Lingvisticæ Investigationes - Volume 32, Issue 2, 2009

This paper offers an analysis of the French integrated consecutive (result) construction, starting out from the correlative system observed in the ancient Indo-European languages, in particular the one which J. Haudry terms « dyptique inverse ». It is shown that the morpheme que, a relative marker, has a content that is identifiable on the basis of the main clause (as in the case of canonical relative clauses) but which functions as a causal principle in relation to the subordinate clause. At the same time, it is also shown that sentences containing consecutive subordinate clauses involve two distinct structures superimposed one on the other: one of these is defined in terms of the hierarchical constituents of distributional analysis, and the other in terms of two independent clauses. The latter structure makes it possible to account for certain ambiguities that may be found in their interpretation. In that respect as well, consecutive clauses are similar to canonical relatives.

In the present article we are studying the relation between the prefix sun- and the simple verbal basis of the Greek lexicon-grammar in order to find the regularities pertaining to the behavior of the prefix sun-, which will contribute to the semi-automatic enrichment of the electronic dictionary DELAGR. A series of observations on the syntactic structures containing these verbs reveal the fact that prefixation has to be considered in terms of predicative relations and arguments. The article is divided in three parts: 1. presentation of some particularities; 2. description of the inflection of prefixed verbs 3. presentation of data and restraints. In conclusion we are providing some statistic elements concerning the formation of prefixed verbs with sun-.

In this preliminary descriptive research, taking one French verb (expliquer), we explore syntactic possibilities of argument realization. For this verb, at least four types of syntactic constructions are specified for a fixed set of semantic arguments.
 When the direct object is realized as a NP (and not as a bare Que P) assuming the role of “object” of explanation, it is observed that the subject can vary between the “agent” and “explanation” roles, in correlation with the change of the origin of “explanation”. This type of difference is to be considered a change of diathesis of the same verb.
 An agentive construction with a complement clause object happens to be a marginal one from the point of view of argument realization, functioning accidentally as a variant of a sentence with verbs of saying. This type of complement clause is a bare Que P in the sense that it does not manifest an alternation with a Que P headed by le fait.

A parallel is drawn between two clause-types carrying an obligatory pro-form: wh-questions and the so-called free relative of pseudo-clefts. In the analysis both types are derived from a single counterpart which is structurally simpler. The constituent targeted in the former and focused in the latter always corresponds to the VP of their counterpart. The mandatory verb fare ‘do’ that occurs in the question and in the free relative is analyzed as a support verb. In both cases its origin is traceable to the syntactic function of the pro-form, which is claimed to be double, in that it simultaneously works as a predicate and an argument (direct object).

The contribution offers a critical review of existing works on nominal predication with light verbs in Czech. It proposes also three key elements in which the principles of nominal predication in Czech differs from Romance languages: absence of nominal determination, obligatory marking of aspectuality and double system of case marking. Analyse of these elements shows that in Czech, variability of light verbs has a high degree of dynamics synchrony.

This paper deals with psychological predicates in Ancient Greek. The corpus is composed of sentences found in a representative collection of Ancient Greek texts, in which nouns designating psychological states show predicative function. Some lexical-syntactic properties here accounted for concern specifically psychological constructions; others more generically refer to the Ancient Greek morphosyntax: in particular the domain of nominal sentences and that of lexical-syntactic expression of the opposition middle vs. non-middle. Differences and similarities with other Indo-European languages, such as Sanskrit and Italian, are also mentioned.

An analysis of all the French expressions presenting a coordination reveals the conditions necessary to convey intensification values. The connector et (and) needs to have an additive value (or the connected value of precision or overbid) to trigger off intensification. This happens when it is combined with coordinated elements that have certain semantic connections with either addition or duplication : synonymy, total or partial referential convergence. Two degrees of intensification can be found : a relative degree (really, vraiment) when re-definition is at stake, and a high degree in the two other cases.

This study examines a corpus of 300 compositional phrasal verbs in English using the particle up. Whereas frozen phrasal verbs clearly do not have the same meaning without the particle and must be listed in the lexicon grammar of idiomatic expressions (e.g., break up the audience “cause to laugh”), compositional phrasal verbs merely add an aspectual element to the simple verb, and particle usage might be better analyzed as a property of the simple verb. Although this meaning is difficult to characterize precisely, it generally indicates direction (ride up the elevator), completion (open up the store), intensity (hype up the book), or combinations (polish up the silverware, completion and intensity). These properties are added to the table of compositional phrasal verbs. Optional particle usage can also identify the meaning of a potentially ambiguous simple verb. For example, the simple verb call “phone” can take the particle up but not out, while the simple verb call “announce” can take the particle out but not up. The simple verb tables, however, will become enormously complex when all English particles (fifteen different particles according to B. Fraser 1976) are taken into account.

In the present paper “Similarities between Comparative and Exclamative Degree Markers: French ce and Spanish lo”, we first analyse the so-called pseudo-relativized phrasal comparative construction X que ce que P in French and X de lo que P in Spanish. We after compare the pseudo-relativized phrasal comparative construction properties with the properties of the exclamative French construction Ce que P! and of the Spanish one ¡Lo X que P ! whose form, degree reading, and some syntactic and semantic features are similar. We eventually conclude about the nature of que/que, introducing both subordinate constructions, as a complementizer (and not a Wh- item).

We present an analysis of opinion in texts based on a detailed semantic analysis of a wide class of expressions. We propose a new annotation schema for a deep contextual opinion analysis using discourse relations. We analyze the distribution of our categories in three different types of online corpora, movie reviews, Letters to the Editor and news reports, in English and French.

In this paper, we describe how pronominal constructions are represented in Dicovalence and in the lexicon-grammar. We introduce a method for extracting and merging lexical syntactic information about these constructions, and integrating it in the Lefff NLP lexicon.

This paper presents a work on extending the adverbial entries of the WOLF, a semantic lexical resource for French, and connecting them with those of the syntactic lexicon Lefff, which were mostly extracted from the lexicon-grammar tables from (Molinier & Levrier, 2000). This work relies on the exploitation of the derivation and the synonyms relations ; the latter are extracted from the DicoSyn synonyms database. The resulting semantic resource, which is freely available, is manually evaluated and validated in an exhaustive manner.