *Ivan kupil pjat' masiny. Ivan bought five cars(ACC.PL) With 'one', both 'one' and the noun receive theirCase from the verb. However, with a higher numerallike 'five', the noun receives genitive.2 This pattem is traditionallyinterpretedas indicating that only higher numeralshave the ability to assign GQ. As for inherentCase contexts, when a numeralNP occurs as an object of an inherent-Caseassigning V, as in (3)-(4), both the noun and the numeral('one' as well as 'five') bear the inherentCase in question.3

I thankSteven Franks,SandraStjepanovi6,anonymousLinguisticInquiry reviewers,andparticipants of FormalDescriptionof Slavic Languages5 andmy seminarat the Universityof Connecticut for helpfulcommentsandsuggestions. 1Case informationis indicatedin the glosses only where it is relevant. 2 There is somewhat of a controversyregardingwhetherpjat' in (2a) is a syncreticnominative/accusative or a Caseless form (for relevantdiscussion, see Franks1994 and Boskovic, to appear,amongotherworks).Whatis important for our purposesis that the numeralalways takes this form when GQ is assigned.Note also thatI ignore the paucalnumerals'two', 'three',and 'four', because of the controversyover whetherthey patternwith 'one' or with 'five'. (CompareBabby 1987 and Franks 1994 with Rakhlin 2003. Rakhlin argues that what is standardlyanalyzed as genitive singular GQ with 'two', 'three', and 'four' is actually paucal nominative/accusative.) 3 All otherCase patterns for the numeralandthe following nounin (3)-(4) are unacceptable.

SQUIBS AND DISCUSSION

523

(3) Ivan vladeet odnoj fabrikoj. Ivan owns one(INsTR) factory(INsTR.sG) (4) a. Ivan vladeet pjat'ju fabrikami. Ivan owns five(INsTR) factories(INsTR.PL) b. *Ivan vladeet pjat' fabrik. Ivan owns five factories(GEN.PL) The descriptive generalizationis that GQ overrides structuralCase (2), but not inherentCase (4). There are many analyses of the basic GQ paradigmin the literature.Most authorsassumethatthe reasonwhy GQ overridesstructural like (2) is thatthe GQ assigner,the numeral, accusativein constructions is closer to the relevantnoun thanthe verb. The strategypursuedwith respect to (4) is to enforce the satisfaction of the Case propertyof the inherent-Case-assigning verb independentlyof the distancefactor. Freidin and Babby (1984) and Babby (1985) do this by appealing to the Principle of Lexical Satisfaction, which requires that lexical propertiesbe satisfied at all levels of representation. Assuming that instrumental Case assignmentis an idiosyncraticpropertyof the verb will that must be specified in the lexicon, it follows that instrumental takeprecedenceover any otherCase-licensingstrategyin (3)-(4) (see, however, Franks 1998, 2002 and Boskovic, to appear,for some arguments againstthis analysis). Franks(1994, 1995) presentsan analysis based on Chomsky's (1986) Case theory, in which structural Case is and inherentCase at D-Structure,and he proassigned at S-Structure poses that GQ is a structuralCase in Russian. 'Factory' then must bear instrumentalin (4) because the verb assigns its inherentinstrubefore the Case-assigningability of the Q mental Case at D-Structure is activated.On the otherhand,adoptingthe gist of Chomsky's (1986) theory of inherent Case (but not requiringthe assumption that DI argue elseStructureand S-Structureare levels of representation), where (Boskovic, to appear)that GQ cannot override inherentCase because, as Chomsky argues, inherentCase is associated with 0-role verbfails to check assignment;as a result,if an inherent-Case-marking its inherentCase againstits object, it will also fail to 0-markit.4 Under this analysis, (4b) ends up violating the 0-Criterion.0-considerations are irrelevantin a structural-Case-assigning context since structural Case is not associated with 0-role assignment in Chomsky's (1986) Case system. Forease of exposition,I will adoptthe analysisdeveloped in Boskovic, to appear,in the discussion below.5

4 The fact that the verb is not the closest potential Case licenser is then irrelevantbecause the more economical derivationdoes not converge owing to a 0-Criterion violation.Note thatin Boskovi6, to appear,I couch my analysis in terms of Case checking, while the other authorsmentioned above assume Case assignment. 5 I am ignoring many importantissues here regardingSlavic GQ since they are not directly relevantto the theoreticalpoints raised in sections 2-3.

