qxls

SN

2.2.2.4 head shift 2.2.3 dependants zero 2.3 cohésion 2.3.1 accord 2.3.2 ordre 2.3.2.1 nucleus & arguments 2.3.2.2 nucleus & modifieurs 2.3.2.3 dependants&dependants 2.3.3 compatibilités 2.3.4 syncrétismes 2.3.5 head attraction 2.3.6 linkers 2.3.7 SN discontinus 2.3.8 apposition

1 cadre 1.1 domaine 1.2 exclusions 1.3 concepts 1.3.1 extension / intension 1.3.2 entité / referent 1.3.3 composantes formelles 2 forme 2.1 nucleus 2.1.1 nature (théorique) du nucleus 2.1.1.1 categorie lexicale 2.1.1.2 categorie fonctionnelle (D) 2.1.2 classe (lexicale) du nucleus 2.1.2.1 PRN 2.1.2.2 nom propre 2.1.2.3 deverbal 2.1.3 nucleus lexical zero 2.1.3.1 uniquement dépendants 2.1.3.2 relative sans tête 2.2 dependants 2.2.1 arguments 2.2.2 modificateurs 2.2.2.1 determinants 2.2.2.2 adjectifs 2.2.2.3 adjoints

3 fonction 3.1 sémantique 3.1.1 qualités 3.1.2 possession 3.1.3 quantification 3.1.4 genre 3.1.5 classes 3.1.6 saillance 3.1.7 aspect-temps 3.2 referentialité 3.3 pragmatique

1

qxls

SN

1 cadre There is surprisingly little typological literature on the structure of noun phrases, despite the fact that this is discussed at length in many descriptive grammars. The most detailed discussion is Rijkhoff (2002). There is also basic discussion in Giv´on (1990, 2001). Giv´on, T. 2001. Syntax: An Introduction. Amsterdam: John Benjamins references (entre autres) dryer dans shopen givon syntax II rijkhoff noun phrase creissels supposons acquise l'existence d'un être grammatical 'syntagme nominal'; syntagme = acception large : peut être fait de plus d'un élément, d'un élément, de Ø element 1.1 definition domaine ce qu'il y a dans SN digne d'interet : ce qui va contre les choses banales 1.2 exlusions cas (sauf genitif, modifieurs adjoints, accord en cas) syntaxe externe relatives (recife) noms relationnels-adpositions (grammaticalisation) morphologie nominale coordination ni apposition : ce sont des sequences de SN complets pragmatique de focus = relation avec autres constituants anaphore Ø mais je peux toucher à tout si ça a une incidence sur le thème, par exemple ·relatives pour les questions d'ordre dans SN ·morphologie nominale pour interaction avec d'autres éléments du SN e.g. headmarking etc. 1.3 concepts

2

qxls

SN

1.3.1 extension/intension the extension of a concept, idea, or sign consists of the things to which it applies, in contrast with its comprehension or intension, which consists very roughly of the ideas, properties, or corresponding signs that are implied or suggested by the concept in question. the comprehension of an object is the totality of intensions, that is, attributes, characters, marks, properties, or qualities, that the object possesses, or else the totality of intensions that are pertinent to the context of a given discussion. This is the correct technical term for the whole collection of intensions of an object, but it is common in less technical usage to see 'intension' used for both the composite and the primitive ideas denotation selon la discipline veut dire intension ou extension; exemple :

moi : intension, ça me permet de l'appliquer aux predicats 1.3.2 entité / referent nom = type d'entité, SN = sous-type d'entité + referenttout ce qui s'ajoute au nom dans SN : modeler l'entite + identifier un referent 1.3.3 composantes formelles = nucleus / dependants nucleus (head) : determine les catégories de tout le constituant (dixon) jarawara : sujet declenche accord en genre sur le verbe dans possession alienable, genre sur verbepossédé = nucléus 3

