Institut de Lingüística Aplicada

Third Mediterranean Meeting on Morphology (MMM3)


Jaume Mateu (Universitat Autònoma de Barcelona)
Complex Denominal Verbs and Parametric Variation: A Lexical-Syntactic Approach

I. The specific purpose of this paper is to provide a lexical-syntactic account (Hale & Keyser (1997) (henceforth HK)) of those complex denominal verbs whose formation can be argued to involve a 'lexical subordination process' (cf. Levin & Rapoport (1988); Spencer & Zaretskaya (1998); Mateu & Amadas (1999)): e.g., cf. the German examples in (1a-b) drawn from Stiebels (1998: 285-286).

Stiebels (1998) argued that HK's syntactic approach appears to be problematic when confronted with complex denominal verbs like those in (1a-b). She pointed out that complex verbs with an integrated adjunct (e.g., cf. the prefixes ver- and er- in (1a-b)) should not occur according to a syntactic approach like that of HK, since adjunct incorporation is argued to be impossible in L-syntax. According to Stiebels (1998: 270), "complex denominal verbs (<like those in (1a-b)>) constitute an important touchstone for HK's proposal". Rebus sic stantibus, she argues for a semantically-based approach: e.g., cf. (2) for her Lexical Decomposition Grammar (LDG) analysis of (1a).

Taking up her challenge, we reply as follows: Stiebels's (1998: 285) requirement that the verbal prefixes in (1a-b) be "lexical adjuncts" (sic) is not to be taken for granted. According to the 'lexical subordination approach' (cf. Spencer & Zaretskaya (1998)), it is precisely the preverb element (e.g., ver- in (1a)) that must be considered as part of the main thematic structure, the surface head element (e.g., [gärtner]V in (1a)) being a subordinate predicate. That is, despite appearances, the "added" element is not the directional/resultative prefix, but the process denominal verb, the latter being considered as a true adjunct to the main thematic structure. The same can be argued to hold for (1b). Notice that it is not coincidental that this subordination analysis goes hand-in-hand with the English analytic translations of the examples in (1). However, unlike Spencer & Zaretskaya (1998), we claim that complex predicates like those in (1) are not to be formed at a lexical-conceptual level of representation (i.e., LCS), but at HK's lexical-syntactic level, the latter level being the locus of parameterization of morphosyntactic facts affecting argument structure (cf. infra). This accepted, we claim that the lexical relational analysis of complex denominal verbs like that in (1a) involves the syntactic composition of two different Lexical Relational Structures (LRSs), the main one being transitive (cf. (3a)), and the subordinate one being unergative (cf. (3b)). Following Hale & Keyser's (1997: 228-229) analysis of 'conflated' structures like Sue danced into the room (i.e., 'Sue went into the room dancing'), we posit that complex denominal verbs like that in (1a) can also be analyzed by means of a 'generalized transformation' (cf. Chomsky (1995)), the verb of (3a) being replaced by the denominal verb in (3b): See (4). As in Hoekstra's (1988, 1992) Small Clause (SC) approach, the directional/resultative prefix (e.g., ver-) is assumed to be the head of the inner 'SC' projection (i.e., P), which turns out to be adjoined to the superior verbal head because of its affixal status.

II. More generally, we want to argue that there is a morphosyntactic reason involved in Talmy's (1991) distinction between 'satellite-framed languages' like German, Dutch, Russian, or English, and 'verb-framed languages' like Catalan, Italian, French, or Japanese (cf. Mateu & Rigau (1999)). In Talmy's (1991) terms, the data in (1) obey the following 'lexicalization pattern', which is shown to be typical of satellite-framed languages: i.e., conflation of AGENTIVEMOVE with [EVENT]SUPPORTING. Let us exemplify it with the analysis of (1a). To put it in our present lexical-syntactic terms, the 'satellite' nature of the Path relation ver- allows an independent verb (e.g., cf. the unergative LRS in (3b)) to be conflated into the main verb (i.e., the V in (3a)), the former providing the latter with phonological content (cf. (4)). Furthermore, an additional step in the derivation of (1a) appears to be involved: the affixal nature of the Path relation forces it to be adjoined to the superior verbal head (vs. cf. Engl. He gambled all his fortune away; but cf. I {outplayed/outran/outcooked} him).

By contrast, Romance languages are 'verb-framed': as noted by Talmy (1985, 1991), the conflation of the Path relation into the verb has a fossilized status (cf. Cat. pujar 'go up'; baixar 'go down'; sortir 'go out': i.e., what corresponds to the verb and what to the Path relation cannot be distinguished any longer). It is precisely this fossilization process what appears to prevent an independent 'Manner' component from being conflated into the verb (cf. Mateu & Rigau (1999)). Crucially, notice that the latter fact explains why Romance languages typically lack satellite-framed constructions such as those 'morphological objects' exemplified in (1) or lexicalized resultatives like Ger. leertrinken 'drink empty' or plattlaufen 'run flat' (cf. Jackendoff (1990) or Goldberg (1995) for the claim that resultatives involve an abstract Path). On the other hand, unlike Marantz's (1997) reductionist proposal, we want to argue that HK's (1993, ff.) distinction between Lexical-syntax and Sentential-syntax appears to be motivated on the basis of the lexical fact that Romance languages (and more generally, verb-framed languages) lack both telic Path of Motion constructions like Mary danced away / He gambled all his fortune away and complex resultative constructions like They talked us into a stupor / The dog barked the chickens awake. Our claim is that these constructions and those morphological objects in (1) are not to be formed in Sentential-syntax (Déchaine's (1996) 'big syntax')) but in Lexical-syntax, the latter level being regarded as the proper place where such parameterizable facts affecting argument structure can be naturally argued to be encoded.

  1. a. Er ver-gärtner-te sein gesamtes Vermögen. (German)
    he VER-gardener-ed his whole fortune 'In gardening, he used up all his fortune'.

b. Sie er-schreiner-te sich den Ehrenpreis der Handwerkskammer.
Sie ER-carpenter-ed herselfDAT the prize of the trade corporation
'She got the prize of the trade corporation by doing carpentry'. Stiebels (1998: 285-286)

  1. I) [ ]V 8Q 8x 8s Q (x) (s)
    II) [gärtner]V 8x 8s GARDENER (x) (s)
    III) ARG (gärtner): 8R 8x 8s [GARDENER (x) (s) & R (s)]
    IV) ver- 8u 8s CONSUME (u) (s)
    V) [ver[gärtner]V ]V 8u 8x 8s [GARDENER (x) (s) & CONSUME (u) (s)] N.B.: The ARG-operation is a semantic argument extension operation, which allows the semantic integration of the prefix into the verb (Wunderlich (1997)).

  1. a. [tree]
    N.B.I: The external argument is not present at l-syntax (HK (1993, 1997, 1998)).
    N.B.II: Directional/resultative prefixes and PPs involving a 'terminal coincidence relation' (HK (1993, 1997)) can be argued to be assigned the same argument structure (both contain a birelational Path element), the difference being that the former involve the conflation of a non-relational element X (i.e., an abstract Ground) into a directional relational element P (i.e., the Path). N in (3a) is to be interpreted as 'Figure/Theme'.

  1. [tree]

Selected references