GB and MP treatments of PVCs

Let me start this discussion with a quote from É. Kiss (1994a), which is very important and informative for the following reasons.³ (i) It appropriately demonstrates “the verbal prefix problem”. (ii) It firmly argues for the syntactic analysis of these particles. (iii) It draws a

1 It is generally assumed that they (just like other VM types) make up complex predicates with the lexical verbs they combine with. Their combinations with lexical verbs also have a frequently used term (particular to them): particle verb constructions (PVCs). In what follows I will also use this term to separate this type from the rest of VM + V combinations.

2 Also see Booij & Jaap (2003).

3 Most of these empirical generalizations go back as far as Soltész (1959).

parallel between idiom chunks and their treatment in idiomatic expressions, on the one hand, and particles and their treatment in PVCs, on the other hand.⁴ (iv) It can be considered to be a classic epitome of the most crucial shared aspects of the assumptions underlying the sweeping majority of GB and MP accounts of these phenomena.⁵ (v) At the same time, part of the argumentation in this text also serves as a trigger for strong critical remarks.

The clearest type of incorporated constituents, whose marked syntactic and semantic properties are the easiest to identify, is the verbal prefix, traditionally analyzed as a kind of Adv. Its main characteristics could be summarized as follows:

(1) The verbal prefix and the V selecting it form a single lexical unit; this unit serves as input to lexical word formation processes. Example: fel-tesz ‘up-put [assume]’; feltétel ‘assumption’.

(2) The verbal prefix and the V form a single semantic unit. Their meaning is often noncompositional, as in be-rúg ‘get drunk’ (lit. ‘in-kick’).

(3) If the verbal prefix immediately precedes the V, they form a sing phonological unit.

(4) The verbal prefix is syntactically free of the V; it moves separately.

(KÉK98) Feli kell, hogy hív-ja-m ti. up needs that call-SUBJUNC-I(.him)

‘It is necessary that I call him up.’

The question is whether the syntactic representation of prefixed verbs should be based on properties (1)-(3), which suggest their unity, or it should be based on property (4), which suggests that the prefix and the V are two separate constituents. Actually, property (3) does not seem to be crucial; any constituent moved into the immediately preverbal Spec,VP position forms a singe phonological unit with the V; that is, whether the prefix and the V should be one or two syntactic constituents, property (3) will fall out. As for properties (1), (2), and (4), they co-occur in the case of idioms, as well.

(KÉK99) a. Mari-t be-húzt-ák a cső-be.

Mary-ACC in-pulled-they the tube-into

‘They pulled Mary into the tube. [They tricked Mary.]’

or:

b. Mari-t cső-be húzt-ák.

Mary-ACC tube-into pulled-they

‘They pulled Mary into the tube. [They tricked Mary.]’

Behúz a csőbe or csőbe húz constitute a single lexical unit; the csőbe húz variant can even be nominalized.

(KÉK100) A csőbehúzás nem sikerült.

the pulling_into_tube not succeeded

‘The tricking did not succeed.’

It forms a single semantic unit, too; its meaning is noncompositional. At the same time, the parts of the idiom are syntactically independent; they need not be adjacent.

4 In Section 3.2 I will elaborate on this parallel from the perspective of the generalized treatment of VMs.

5 I change her example numbers in the following way: (98)  (KÉK98). Thus, the original ones can be identified, but they do not interfere with my example numbers in this chapter. Where necessary, in the glosses I change her hyphens to the more appropriate underscores, as in: pulling-into-tube  pulling_into_tube.

(KÉK101) Be akar-ják húz-ni Mari-t a cső-be.

in want-they pull-INFIN Mary-ACC the tube-into

‘They want to pull Mary into the tube.’

In spite of their lexical and semantic unity, idioms have never been analyzed as syntactic units dominated by a single V or V’ node. I shall adopt a similar treatment for the verbal prefix, too; that is, I shall assume that the verbal prefix is an idiomatically selected complement of the V. It has no role; it is not an argumental but a predicative complement. Its role is similar to the role of the adverb in the phrase behave well.