qxls

SN

dans possession inalienable, genre sur verbepossésseur (=genitif) = nucléus le nom nucleus donne un type d'entité, les modificateurs raffinent le type d'entité (les adjoints aident à ça s'ils caractérisent), les déterminants identifient un referent au SN (les adjoints aident à ça s'ils localisent), les compléments saturent des places d'argument (les pronoms n'ont pas de compléments, il ne sont que des référents) dependants lusekelo sur nyakyusa : dependents of the head noun attested in the Nyakyusa data are possessive, demonstrative, adjective, numeral, quantifier, associative, intensifier, relative clause and distributive = arguments / modificateurs the fact that modification is so dispensable in the case of pronouns and names already hints at the major function of noun modifiers — they are used to further specify or narrow down the domain of reference of their head nouns. Such further specification is necessary because nouns, unlike pronouns and names, do not of themselves refer to unique token entities. Rather, they connote types of entities. They thus require further modification in order to become uniquely referring expressions. givon modificateurs = grammaticaux / lexicaux grammaticaux = article, demonstratif, etc. lexicaux = adjectif, adjoint (syntagme adpositionnel) 2 forme forme du SN modelée selon nucleus ou selon dépendants  (i) simple noun phrases, which contain only pronouns or nouns plus simple modifiers like articles, adjectives, demonstratives, or numerals; (ii) complex noun phrases, which contain more complex sorts of modifiers, like genitive or possessive modifiers and relative clauses 2.1 nucleus 4

qxls

SN

2.1.1 nature (théorique) du nucleus 2.1.1.1 nucleus lexical = tradition mais nominalisation sik, classificateurs maya, chinois 2.1.1.2 nucleus fonctionnel (D) justification de DP au lieu de NP (gustave guillaume 1917 : l'article est le nucleus du SN) : (bhatt) : king of england est un constituantdifferents tests dont (1) The present [king of England] is more popular than the last [one]

DP = D + NP NP = N + etc. parallèle structure proposition / phrase  parallele nom /verbe : movima, kat, maya, philippines, eskimo (langues ergatives à origine genitive) TG, nahuatl (langues omnipredicatives) maya : monod & vapnarsly maya ency 1385 Ressemblances structurelles et fonctionnelles des syntagmes nominaux et verbaux : Autour d'un centre radical, se deploie a droite et a gauche un ensemble d'indications touchant a ·l'orientation et a la direction des actions et des mouvements,

5

qxls

SN

·la position dans l'espace des objets selon des calculs absolus (haut et bas, directions cardinales ) aussi bien que relatifs (socio- ou egocentres) [allocentrés/egocentrés serait bien], ·au deroulement dans le temps des processus.  si predicat  prennent arguments si argument  prennent determinants. deictiques, classificateurs parallele structure proposition / phrase  X-barre (fukui) et DP (bernstein) ¡fukui handbook baltin : NAOKI FUKUI phrase structure

DP

Chomsky (1970), however, refutes this ―Transformationalist Hypothesis,‖ [deverbal genre knowledge] and argues that the theory of grammar should not allow a nominalization transformation (or any other transformation with similar expressive power) because it performs various operations that are never observed in any other well-argued cases of transformations. Thus, the alleged nominalization transformation (i) changes category types (it changes S to NP and V to N), (ii) introduces the preposition of, (iii) changes the morphological shape of the element (destroy is changed to destruction; refuse is changed to refusal, etc.), (iv) deletes all auxiliaries, and so on. These are the operations that other wellattested transformations never perform, and hence should not be allowed, [ les nominalisations ne viennent pas de transformations] (i) that derived nominals are really ―nounlike,‖ not sharing various essential properties with sentences, and (ii) that the relationship between derived nominals and their sentential counterparts is rather unsystematic and sometimes unpredictable (see Chomsky 1970 for more arguments establishing these points). He then concludes that derived nominals should be handled in the lexicon, rather than in terms of transformations which deal with formal and systematic relationships between phrase structure trees. This proposal defines the ―Lexicalist Hypothesis,‖ [d'où l'idée que le changement de classe est toujours de la dérivation] Once we adopt the Lexicalist Hypothesis, however, an important problem immediately arises as to how to capture certain similarities and parallelisms holding between verb/noun and sentence/nominal pairs. More specifically, the strict subcategorization properties of a verb generally carry over to the corresponding noun, and the identical grammatical relations are observed in 6