524

SQUIBS AND DISCUSSION

Turningto Serbo-Croatian (SC), we find that GQ in SC patterns with GQ in Russian with respect to structuralCase environments. NP cannot However, SC differs from Russianin thata higher-numeral verb.6(Fromnow on, occur as an object of an inherent-Case-assigning NPs as QPs, following the standard I will refer to GQ higher-numeral assumptionthat QP dominatesNP in such structures.) (5) Ivan se zahvaljujejednom covjeku. Ivan REFL thanks one(DAT) man(DAT.sG) 'Ivan is thankingone man.' (6) a. *Ivan se zahvaljujepet ljudi. Ivan REFL thanks five people(GEN.PL) b. *Ivan se zahvaljujepet ljudima. five people(DAT.PL) Ivan REFL thanks 'Ivan is thankingfive people.' (7) On je ovladao jednom zemljom. he is conqueredone(INsTR) country(INsTR.sG) 'He conqueredone country.' (8) a. *On je ovladao pet zemalja. he is conqueredfive countries(GEN.PL) b. *On je ovladao pet zemljama. he is conqueredfive countries(INsTR.PL) of (6a) and (8a) is not surprisingand can be The ungrammaticality accountedfor on a parwith thatof Russian(4b). As for the impossibility of "repairing"these examples in SC (cf. (6b) and (8b)), note that, in contrastto higher numeralsin Russian, higher numeralsin SC do not decline-they arefrozenforms.7In Boskovic, to appear,I interpret this as indicatingthat they are Caseless (see also Franks 1975). Since the head of the numeralphrase, the numeral,is Caseless in SC, the verb cannot check its Case against it. A QP, viewed as an object of an inherent-Case-assigning verb in SC, then inevitably causes a 0Criterionviolation. 2 Case Checking versus Case Assignment Having discussed the basic propertiesof GQ, I now turn to a very interestingparadigminvolving the SC prepositions(a), also discussed
6 Inherent-Case-assigning verbs may assign dative, instrumental, or genitive. Genitive-assigningverbs, however, might be exceptional in this respect. With such verbs it is actually difficult to tell whetherthe verb or the numeral assigns genitive to the noun, so I will put them aside here. I am also putting aside contextsin which a numeralNP functionsas the complementof a preposition. Franks(1994, 1995) argues that the prepositionin such contexts behaves differentlyfrom the verb. (The analysis of (28) below can actuallybe extended to much of the P-complementevidence Franksexamines.) 7The non-GQ-marking numeral 'one', on the other hand, does decline (in fact, SC 'one' behaves like its Russiancounterpart in all relevantrespects). Below I will disregard'one'.

