BeNeCLA Blog

De laatste bijeenkomst voor de zomer van de Construction Grammar Discussion Group vindt plaats op vrijdag 19 juni a.s. vanaf 15.30 uur in zaal 307 van het Lipsius-gebouw v/d Universiteit Leiden.

Nederlands als constructietaal

Ton van der Wouden

Dankzij corpora zoals het Corpus Gesproken Nederlands kunnen we meer te weten komen over eigenaardigheden van gesproken Nederlands. In mijn voordracht zal ik een aantal typische spreektaalfenomenen van zeer uiteenlopende aard bespreken, en ik zal proberen aannemelijk te maken dat een constructiegrammatica-aanpak daarbij dikwijls zijn vruchten afwerpt.

Date and time: Friday April 24, 15h30-17h

Location: Leiden University, Eyckhof 1, 001A.

Syntax or semantics?

A construction based approach to verb ellipsis & voice switch

Egbert Fortuin

In construction based approaches to ellipsis, elliptical structures are treated as constructions that are associated with a specific meaning (e.g., Culicover & Jackendoff 2005). This differs from generative approaches that account for ellipsis in terms of non-audible syntactic structure. Recently, Merchant (2007) has argued that the ‘what-you-see-is-what-you-get’ approach to ellipsis by Culicover & Jackendoff (2005) is fundamentally flawed. Merchant points at the relation between elliptical constructions where the verb is elided, and voice switch. Merchant notices that in some types of elliptical constructions, for example sluicing, the elided material and the antecedent phrase must match in voice, e.g.:

(1) *Joe was murdered but we don’t know who Joe murdered.

(2) *Someone murdered Joe, but we don’t know who Joe was murdered by.

This is, however, not a general phenomenon of verbal ellipsis. In the case of VP-ellipsis (verbal phrase ellipsis) , voice switch is acceptable:

(3) The janitor must remove the trash whenever it is apparent that it should be removed.

(4) The system can be used by anyone who wants to use the system.

According to Merchant, models that try to explain ellipsis from a semantic point of view cannot deal with the phenomenon of voice switch in the context of ellipsis in an adequate way. In this talk, I will explain the data within a construction based approach to ellipsis My explanation is based on the following principles:

I. Semantic-syntactic incompatibility

Verbal ellipsis presupposes that the verbal information is easily retrievable from the context. Voice switch disturbs the principle of ‘ease of retrieval’ because of the difference semantics and correlated structure associated with passive and active sentences.

II. Semantic-syntactic dependency

The more semantically-syntactically independent the elliptical construction is from its antecedent, the easier it is to have ellipsis under voice switch:

- The more semantic-syntactic integration between elliptical clause and antecedent, the stronger the requirement of semantic-syntactic identity between antecedent and ellipsis site.

- If the elliptical construction contains a verb with information about voice, voice switch is possible.

Culicover, P. and R. Jackendoff. 2005. Simpler syntax. Oxford : OUP

A Constructional Analysis of Quasi-Incorporation in Dutch

Geert Booij (University of Leiden)

Date and time: March 27, 15.30-17.00
Location: Utrecht University (precise location will be added)

Reading: If you would like to read the article we will be discussing, please send an e-mail to Elena Tribushinina.

Abstract: This paper deals with quasi-incorporation, the phenomenon that bare nouns form tight phrasal lexical units with verbs. The data are mainly from Dutch. The semantics of quasi-incorporation is similar to that of real noun-incorporation: the bare noun receives a generic interpretation, and the NV combination denotes a conventional activity. However, the lexical units are phrasal since they are separable in root clauses, and in verb raising clusters. It is shown that such Dutch bare Noun + V combinations (either singular or plural nouns) have two structural interpretations. They are either VPs with an NP that consist of the bare noun only, or they are units of the form [N⁰ V⁰]_V0, and thus a case of adjunction of the bare noun to the verb. The behaviour of these NV combinations under verb raising, in the Dutch periphrastic progressive construction, and with respect to the choice of the correct negative word (geen or niet) is shown to follow from the assumption of two possible structural analyses. Thus, Dutch quasi-incorporation is parallel to the analysis of similar cases of incorporation in Japanese, presented in Iida and Sells (2007). If the bare noun has no argumental role with respect to the verb, it is only the second structural option that is available, that of adjunction.

Quasi-incorporation can be seen as a construction, in which a specific syntactic pattern receives a specific semantic interpretation, that of conventional activity. Thus, the paper argues that the notion ‘construction’ is essential for giving a proper account of the semantics of quasi-noun incorporation.