I assume that in D-structure, the prefix is one of the XPs generated postverbally as sisters to the V and to each other. (It is of the category [AdvP [ Adv]].) This accounts for why its postverbal order is basically free – although recall that unstressed, clitic-like elements are more felicitous if they are ordered immediately after the V in V’.

(KÉK102) a. Imre tavaly ismerkedett meg Erzsi-vel.

Imre last_year got_acquainted PREV Lisa-with ‘Imre got acquainted with Lisa last year.’

b. ?Imre tavaly ismerkedett Erzsi-vel meg.

c. Imre tavaly ismerkedett vele meg.

Imre last_year got_acquainted with_her PREV

‘Imre got acquainted with her last year.’

If the verbal prefix is one of the major postverbal constituents, it is a potential target of Focus Movement; hence the complementary distribution of the prefix and a focused XP in preverbal position immediately falls out. If the prefix is moved into Spec,VP, then there is naturally no room there for another focus, too.⁶

(KÉK103) a. [TP Imrei [VP Erzsi-velj [_V’ ismerkedett meg ti tj]]]

Imre Lisa-with got_acquainted PREV

‘It was Lisa who Imre got acquainted with.’

b. [TP Imrei [VP megj [_V’ ismerkedett Erzsivel ti tj]]]

‘Imre got acquainted with Lisa.’

It might seem implausible at first sight to assign parallel analyses to the sentences in (KÉK103a) and (KÉK103b), as they have very different interpretations: (KÉK103a) involves a focus operator; (KÉK103b), on the other hand, is a neutral sentence. In the Spec,VP of (KÉK103a) Erzsivel expresses identification with exclusion, while in the Spec,VP of (KÉK103b) meg merely bears an emphasis, which we understand as the receive the [+F] feature from the verb in Spec,VP) and there is also a verbal prefix (or any other VM) in the postverbal domain then it is the former that will be moved into the preverbal position, and the “default” VM type focus movement is blocked. In addition, notice that if É. Kiss assumes that both an ordinary constituent to be focused and a VM are potential foci then she would need an explicit (perhaps OT style) rule to ensure that the presence and movement of the former blocks the movement of the latter. (Compare the two views above in this context.) At several points in this dissertation I claim that it is inappropriate to assume that VMs in Spec,VP are always and necessarily focused constituents. É. Kiss herself changed her (1992, 1994a) view and as of É. Kiss (2002) at least she has assumed that VMs and foci do not occupy the same preverbal position. For details, see Section 2.1.1 in Chapter 2, the subsequent part of the quote from É.

Kiss (1994a) and my remarks on it in the next footnote.

focus outlined in Sections (KÉK)5.1 and (KÉK)5.2.⁷ Recall that an XP in Spec,VP receiving the feature [+F] expresses identification with exclusion only if it denotes an entity. This is what happens to Erzsivel in (KÉK103a). If the feature [+F] percolates onto an XP which is not referential, the feature [+F] will only express identification. […]

Since a verbal prefix is not an independent semantic entity, the identifying role assigned to it via the feature [+F] is associated with the whole Prev + V unit.

Similar to other adverbials in Spec,VP, a verbal prefix can actually act as an operator expressing identification with exclusion if it is set into an explicit contrast, that is, if the context provides a set of two (or more) contrasted elements, one of which can be identified through the exclusion of the other. This is only possible in the case of verbal prefixes which express direction.

(KÉK104) János nem [VP KI [V’ szaladt]], hanem [VP BE [V’ szaladt]]

John not out ran but in ran

‘Jon did not ran OUT, but he ran IN.’

(É. Kiss 1994a: 42-44) From the perspective of this dissertation, É. Kiss’ (1994b) discussion of foci in Spec,VP is also very important. She uses the following two examples (1994b: 132).⁸

(1) a. [VP JÁNOS [V’ ette meg a süteményt]]

John ate PERF the cookie

‘JOHN ate the cookie.’

b. [VP Egy ’autó [V’ állt meg a ház előtt]]

a car stopped PERF the house in-front-of

‘A car stopped in front of the house.’