SQUIBS AND DISCUSSION

525

by Leko (1986), Franks(1995, 2002), andWechslerandZlatic (2003). Case; no Case but instruS(a) is a prepositionthatassigns instrumental mental is possible in (9).8 (9) s(a) njim with him(INsTR.sG) Bearing this in mind, consider (10)-(13). (10) On je ovladao zemljom. he is conqueredcountry(INsTR.sG) 'He conqueredthat country.' (11) *On je ovladao s(a) zemljom. he is conqueredwith country(INsTR.sG) (12) *On je ovladao pet zemalja. he is conqueredfive countries(GEN.PL) (13) On je ovladao s(a) pet zemalja. he is conqueredwith five countries(GEN.PL) 'He conqueredfive countries.' Apparently,sa can (and must) occur in the complement of ovladati 'to conquer' only when the complement contains a higher numeral. Recall now that,being an inherentCase assigner,ovladati must check its instrumentalCase against its object argument.(12) is ruled out because ovladati fails to do so. Regarding (11) and (13), I propose thatwe are dealing here with a last resortsa-insertionthat takes place so that ovladati can check its instrumentalCase against its object argument.The complement of ovladati in (13) is a PP, and ovladati checks its Case against the head of its PP complement, namely, sa (see also footnote 14 for an importantpoint). Given that sa-insertion is a last resort operation,since there is no need for it in (11) it is not allowed to take place. Alteruatively,(11) can be ruled out via Case theory: if ovladati and sa check Case against each other, the object NP cannot be Case-checked.(In other words, the Case featureof one of the relevant elements must remain unchecked.)What we have in (13), then, is a situation where a traditionalCase assigner (a verb) checks Case against a traditionalCase assigner (a preposition).This state of affairscan be easily accommodatedunderthe Case-checking theory, but not under the Case assignmenttheory. Under the former betweenthe elementsinvolved in a Case-licentheory,the relationship sing relationis inherentlysymmetrical.The distinctionbetweena Case assigner and a Case assignee cannot in fact be stated in naturalterms without additionalstipulationsin this theory. On the other hand, in the Case assignmenttheory, the relationshipin question is inherently asymmetrical(i.e., unidirectional),with one element (Case assigner)

8 Whether in thesquib s orsa is usedin theexamples discussed doesnot I will referto thepreposition as in question affecttheirgrammaticality status. sa.

526

SQUIBS AND DISCUSSION

essentially transferringits Case property to another element (Case in the Case assignassignee). Obviously,given the inherentasymmetry ment process, a Case assigner cannot assign Case to a Case assigner. This problemdoes not arisein the Case-checkingtheory,wherethereis no distinctionbetween Case assignersand Case assignees. I conclude, therefore, that the paradigmin (10)-(13) provides an argumentfor the superiorityof the Case-checkingtheoryover the Case assignment theory. I now turn to additional data regardingGQ as well as certain dataconcerningundeclinednouns thatconfirmthe currentanalysisby showing that sa-insertion with QPs is part of a broaderpatternfor resolving Case problems,more precisely, for Case-checkinginherent Case verbs. Considerthe following sentence:9 i jos pet (14) ?Oni su ovladali Andorom and anotherfive they are conqueredAndorra(INsTR) zemalja.
countries(GEN.PL)

'They conqueredAndorraand anotherfive countries.' (14) clearly contrastswith (12). Recall that (12) is ruled out because ovladati fails to check its instrumental Case owing to the inability of the object QP to bear Case. The object in (14) is a coordinated phrase, where only the second conjunct is a Caseless QP. The first conjunct is an NP bearinginstrumental Case, which is capableof checking the instrumentalCase of ovladati. Insertionof sa in (13) apparentlyhas the same rescuing effect on (12) as the presence of the instrumentalmarkedNP in (14). Since it seems quite clear that the instrumental NP in (14) Case-checksthe verbovladati, it seems naturalto conclude that sa in (13) performsthe same task. Interestingly,(15) contrastswith (14). (15) *Oni su ovladali Andorom.
Andorra(INSTR)

pet zemalja

they are conquered five countries(GEN.PL)and

One possibility is thatthe NP Andoromis too deeply embeddedwithin the coordinationto be accessible for Case checking by the verb. (For argumentsin favor of a hierarchicalstructureof coordination,see Munn 1993, Kayne 1994, and Johannessen1998, among others.)Note that if indeed the coordinationphraseis a phase (see Boskovic 2005), Chomsky's (2000) Phase Impenetrability Condition, which says that only the head andthe specifierof a phase can be accessed from outside the phase, would prevent the verb from entering into the required relation with Andorom. In this respect, notice example (16), where
9 SandraStjepanovic(pers. comm.) suggested checking the coordination patternsin (14), (15), (23), and (24).

