Lingvisticæ Investigationes - Volume 12, Issue 2, 1988

Orthographic variations of compound words in le Petit Larousse and le Petit Robert.The analysis of the two most common French dictionaries — le Petit Larousse and le Petit Robert — brings into relief the orthographic instability of compound words. This study notes many disagreements between the two works, as well as a number of internal contradictions. Hesitation concerns mainly the use of the hyphen (audio-visuel or audiovisuel, fille-mère or fille mère) and plural marks (un corps de troupe/un corps de troupes, des essuie-glace/des essuie-glaces). Whatever may be the best way to correct such contradictions — admission of variants or normalization — it is urgent that dictionaries be coherent.

In this paper, I distinguish three different uses of the verb cuta: dative verb, support verb, and causative operator verb. The syntactic properties of a sentence containing the verb cuta vary with the lexical choice of the direct object. The subject of the sentence in which cuta is a support verb is seman-tically the subject of the direct object. This special relationship subject -direct object doesn't exist in the sentence with a causative operator verb cuta in which the indirect object is semantically the subject of the direct object. Furthermore, the distribution of the subject in the sentence with a dative verb cuta is different from that in the sentence with a causative operator verb cuta. The causative operator verb cuta takes the subject of the type "unrestricted noun". I present the principal syntactic properties of the verb cuta in the columns of the table. The sign "+" indicates that the verb has the corresponding property: the sign "-" that the verb does not have this property.

It is well known that syntactic configurations such as active vs passive are not interchangeable in all contexts and that they can be interpreted adequately only within the framework of the discourse. It is our contention that precisely the same phenomenon underlies the way in which possession is expressed in Romanian, and we argue that Romanian exploits the possibility of two distinct predicative relationships in this respect: "Possessor possesses possessum" and "possessum belongs to Possessor".A similar explanation is given for the choice between lui and sau with a third person Possessor, a fact of Romanian grammar not hitherto understood.

The paper proposes a new approach to temporal clauses and temporal adverbs, which are mostly said to "localize" the main verb. In recent work of Vicenzo Lo Cascio and others, temporal clauses and adverbs are treated completely differently and independently. We want to show that they should be handled together. Therefore, we start from a different conception of temporal localization, based on non-linguistic localization, as used in f.e. geography. So, we are able to show similarities between temporal clauses and adverbs. Our conclusion is that the localizer is always the time interval which is situated at the background, although it is not necessarily expressed by the temporal clause or adverb. On the contrary, a temporal clause or adverb can also be localized by the verb of the main clause.

Several types of duration measurement verbs (Vmtd) can be isolated in french as the result of the description of different types of constructions that can occur with temporal nouns (Ntps) and duration nouns (Ndur). A series of syntactic and semantic properties can be gathered concerning these constructions, leading to the constitution of four main categories, for which durer, prendre, passer, mettre are the best representative verbs. In this study, we try to show the difference that can be made between this specific verbal complementation and durative adverbial complementation in general.

In the tradition of generative grammar, two different definitions of "subject" have been proposed: we name the first one "structural", since it is based on the constituent structure of the sentence, and the second one "thematic", because it makes use of the concept of thematic role. We argue that both definitions are necessary, since they deal with two different entities, i.e. the structural subject and the thematic subject, which are to be kept distinct. In particular, we show that opacity phenomena are induced by the "thematic" subject, and not by the "structural" one (in showing this, we make use of the notion of "Complete Functional Complex" recently proposed by Chomsky); this kind of analysis allows us to dispense with the so-called "i-within-i condition", which appears theoretically unsound. Some cases of behavior of anaphors which seem to run against our proposal are also discussed, and it is shown that they can be accounted for on the basis of independent principles.

A description of 1600 French verbs shows that three types of control verbs should be distinguished in French: subject control, object control, and variable control verbs. The control properties of these verbs can be related to their syntactic or distributional characteristics.It appears that object control verbs and the different subclasses of variable control verbs are semantically coherent: each class of variable control verbs can be described as a particular type of transfer (promettre, garantir/ demander supplier/ proposer, offrir). These types of transfer are described by a highly elaborate semantic description that allows for deontically and temporally defined Source / Goal relations.This semantic approach also incorporates a description of the agentive properties necessary for a complement to qualify as a controller by the introduction of the notion of direct and indirect agentivity. Thus, we are able to explain why the subject of object control verbs cannot qualify as a controller, and how control shifts come about.On the theoretical side, this analysis shows that the GB distinction between obligatory and non obligatory control is not relevant, and that the empty element PRO is superfluous.