7 TL’s remark: the essence of the assumptions in those sections (KÉK5.1 and KÉK5.2) relevant to our discussion here is that É. Kiss assumes a single [+F] feature (thus, my remarks in the previous footnote are still valid), and it expresses either identification with exclusion or identification. In the latter case, this identification percolates up to the entire VP. I think the greatest problem with this approach is that a prev + V complex can systematically behave in two different ways, compare (i) and (ii). Meg ‘PERF’is a perfectivizing verbal prefix. In (i), it does not receive any special (focal) stress, it simply makes up a phonological word with the verb, and the entire sentence has the regular, neutral sentence level wanted to assume that (i) was also an instance of VP/verum (identificational) focusing, then the challenge would be to capture the obvious differences between the two “ID-foci” in (i) and (ii). On the basis of these considerations, it seems understandable why later GB/MP approaches (including those developed by É. Kiss) account for the preverbal complementarity of VMs and foci by postulating different syntactic positions, and by capitalizing on the aspect encoding and/or complex predicate forming potential of VMs. By contrast, one of my main claims in this dissertation is that LFG’s architecture and assumptions make it possible to capture this VM vs. focus complementary by postulating a single designated preverbal position (in the spirit of the what-you-see-is-what-you-get principle).

(i) János meg érkez-ett Debrecen-be.

John.NOM PERF arrive-PAST.3SG Debrecen-into

‘John arrived in Debrecen.’

(ii) János MEG érkez-ett Debrecen-be.

John.NOM PERF arrive-PAST.3SG Debrecen-into

‘John DID arrive in Debrecen.’

8 I keep the format, the glossing and the translations of these examples intact. Her example number is (59).

É. Kiss makes the following observations.

Whereas (1a) can only be used as an answer to the question Who ate the cookie?, (1b) can also answer the question What happened?. While (1a) expresses identification with exclusion, (1b) expresses identification only. (In fact, (1b) is ambiguous: it could also be an answer to What stopped in front of our house?, or its focus could be set into a contrast; that is, it is also capable of expressing identification with exclusion.) […] The focus of (1a) is interpreted contrastively because it is assumed that the situation described in the sentence involves a closed set of persons who were in the position of being capable of eating the given cookie. […] In the case of (1b), it is very likely that there is no closed set of relevant entities in the domain of discourse that could have performed the act of stopping in front of the house. The set being open […], the identification operation performed by the focus operator does not go together with an exclusion operation; so no contrast is implied (1994b: 132-133).

On the basis of the foregoing discussion of É. Kiss (1994a, 1994b) the following “focusing picture” emerges in her 1994 approach:

(a) identification of the constituent in Spec,VP with exclusion: (KÉK103a) and (1a), (b) identification of the VP: (KÉK103b),

(c) identification of the constituent in Spec,VP without exclusion: (1b).

Later (a) was separated from the rest and became the standard id/exhaustive focus type. As I pointed out above, (b) is actually Janus faced: (i) when there is no focus stress, this is an instance of ordinary VM + V combination, and a syntactically and semantically different (nonfocused) analysis was developed along the aspectual/complex-predicate lines; (ii) when there is focus stress, we are dealing with VP-focusing. The (c) type does not seem to have received due theoretical attention. Prosodically and semantically it seems to manifest a clearly distinct focus type, which can be taken to be presentational focus, and it can be regarded as different from “real” id-focus without exclusion. It can be argued that this construction type is the last (and “weakest”) type on the following focus scale:⁹

(2) (A) id-focus with exclusion > (B) id-focus without exclusion > (C) presentational focus As regards the categorial and phrase structural status of the verbal prefix, in comparison with other VM types, É. Kiss (2002) offers the following detailed and classic discussion. She starts out by giving the following examples¹⁰ illustrating the syntactic independence of the verbal prefix. Although in neutral sentences it immediately precedes the finite verb: (3), under clearly definable circumstances they are separated. In the presence of a focused constituent or in the case of clausal negation, the preverb has to follow the verb, see (4a) and (4b), respectively. They can also be separated in the unmarked order by a particle, see (5).

Moreover, the verbal prefix can move further away from the verb: it can occupy a contrastive topic position: (6a), and it can even be raised into a superordinate clause: (6b-c).

(3) János fel-olvasta a verseit.