SQUIBS AND DISCUSSION

527

the second conjunctis clearly not Case-checkedby the coordinationexternalCase checker I(nfl). (16) John and me left. What about examples like those in (17), then? (Note that the default Case in SC is nominative;see Boskovic 2005a.) (17) a. John and I left. b. On je kupio kucu i kola. he is bought house(Acc.sG) and car(ACC.SG) 'He bought a house and a car.' i c. Oni su ovladali Andorom and they are conqueredAndorra(INsTR) Luksemburgom.
Luxembourg(INSTR)

'They conqueredAndorraand Luxembourg.' of (15) One conclusion that can be drawngiven the ungrammaticality is that the second conjunct in (17) is not directly Case-licensed by V/I. Rather, it is Case-licensed via the first conjunct.10Under this Case-licensing strategy, the second conjunct can only get the same of (15) then straightCase as the first conjunct.The ungrammaticality forwardlyfollows given that QPs (i.e., the first conjunctin (15)) are Caseless. Alternatively,it is possible that the interveningQP in (15) prevents the verb from entering into an Agree relation with the second conjunct NP, which would have resulted in its Case checking, via Regarding Chomsky's (2000) Defective Intervention Constraint.11 (17), underthis analysis we may be dealing here with an application of Hiraiwa's (2001) Multiple Agree, where the outside Case checker (V/I) enters into a simultaneousAgree relation with both conjunct NPs. As discussed by Hiraiwa, this kind of simultaneousagreement voids potentialinterventioneffects, so that the first conjunctdoes not prevent the second conjunct from being Case-licensed by V/I. One way or another,it seems plausible that the contrast in (14)-(15) is another instantiationof well-known first-second conjunct asymmetries, which may be analyzableby appealingto the hierarchicalstructure of coordinationalong the lines suggested above; hence, it does not affect the point made above regarding(14). Notice also that (15) can be improved with sa-insertion, as expected given the above discussion of sa. (18) also confirms that the

10We would be dealing here with some kind of Case agreement.I leave its precise formal implementationfor anotheroccasion. 11The gist of defective interventionin this case is that although QP is itself Caseless, it still prevents the verb from entering into an Agree relation that would result in Case checking with the second conjunctNP.

528

SQUIBS AND DISCUSSION

problem with (15) is that ovladati is unable to undergo Case checking.12 (18) ?Oni su ovladali Andorom.
Andorra(INSTR)

s(a) pet zemalja

they are conquered with five countries(GEN.PL)and

Data regardingundeclinednouns are also relevanthere. Some loaned female names that do not end in a do not decline in SC. Thus, while Nada declines, Meri does not; it has only one form. b. Meri (19) a. Nada Nominative:Nad-a Accusative: Nad-u Genitive: Nad-e Dative/Locative:Nad-i Nad-om Instrumental: Vocative: Nad-o Let us assume thatundeclinednames are actuallyCaseless. From this perspective, the patternin (20)-(21), noted by Wechsler and Zlatic (2003), is not surprising. je kupio Meri. (20) Uzgajac konja breeder horses(GEN) is bought Meri 'The horse breederbought Meri.' (21) *Dzokej je ovladao Meri. jockey is conqueredMeri 'The jockey conqueredMeri.' Like Caseless QPs, undeclinednouns can functionas objects of verbs that normallyassign accusative Case, but not as objects of verbs that Case. This can be straightforwardly assign instrumental accountedfor if such nouns are Caseless, given that only inherent-Case-assigning verbs must assign their Case for reasons discussed above. Significantly, undeclined nouns also patternwith QPs with respect to both sa-insertionand the coordinationtest performedabove.

12 (i) appearsto contradictthe last resortnatureof sa-insertion,given the grammaticality of (14).

i sa jog (i) Oni su ovladali Andorom pet and with anotherfive they are conqueredAndorra(INsTR) zemalja.
countries(GEN.PL)

However, there is anotherderivationfor (i), which refutes this conclusion. As shownin Bogkovi61997, 2001, SC participlesmay undergoshortV-movement. (i) can then involve VP coordination,with the participleundergoingacrossthe-boardmovementfrom each conjunct.On this derivation,ovladati takes the QP as its complementin the second conjunct,just as it does in (13).

