The light verb hypothesis and the construction of syntactic meaning: Flavors of v
In work on the lexicon-syntax interface, traditionally the study of verb classes and alternations is a fundamental field of investigation, because the identification of common syntactic properties belonging to verbs with common semantic characteristics has supported the hypothesis that important generalisations are indeed possible. The variable behavior of verbs is a complicated issue whose explanation has been at the centre of two decades of work on aspectual classes and argument projections. It would seem that either we believe that in the lexicon of a language we have different entries (possibly derived) for each of the alternating verbs, and that therefore the syntactic computation is working with only one of the possible entries each time; or we have to make the derivation of the different forms a matter of syntactic computation. This latter route is attractive if we assume that the lexicon is optimally minimal, and it becomes essential if we accept the arguments in favor of eliminating the generative lexicon altogether (see, e.g., Marantz 1997).
In the literature these questions have been tackled from both sides. Given the agreement on the necessity of making generalisation about verb classes and syntactic structure, different proposals have been placing the burden of the explanation in either one of the two components, the lexicon (Chomsky 1981, Pustejovsky 1988, Levin and Rappaport-Hovav 1995, Jackendoff 1990, Baker 1988, among others) or the syntactic computational system (Hale & Keyser 1993, Halle & Marantz 1994, Borer 1994, 1998, Travis 1994, Harley 1995, Kratzer 1996, Ritter and Rosen 1998, among others). Both positions base their argumentations on the two fundamental classifications that studies on verbs have produced: a classification in terms of the aspectual structure they encode (Kenny 1963, Vendler 1967, Smith 1991, among others) and a classification of verbs in terms of the argument structure(s) they allow (Perlmutter 1978, Burzio 1986). Theories of the lexicon-syntax interface take advantage of these classifications and try to address a further question, that is how these distinctions are represented in the grammar and what is the division of labour between the lexical and the syntactic module.
In this paper we pursue a constructionalist-type explanation. According to such an approach, it is not entirely the lexical semantics of a verb that determine its syntax. Rather, the functional/aspectual structure in which a verb root is inserted and therefore the syntactic positions in which its arguments are realised determines its interpretation. For such theories, the construction of functional event structure on top of the predicate is responsible for the assignment of event roles to the participants in the event, and accordingly the construction of different event structures gives rise to verb alternations.
The fundamental motivations for pursuing this kind of approach are both theoretical and empirical. Starting with the theoretical strength of a syntax-based approach, it has been argued that there are a number of syntactic phenomena that a lexicalist approach would have difficulty explaining (see for example Rosen 1984 for a discussion of the phenomenon of unstable valency of unaccusative/unergative verbs, Hoekstra and Mulder 1990 on the alternating behaviour of motion verbs in Italian and Dutch). But even more implausible is to imagine that all the syntactic correlates of the unaccusative/unergative distinction (ne-cliticization, auxiliary selection, availability of resulative construction, er-nominalisation in English, and possessor datives in Hebrew etc.) are not a reflection of structural facts (as suggested, e.g., by Dowty 1991). This approach seems strongly supported from psycholinguistic argumentation showing the explanatory power of the syntactic bootstrapping account of verb acquisition (Gleitman 1990, Borer 2000, van Hout 2000, 2001) and from the evidence that event structure is accessed before argument structure in the process of sentence comprehension (O’ Bryan et al. 2003).
However, the generation of alternations is only one side of the problem that theories of the lexicon-syntax interface have to solve. The other, of course, is the absence of certain forms that an extreme constructionalist position would predict should occur. Why can we not say #The city destroyed, or #The giraffe fell that the monkey ate a banana? Different factors have been exploited in explaining this problem; Marantz 1997 and Harley & Noyer 2000 argue that the lexical semantics of the roots themselves are responsible in at least the former case; the poorness of #The city destroyed is thus analogous to that of #Colorless green ideas.
In this paper, however, we wish to address a different class of cases, where certain valency alternations are not forbidden, but rather have unexpected syntactic consequences. The first class of cases concern the effect of varying the animacy of an agent on the syntax verbs of consumption; the second class concerns some subtle problems in tthe famous Faire Par/Faire Infinitif alternation of Kayne. The way we propose to resolve both issues is by proposing that there are different flavours of little v. There is a consensus that there is at least causative v and inchoative v, which we will notate vCAUSE and vHAPPEN. We will argue, however, that this typology of v is inadequate to account for the alternations mentioned above.
To take the first case, consider consumption verbs. These verbs, unlike non-alternating destroy-class verbs, do not generally allow inanimate agents.
1. a. The sea destroyed the beach. /The groom destroyed the wedding cake.
b. *The sea ate the beach. /The groom ate the wedding cake.
c. Il mare ha distrutto la spiaggia. /Lo sposo ha distrutto la torta nunziale
The sea has destory.PST the beach /The groom has destroy.PST the cake nuptial.
d *Il mare ha mangiato la spiaggia / Lo sposo ha mangiato la torta nunziale.
The sea haseat.PST the beach /The groom has eat.PST the cake nuptial.
In order to capture this restriction (which is quite general for the class of verbs in question), we propose a third flavor of v: vDO, following Hale and Keyser 1993. This light verb has different selectional properties than vCAUSE: a) While vCAUSE allows any kind of causer subject, vDO only allows animate agent subjects, and b) vCAUSE selects for a complex complement describing a state of affairs, vDO selects a simple DP complement. Consumption verbs, based on the vDO structure, therefore do not allow simple substitution of inanimate subjects, as it would violate a).