John up-read his poems 'John read out his poems.'

9 It can be claimed that (B) requires situational/contextual support, and it is the type dubbed “hocus” by Kálmán et al. (1986), while (C) can be used in entirely out of the blue situations. For examples and discussion, see Chapter 7.

10 I have kept her examples, glosses and translations intact, except for the numbers of the examples. I have changed them in such a way that they should fit into the numbering of the examples in this chapter.

(4) a. János \tegnap olvasta fel a verseit.¹¹ John yesterday read up his poems

'It was yesterday that John read out his poems.' b. Péter nem olvasta őket fel.

Peter not read them up 'Peter did not read them out.'

(5) János fel akarta olvasni a verseit, és fel is olvasta őket.

John up wanted to.read his poems and up also read them 'John wanted to read out his poems, and out he read them.' (6) a. Fel csak János olvasta a verseit.

out only John read his poems 'Loudly, only John read his poems.' b. János fel szeretné olvasni a verseit.

John up would.like to.read his poems 'John would like to read out his poems.' c. Fel akarom, hogy olvasd a verseidet.

up I.want that you.read your poems 'I want that you should read out your poems.'

É. Kiss concludes that the prefix has to be inserted as an independent syntactic unit that is a lexically selected complement of the verb, and she adds that there are further questions as well.¹² (A) Should it be inserted preverbally or postverbally? (B) Should it be treated as a phrase or as a head? É. Kiss’ answers are as follows.

(A) It is an argument for the postverbal base generation of the prefix that its preverbal occurrence in constructions like (6b,c) is definitely not base generated, given that the prefix is not selected lexically or semantically by the verbs adjacent to it, so its position in (6b,c) is a derived position, the landing site of prefix movement. É. Kiss adds that verbal prefixes are a subtype of a larger class of elements, collectively called verb modifiers (VMs), which includes objects, oblique goal complements and nonagentive subjects, see (7).

(7) a. János újságot olvas.

John newspaper-ACC reads

'John is engaged in newspaper-reading.' b. János iskolába ment.

John to.school went 'John went to school.' c. Víz szivárog a falból.

water oozes the wall.from 'Water is oozing from the wall.'

11 The \ symbol indicates heavy (focus) stress.

12 In this connection, advocates of lexicalist approaches very often point out that the uniform, generalized assumption of a predicate-complement relationship between the verb and the particle becomes much less feasible if we take noncompositional PVCs into consideration (the more noncompositional, the less feasible). Below I will discuss what I believe an example of the most extreme case: meg-vet PERF-throw ‘despise’. Here the perfectivizing particle does not even perfectivize; thus, the meaning of this PVC is absolutely noncompositional.

É. Kiss claims that although the VMs in (7) are not referential, they bear case and they appear to be associated with theta roles; thus, they should be treated as arguments, and, just like ordinary arguments, they are best base generated postverbally. Verbal prefixes are adverbs and they can be analyzed as AdvPs that merely consist of a head. The basic question is what type of movement they are involved in: phrasal movement or head movement (in other words, whether their landing site is a specifier position or a head-adjoined position). É. Kiss gives the examples in (8) to show that this prefix movement can be nonlocal (also see 6b,c above), and she points out that head movement is generally assumed to be local; therefore, prefix movement can only be taken to be phrasal in these examples.

(8) a. János fel akarja olvasni a verseit.

John up wants read-INF his poems 'John wants to read out his poems.'

b. Jánosnak fel kell olvasnia a verseit.

John-DAT up needs read-INF-3SG his poems 'John needs to read out his poems.'

c. János fel szeretném, hogy olvassa a verseit.

John up I.would.like that read-SUBJ-3SG his poems 'I would like that John read out his poems.'

d. János fel kell, hogy olvassa a verseit.

John up needs that read-SUBJ-3SG his poems 'It is necessary that John read out his poems.'

É. Kiss adds that the prefix can also undergo topicalization and focusing, see (9), and both operations are phrasal.

(9) a. [_TopP Fel [_FP János olvasta a verseit]]¹³ up John read his poems 'Loudly, JOHN read his poems.'

b. [_TopP János [_NegP nem [_FP fel ment a lépcsőn]]], hanem le ment.