SQUIBS

AND

DISCUSSION

529

(22) Dzokej je ovladao s(a) Meri. jockey is conqueredwith Meri 'The jockey conqueredMeri.' i Meri. (23) ?Dzokejje ovladao Kraljicom jockey is conqueredQueen(INsTR.sG) and Meri 'The jockey conqueredQueen and Meri.' Given the above discussion, the contrastbetween (21) and (22)-(23) accountedfor. While in (21) ovladati fails to can be straightforwardly check its instrumental Case, in (22)-(23) it does not:in (23), Kraljicom Case and in (22), sa does. checks its instrumental As in the case of QPs, a first-second conjunct asymmetryis at work here. Thus, (24) contrastswith (23). Furthermore, as with QPs, (24) can be improvedwith sa-insertion. (24) *Dzokej je ovladao Meri i Kraljicom. (25) ?DMokej je ovladao s(a) Meri i Kraljicom. jockey is conqueredwith Meri and Queen(INsTR.sG) There is anotherway of rescuing (21), noted by Wechsler and Zlatic (2003). The example becomes acceptable if the undeclined noun is modified by an adjective bearing instrumentalCase. (Note that possessives are morphologicallyand syntacticallyadjectives in SC; see Zlati6 1997 and Bogkovi6 2005a. I will not distinguishbetween the two below.) (26) Dzokej je pokusao ovladati jockey is tried to-conquer nasom/neukrotivom
our(INSTR.SG)/untamable(INSTR.SG)

Meri.
Meri

'The jockey wanted to conquerour/untamable Meri.' We are dealing here with the same type of rescuing effect as the one illustratedin (22)-(23): the presence of an instrumental-marked adjective makes it possible for the verb to check its instrumental 13 The example in questionalso bears on the issue of how adjecCase. It is often assumedthatCase-marked tives areCase-marked. adjectives get theirCase throughagreementwith the noun,which is Case-licensed from outside the NP. Given the above discussion, (26) provides evidence againstthis assumption.Since the nounis Caseless, the adjective in (26) cannot get Case throughagreementwith the noun. Rather,it must be directly Case-checkedby the verb. To summarizethe discussion in this section: The data regarding sa-insertion in SC provide evidence for the superiorityof the Casecheckingtheoryover the Case assignmenttheory.It can also be argued that a Case-markedmodifying adjective is Case-checkeddirectly by

13 Note thatthis type of example cannotbe constructedwith QPs because adjectives can only bear genitive in GQ environmentsin SC.

530

SQUIBS AND DISCUSSION

the NP-externalCase checkerrather thangettingits Case fromthe noun it modifies. Additionally,we have seen thatwhen two coordinated NPs bear the same Case, which they appearto get from an outside Case checker,eitherthe first NP gets its Case fromthe outside Case checker and the second NP is Case-licensed via the first NP, or both NPs are Case-licensed throughan applicationof Multiple Agree. 3 Case of Bare NP Adverbials Interestingly,Franks(2002) notes that sa-insertioncannot occur with instrumental adjuncts. (27) a. Trcaoje iumom. run is forest(INsTR.sG) 'He ran througha/the forest.' b. *Trcaoje pet suma. run is five forests(GEN.PL) 'He ran throughfive forests.' c. *Trcaoje s(a) pet suma. in (27a) is not assigned We have evidence here that the instrumental by a null head. If that were the case, (27c) could be analyzedon a par with (13), with sa insertedto check the instrumental Case of the null head.'4 Rather,I suggest that adjunctNPs are not subject to the Case Filter;that is, they are not Case-checkedlike argumentNPs (see also Babby 1986, 1994). Instead,the function of the instrumental Case in (27a), which is not checked/assignedby anything, is to identify the precise semantic role of the adverbial (see Larson 1985 regarding adverbial0-roles), which sa cannotdo. On the otherhand, because of its semantics, kroz in (28) can do that. (28) Trcaoje kroz pet suma. run is throughfive forests(GEN.PL) 'He ran throughfive forests.' Underthis analysis,the inherentCase of adjunctsis semanticallyinterCase Filter (see Chomsky pretable,hence not subjectto the traditional 1995 for the proposal that interpretablefeatures do not need to be checked (as long as they are not strong)):its functionis to identify the precise semantic role of the adjunct,with differentCases identifying different semantic roles. In this respect, note the paradigm in of the adverbialNP de(29)-(3 1), where the precise interpretation pends on the Case it bears.'5
14 As notedby Franks(2002), the ungrammaticality of (27c), also provides evidence thatsa in the constructions underconsideration shouldnot be analyzed as a mere realization of instrumentalCase, that is, as an instrumentalCase marker. '5 Particularly notoriousin this respectis Hungarian, which, for example, has ten Cases for denotingdifferentlocationalrelations.Thus,asztal-on ('table' in superessive Case) means 'on table', asztal-rol ('table' in delative Case) means 'off table', and so on (see Babby 1994).

