In contrast, I claim there is no reasonable analysis of these facts under a pure syntax approach. Given that clitics contrast with non- clitics in their placement options, a pure syntax approach must pos- it two different kinds of syntactic movement for the two cases and explain why they correlate with different kinds of intervening ma- terial.
Getting this contrast requires an arbitrary stipulation under any pure syntax ac- count of the PP paradigm, because of the basic descriptive fact that clitics go where nothing else can: to accomplish this in syntax re- quires a type of movement for which there can in principle be no independent motivation. It ought to be possible to construct the same kind of argument based on other constructions in SC as well.
I have come across two paradigms that might serve as a starting point: The first involves a modified predicate adjective phrase: 31a vs. If it is re- placed with a wh-word that can extract, other material can more easily intervene: 32b , Vrlo je visoka Bojanova sestra. Vrlo je Bojanova sestra visoka. Koliko tvrdis da je visoka Bojanova sestra? Koliko tvrdis da je Bojanova sestra visoka? Koliko je Bojanova sestra visoka?
The second involves conjoined NPs with pre-nominal modifiers: 34 a. Tvoja su ti ga mama i tvoja sestra kupile. In this section the focus is on predicative constructions, which have been claimed to disallow 1C placement and require 1W placement. Here the alternative of putting them after a whole phrase is not open…In this position when an adjec- tive modifies a noun, or an adverb modifies an adjective, the two together form a phrase, and the enclitics, again, must come after the first word of the phrase. Browne 35 a. Jako si mi dosadan.
Jako mi je dosadna njegova posljednja knjiga. U drugoj su sobi.
Serbo-Croatian clitic placement: An argument for prosodic movement
Under my theory, we have to say that the adjective phrase in 37 cannot front ahead of the clitics in the syntax; in particular, it cannot front to Spec-CP. If the Adjective Phrase always follows the clitics syntactically, perhaps sitting in a Focus position between Comp and IP, the clitics must move right- ward in the phonology to derive a valid sentence, and since I have claimed that they never move more than one PWd in the phonolo- gy, the ungrammaticality of 37b would be explained.
A pure syntax ac- count would be hard-pressed to explain why part of a copular predicate can front but the whole predicate cannot. In all cases, the variant with the clitic following the entire initial NP is fine. Under such an account, clitics can only appear within an NP if the part that precedes them is syntactically extractable.
Thus, she claims this fails to be the case in 39 — 42 : at least according to her intuitions, none of these elements independently allows extrac- tion. The data for one of the fortress types is given in Extractability could be the reason why people who allow clitics in fortresses do so, but we must still explain what blocks them for those who do not allow them.
Therefore, Halpern at- tempts to account for the degraded nature of these sentences pro- sodically. Specifically, it would have to be that these constructions have a different prosodic structure from good cases of interrupted constituents, and that this difference blocks the operation of PI or subsequent cliticization. We want a constraint that rules out the structures in 39 — 42 , and rules in clear cases of PI, discussed in sections 3 and 4. Halpern proposes the phrasing principle in 44 : 44 The left edge of the head of a branching constituent corre- sponds to the left edge of a prosodic [phonological] phrase.
Halpern 96 plus the constraint that PI cannot cross a phonological-phrase boundary. A clitic that originates in Comp, to the left of this NP at S-structure, would then be outside that phono- logical-phrase after prosodic mapping, and PI would require it to cross that phrase edge if it were to invert with and cliticize to pri- jatelji, which Halpern disallows. For one thing, examples like 46 are derivable without PI anyway, so this exam- ple is actually irrelevant to the proposed constraint.
It is also not at all clear that his proposal will extend to cover the various other types of fortresses while allowing cases like 27 a PP with modi- fiers and 37a a predicative AP above, and there are other prob- lems. Nonetheless, I believe that something along these lines is right. A possible generalization is that PI cannot move clitics across the head noun of an NP, regardless of branching.
This would at least unite 40 — 42 ; something special would have to be said about proper names like To the extent that we can find a natural prosodic constraint on PI, this supports the mixed approach to clitic placement if an alter- native syntactic constraint would be unappealing or unstatable.
One intriguing fact that supports this reasoning is the following, noted by Percus : postnominal PP fortresses become better when the PP portion is made heavier—compare 48 with 42 above. Getting these facts, if they turn out to be fully general across fortress types, evidently requires a more complex constraint on PI than the ones I have considered.
Perhaps the first noun of the NP likes to phrase with following material, but cannot do so if that material is set off due to heaviness. This in turn could be because phonological phrases prefer to be binary branch- ing Dresher Whatever the explanation, the fact that the cru- cial contrasts come from presumably identical structures that differ only in heaviness or pause strongly supports the idea that the con- straint must be a prosodically-based one.
I have shown that his proposals can be extended to cover a substantially wider range of facts in SC than he or others have discussed.
Serbo-Croatian Second Position Clitic Placement and the Phonology-Syntax Interface (1994)
The notion that clitics can be re-ordered with respect to an ad- jacent word in the way proposed by Halpern is key to understand- ing constraints on clitic placement. We have seen considerable evi- dence that this is a phonological process. One would obviously like to study other instances of prosodic movement to see what generalizations can be made about it. An important implication of this study bears on the nature of the phonology-syntax interface more generally.
The facts of SC were used by Zec and Inkelas to support their view of this interface as a co-present, non- derivational one. In this paper I have striven to clarify the role that the phonology plays in this system. References Bennett, David C. Word-order change in progress: the case of Slovene and Serbo-Croat and its relevance for Germanic. Journal of Linguistics 23 2 — Browne, Wayles. On the problem of enclitic placement in Serbo-Croatian.
In Slavic transformational syntax, Michigan Slavic Materials 10, ed. Richard D. Brecht and Catherine V. Chvany, 36— Serbo-Croatian enclitics for English- speaking learners. Long head movement?
Verb- movement and cliticization in Croatian. In Sprachwissenschaft in Frankfurt Arbeitspapier 7. Verbal and pronominal clitics in Croatian. Paper presented at the Geneva Workshop on the Wackernagel Position.
Dresher, Bezalel Elan. The prosodic basis of the Tiberian Hebrew system of accents. Language 70 1 :1— Contrastive analysis of English and Serbo-Croatian. Volume One.
- The Rise and Fall of Second-position Clitics | SpringerLink!
- Sorria, você está praticando inglês!.
- Adventures in Birdland.
Halpern, Aaron Lars. Topics in the placement and morphol- ogy of clitics. Dissertation, Stanford University.
- Publisher Description!
- Clitic Placement in South Slavic.?
- Papers selected for presentation.
Hock, Hans Heinrich. Syntax vs. Studies in the Linguistic Sciences 22 1 — Syntactic vs. Clitics, morphological merger, and the map- ping to phonological structure. In Theoretical morphology: Ap- proaches in modern linguistics, ed. San Diego: Academic Press. Marantz, Alec. Clitics and phrase structure. In Alternative conceptions of phrase structure, ed. Mark R. Kroch, 99— Chicago: University of Chicago Press.
Slavic versus Balkan: An analysis of the Balkan Slavic major clitic clusters. Percus, Orin. The captious clitic: Problems in Serbo-Croat- ian clitic placement. Phonology generals paper, MIT. Progovac, Ljiljana.
Serbo-Croat-Bosnian clitics and Word Grammar * „AMDŽI‚ and RICHARD HUDSON
Paper presented at the LSA, Boston. The grammar of Serbo-Croatian clitics: A synchronic and diachronic perspective. Disserta- tion, University of Illinois at Urbana-Champaign.