John not up went the stairs but down went 'It was not up but down that John went the stairs.'

13 Two remarks are in order here. (i) In her more recent model, É. Kiss (2002) uses the Focus Phrase (FP) functional projection, so the focused constituent in (9a) has been moved to Spec,FP, as opposed to Spec,VP in E. Kiss’ (1994a) framework, see (KÉK103a) above. (ii) I do not think that (9a), which is the same as (6a) above, is an appropriate (felicitous) example for the following reason. Elsewhere É. Kiss points out that the verbal prefix can be focused if it is meaningful enough (e.g. when it expresses direction), see her example in (KÉK104) above. It stands to reason that the condition on the availability of the contrastive topic status to the prefix is the same kind of contentfulness (only something contentful enough can be contrasted). The problem with (6a) is that fel ‘up’ in this example does not express direction: it expresses a mode of reading: ‘loudly’, cf. read out in English; therefore, it cannot be felicitously

“contrastive-topicalized”, as it cannot be naturally (and minimally) contrasted with something else (cf.# le olvas down read ‘read in’). The following example with fel ‘up’ as a contrastive topic expressing direction would be appropriate.

(i) Fel JÁNOS dob-ta a labdá-t és nem KATI. up John.NOM throw-PAST.3SG.DEF the ball-ACC and not Kate.NOM

ca. ‘As regards upward motion, it was John who threw up the ball, and not Kate.’

É. Kiss also observes that the prefix can constitute an elliptical sentence on its own, which is yet another phrasal property, see (10).¹⁴

(10) Fel olvasta János a verseit?

'Did John read out his poems?' Fel.

'He did.'

On the basis of the above facts and considerations, É. Kiss concludes that the prefix is an AdvP which only consists of a head, it is base-generated postverbally, among all the other arguments of the verb, and it is moved into the specifier position of a functional projection, that is, phrasal movement takes place. Consider her structure in (11).

(11) XP Spec X' X VP V DP fel olvasta a verseit up read his poems

She points out that assuming this structure would receive considerable support if the VP in it could undergo coordination, one of the classic constituency diagnostics. However, this expectation is not satisfied:

(12) a. János [fel [hívta Marit]] és [fel [olvasta neki a versét]]  John up called Mary and up read her his poem b.*János [fel [hívta Marit] és [olvasta neki a versét]]

(12b) is ungrammatical on the reading on which the prefix is interpreted as also belonging to the second conjunct. É. Kiss remarks that if we assume the following structure for the predicate part of (3) then the ungrammaticality of (12b) can be naturally accounted for.

(13) XP X VP VM V t DP

Furthermore, from this structural approach the grammaticality of the coordination in (14) also follows.

14 It is to be noted, however, that short answers of the (10) type also admit clearly word-level elements, for instance parts of compounds:

(i) A: Ez fehérbor vagy vörösbor?

this white_wine or red_wine ‘Is this white wine or red wine?’

B: Vörös.

red

(14) János [[le írta] és [fel olvasta]] a verset.

John down put and up read the poem 'John put down and read out the poem.'

É. Kiss adds, however, that the grammaticality of (14) does not provide strong evidence for the structure in (13), because (14) can also be analyzed as an instance of right node raising.

Then she continues in the following way. “The contradictory evidence indicating the phrasal nature of prefix movement, and the incorporated head status of the prefix in surface structure can be resolved if we analyze the prefix as a projection both minimal and maximal, capable of acting either as a phrase or as a head. Suppose that the verbal prefix is generated in postverbal position as an AdvP complement of the V; then it undergoes phrasal movement into the

Then she continues in the following way. “The contradictory evidence indicating the phrasal nature of prefix movement, and the incorporated head status of the prefix in surface structure can be resolved if we analyze the prefix as a projection both minimal and maximal, capable of acting either as a phrase or as a head. Suppose that the verbal prefix is generated in postverbal position as an AdvP complement of the V; then it undergoes phrasal movement into the

In document Laczkó Tibor Aspects of Hungarian Syntax from a Lexicalist Perspective (Pldal 135-154)