SQUIBS AND DISCUSSION

531

(29) On ide. he walks 'He is walking.' (30) On ide Ivanu. he walks Ivan(DAT) 'He is walking towardIvan.' (31) On ide ulicom. he walks street(INSTR.sG) 'He is walking down the street.' SC prepositions can also perform the task in question. When they functionas adjunctsemanticrole identifiers,they are not Case checkinherentCase markers.16 ers; they are essentially interpretable We have seen that prepositionsexhibit varied behaviorwith respect to Case properties.In fact, on the basis of the data discussed in this squib,we can distinguishthreetypes of prepositionsin this respect (the typology is not intendedto be exhaustive): * prepositionsthat take an argumentthat they Case-check (see (9) and (i) in footnote 16); * prepositions that Case-check traditionalCase assigners (i.e., that Case-check a head that c-commandsthe PP in question; see (13), (18), (22), and (25)); * prepositions that function as interpretable Case markersand do not undergoCase checking (see (28)). Whatbehaviora particular prepositionwill exhibit with respectto the above typology depends on its semantic (includingthematic)as well as Case properties. To summarize,I have arguedthat the genitive of quantification providesempiricalevidence for the superiorityof Case checking over Case assignment.I have also proposedthatbareNP adverbialsare not subjectto the traditional Case Filter (i.e., in termsof checking theory, theirCase does not have to be checked)becausetheirCase is interpretable. Prepositionsoccurringwith adverbialNPs can in some contexts performthe task of interpretable Case features.Finally, there is some evidence thata Case-marked modifying adjectiveis Case-checkeddirectly by the NP-externalCase checker (it does not get its Case from the noun it modifies) and that when two coordinatedNPs bear the same Case, which they appearto get from an outside Case checker, either the first NP gets its Case from the outside Case checker and the second NP is Case-licensedvia the first NP, or both NPs are Caselicensed throughan applicationof Multiple Agree.
16 Notethat when doesassign kroz Case(i.e.,when it functions asa regular rather than Case-checking preposition, aninterpretable Casemarker), it assigns accusative. Caseformsarenotpossible (Other in (i).) (i) kroz Francusku through France(Acc)