Nevertheless, it is possible to use an inanimate subject argument with these verbs. Crucially, however, an inanimate subject forces a change in the event structure of the predicate. Consider the examples below:
2. a. The sea ate away the beach. /*The sea ate the beach.
b. The wind carved away the beach. /*?The wind carved the beach
c. Il mare si é mangiato la spiaggia /*Il mare ha mangiato la spiaggia
The sea REFL is eat.PST the beach /The sea has eat.PST the beach
d.Il vento si é ritagliato un pezzo di spiaggia /*Il vento ha ritagliato un pezzo…
The wind REFL is carve.PST a piece of beach /The wind has carve.PST a piece…
In the English examples, the particle away is the syntactic realization of the resultant state. In Italian, we argue, the clitic verb si realizes the change to vCAUSE, accounting also for the shift in auxiliary selection (Folli 2002, Lidz 1998, Zubizarreta 1987).
We can use the same device with profit in analyzing the FI/FP alternation. It is well known (Kayne 1975, Guasti 1996) that the two constructions display a number of syntactic and semantic differences. We argue that these differences reflect a difference in flavor of vP. In FI constructions, fare is itself a vCAUSE and embeds a vDO, in FP constructions fare is itself a vDO embeding a nominalized VP (Travis 1998). We show (a) that the semantic characteristics of the complement vP in the FI construction explain an 'obligation' requirement on the FI Causee, which is absent in the FP construction; (b) the vDO semantics for fare in the FP explains why inanimate subjects of fare (‘Causes’) are impossible in the FP; and (c) the vDO proposal explains a previously unobserved generalization: the fact that causatives of unergative intransitive verbs may not be passivized, while causatives of other types of verbs may occur in the passive.
The advent of the little-v hypothesis in the early 90s allowed theoreticians to come to grips with the manifestly syntactic nature of verb-frame alternations. The present line of research delves deeper into the nature of little-v, searching for explanations for apparently lexical restrictions on such alternations. It brings to light mechanisms that can account for quite subtle syntactic and semantic effects of verbal alternations, and has allowed us, in some cases, to discover new paradigms of facts which previous approaches failed to predict or observe.
Baker, M. 1988, Incorporation: a Theory of Grammatical Function Changing. Chicago: University of Chicago Press.
Borer, H. 1994, The Projection of Arguments. In Benedicto, E., and Runner, J. (eds.). Functional Projections, 19-47. Ameherst: Umass.
Borer, H. 1998, Deriving Passives without Theta-Grids. In Lapointe, S., Brentary, D., and Farrell, D. (eds.). Morphology and its Relations to Phonology and Syntax, 60-99. Stanford: CSLI
Borer, H. 2001, The Grammar Machine, Ms., USC.
Burzio, L. 1986, Italian Syntax: a Government and Binding Approach. Dordrecht: Reidel.
Chomsky, N. 1981, Lectures on Government and Binding. Dordrecht: Foris.
Guasti, M.T. 1996, Semantic Restrictions in Romance Causatives and the Incorporation Approach, Linguistic Inquiry, 27, 294-313.
Folli, R. 2002, Constructing telicity in English and Italian. Unpublished doctoral dissertation, University of Oxford.
Gleitman, L. 1990, The Structural Source of Verb Meanings. Language Acquisition, 1, 3-56.
Hale, K. and Keyser, S.J. 1993, On Argument Structure and the Lexical Expression of Syntactic Relations. In Hale, K. and Keyser, S.J. (eds.). The View from Building 20: Essays in Linguistics in Honor of Sylvian Bromberger, 53-109. Cambridge Mass.: MIT Press.
Hoekstra, T. and Mulder, J. 1990, Unergatives as Copular Verbs. The Linguistic Review, 7, 1-79.
van Hout, A. 2000, Projection Based on Event Structure. In Coopmans, P., Everaert, M., and Grimshaw J. (eds.). Lexical Specification and Insertion, 403-427. Amsterdam: Benjamins.
Jackendoff, R. 1990, Semantic Structure. Cambridge Mass.: MIT Press.
Kayne, R. 1975, French Syntax, Cambridge Mass, MIT Press
Kenny, A. 1963, Action, Emotion and Will. London: Routledge & Kegan Paul.
Kratzer, A. 1996, Severing the External Argument from its Verb. In Roorych, J. and Zaring, L. (eds.). Phrase Structure and the Lexicon, 109-137. Dordrecht: Kluwer.
Levin, B. and Rappaport-Hovav, M. 1995, Unaccusativity. At the Syntax-Lexical Semantics. Cambridge Mass.: MIT Press.
Lidz, J. 1999, Causativity, late insertion and vP. In L. Pylkkänen, A. van Hout & H. Harley (Eds.), Papers from the UPenn/MIT Roundtable on the Lexicon, MITWPL 35 (pp.117-136). Cambridge, MA: MIT Working Papers in Linguistics.
Marantz, A. 1997, No Escape from Syntax: don’t Try Morphological Analysis in the Privacy of your own Lexicon. In Dimitriadis, A., and Siegel, L. (eds.). University of Pennsylvania Working Papers in Linguistics, 4.2. 201-225.Perlmutter 1978,
Pustejovsky, J., 1988, The Geometry of Events. In Tenny, C. (eds.). Studies in Generative Approaches to Aspect, 34-62. Lexicon project working papers 24, Centre for Cognitive Science, MIT.
Ritter, E., and Rosen, S.T. 1998, Delimiting Events in Syntax. In Butt, M., and Geuder, W. (eds). The Projection of Arguments, 135-164. Stanford: CSLI.
Smith, C. 1991, The Parameter of Aspect. Dordrecht: Kluwer
Travis, L. 1994, Event Phrase and a Theory of Functional Categories. In 1994 Annual Conference of the Canadian Linguistic Association, 559-570. Toronto Working Papers in Linguistics.