532

SQUIBS

AND

DISCUSSION

References and the DirectCase Babby,LeonardH. 1985. Prepositional quantifiers Condition.In Issues in Russianmorphosyntax, ed. by Michael S. Flier and Richard D. Brecht, 91-117. Columbus, Ohio: Slavica. Babby, LeonardH. 1986. The locus of Case assignmentand the direction of percolation:Case theoryandRussian.In Case in Slavic, ed. by RichardD. Brecht and James S. Levine, 170-219. Columbus, Ohio: Slavica. Babby, Leonard H. 1987. Case, prequantifiers,and discontinuous agreementin Russian.Natural Language & LinguisticTheory 5:91-138. Babby, LeonardH. 1994. Case theory. In Noam Chomsky:Critical assessments,ed. by CarlosP. Otero,630-652. London:Routledge. An Boskovic, Zeljko. 1997. Thesyntaxof nonfinitecomplementation: economy approach. Cambridge,Mass.: MIT Press. interBoskovi6, Zeljko. 2001. On the nature of the syntax-phonology face: Cliticizationand relatedphenomena.Amsterdam:Elsevier Science. Boskovi6, Zeljko. 2005a. On the locality of left branchextractionand the structureof NP. Studia Linguistica59:1-45. Boskovi6,Zeljko.2005b. On the localityof Move andAgree:Eliminating the Activation Condition, GeneralizedEPP, Inverse Case Condition.UConn OccaFilter, and the Phase-Impenetrability sional Papersin Linguistics3. Cambridge, Mass.:MIT,Department of Linguistics and Philosophy, MITWPL. Boskovic, Zeljko. To appear. A minimalist account of genitive of In Proceedings of FormalDescriptionof Slavic quantification. Languages 5. Chomsky, Noam. 1986. Knowledge of language: Its nature, origin, and use. New York: Praeger. Chomsky,Noam. 1995. TheMinimalistProgram. Cambridge,Mass.: MIT Press. Chomsky, Noam. 2000. Minimalistinquiries.In Step by step: Essays on minimalistsyntaxin honor of HowardLasnik,ed. by Roger Martin,David Michaels, and JuanUriagereka,89-155. Cambridge, Mass.: MIT Press. Franks, Steven. 1994. Parametricproperties of numeral phrases in Slavic. Natural Language & Linguistic Theory 12:570-649. Franks,Steven. 1995. Parametersof Slavic morphosyntax. New York: Oxford University Press. Franks,Steven. 1998. Parameters of Slavic morphosyntax revisited: A minimalist retrospective.In Formal Approaches to Slavic Linguistics: The Connecticut Meeting, 1997, ed. by Zeljko Boskovic, Steven Franks,and William Snyder, 134-165. Ann Arbor, Mich.: Michigan Slavic Publications.

SQUIBS

AND

DISCUSSION

533

Franks, Steven. 2002. A Jakobsonianfeature based analysis of the Slavic numericquantifiergenitive. Journal of Slavic Linguistics 10:141-181. Freidin, Robert, and LeonardH. Babby. 1984. On the interactionof lexical and syntactic properties:Case structurein Russian. In Cornell workingpapers in linguistics VI, ed. by Wayne Harbert, 71-104. Ithaca,N.Y.: Cornell University, CLC Publications. Hiraiwa, Ken. 2001. Multiple Agree and the Defective Intervention Constraintin Japanese.In Proceedings of HUMIT 2000, ed. by Ora Matushanskyet al., 67-80. MIT Working Papers in Linguistics 40. Cambridge,Mass.: MIT, Departmentof Linguistics and Philosophy, MITWPL. Johannessen,Janne Bondi. 1998. Coordination.New York: Oxford University Press. Mass.: of syntax.Cambridge, Kayne,Richard.1994. Theantisymmetry MIT Press. Larson,Richard. 1985. Bare-NP adverbs.LinguisticInquiry 16:595621. in Serbo-CroLeko, Nedzad. 1986. Syntax of noun headed structures in English.Doctoral atianandcorresponding phrasalstructures dissertation,IndianaUniversity, Bloomington. Munn, Alan. 1993. Topics in the syntax and semantics of coordinate Doctoraldissertation,Universityof Maryland,Colstructures. lege Park. Rakhlin,Natalia.2003. A case againstCase conflicts. Paperpresented at Formal Approachesto Slavic Linguistics 12, University of Ottawa,May. Wechsler, Stephen,and LarisaZlatic. 2003. Themanyfaces of agreement. Stanford,Calif.: CSLI Publications. of the Serbiannounphrase.Doctoral Zlatic,Larisa.1997. The structure dissertation,University of Texas, Austin.