qxls

SN

both sentences and the corresponding nominals [j'appelle ça la nominalisation de la structure argumentale ](see Lees 1960 and Chomsky 1970 for detailed illustrations of these points; see also van Riemsdijk and Williams 1986 for a lucid summary). Under the Transformationalist Hypothesis, these parallelisms are captured by the nominalization transformation. With the elimination of such a transformation under the Lexicalist Hypothesis, we now have to seek an alternative way to express the parallelisms in the grammar. Chomsky (1970) proposes that these parallelisms can be successfully captured if the internal structure of noun phrases is made to be sufficiently similar to that of sentences so that the strict subcategorization properties and grammatical relations can be stated in such a general form as to apply to both verbs/sentences and nouns/nominals [la structure argumentale ne se nominalise pas, il faut fornuler les deux structures d'une seule façon mais suffisamment abstraite] . As a concrete means to express these cross-categorical generalizations, Chomsky introduces a preliminary version of X′-theory [2= '] The ―X‖ in (14) is a variable ranging over the class of lexical categories N(ouns), V(erbs), A(djectives), and (perhaps) P(repositions). The symbol X′ primes rather than bars) stands for a constituent (phrase) containing X as its ―head‖ (the central and essential element of the phrase), as well as those elements appearing in the place indicated by ―. . .‖ in (14a), the elements called the ―complement‖ of X. The schema (14b) introduces a still larger phrase X″ -head elements associated with X′ as [Spec, X′] ).2 Examples of specifiers include, according to Chomsky, determiners as [Spec, N′], auxiliary elements as [Spec, V′], [donc le déterminant occupe, comme le sujet, la position de Spec] comparative structures and elements like very as [Spec, A′], ―projections‖ of X, with the latter (X″) referred to as the ―maximal projection‖ of X (since it does not project any further). The version of X′-theory presented in Chomsky (1970) was in a preliminary form, and there certainly remained details to be worked out more fully. However, it is also true that all the crucial and fundamental insights of X′-theory were already presented in this study and have been subject to little substantive change in the following years. ―adjuncts‖ (modifiers) Chomsky‘s (1986b) [barriers] version of X′-theory has the following characteristics. First, it includes two ―non-lexical‖ categories, I and C, as 7

qxls

SN

members of ―X‖ relevant for X′-theory, so that a full clausal structure is now in full conformity with the principles of X′-theory and ―sentences‖ are no longer exceptions to the theory, a great improvement over earlier versions of X′-theory for which ―sentences‖ have always been treated as exceptions. Second, X′theory is now parametrized in accordance with the general guidelines of the principles- and-parameters approach, and the theory no longer specifies the linear ordering of elements in the scheme. An obvious point in Chomsky‘s (1986b) version of X′-theory that calls for further improvement is the incomplete parallelism it expresses between noun phrases and clauses/sentences. The source of the problems is the fact that in (27a),[Q : SN] all the ―arguments‖ (subject and object) are located within the maximal projection of a single head destruction [ennemy's /destruction/of the city), while in (27b), [Q : proposition] subject and object are split in two different projections Two proposals were made in the mid- to late 1980s which played important roles in resolving these problems. They are (i) the ―DP-analysis‖ (Fukui and Speas 1986, Abney 1987; see also Brame 1981, 1982), and (ii) the ―PredicateInternal Subject‖ Hypothesis (see Hale 1978, Kitagawa 1986, Koopman and Sportiche 1991, Kuroda 1988, among others, for various versions of the ―VPInternal Subject‖ Hypothesis; The DP-analysis claims that ―noun phrases‖ are in fact ―determiner phrases‖ (DP) headed by the head D which takes a noun phrase as its complement the DP-analysis of noun phrases received much justification from the semantics of nominal expressions (a similar analysis had in fact been assumed in Montague semantics before the syntactic DP-analysis was proposed). Thus, this analysis has become more or less a standard analysis of noun phrases and is assumed in much current literature. the DP-analysis provides a ―two-story‖ structure for noun phrases that looks quite similar to the structure of sentences: in both structures, a nonlexical category (I in a sentence, D in a noun phrase) heads the whole phrase, taking a complement headed by a lexical category (V in a sentence, N in a noun phrase). Given the DP-analysis, then, the parallelism between sentences and noun phrases becomes much more visible and easy to capture than in the traditional analysis of noun phrases. there seem to be two apparently conflicting sets of evidence regarding the status of the subject in a sentence: one type of evidence (most of which has to do with 8

qxls

SN

theta-theoretic considerations) indicates that the subject should be inside the verb‘s projection, while the other type of evidence (having to do with Case, agreement, government, etc.) suggests that the subject must occupy [Spec, I″]. The ―VP-Internal Subject‖ Hypothesis was proposed mainly to reconcile these two types of evidence. The crucial and novel part of this hypothesis is the movement process that raises the subject (which is generated inside the verb‘s projection) to [Spec, I″]. This movement is driven by the need for Case assignment. Thus, the subject of a sentence is generated in [Spec, V″] (in some versions of the VP-Internal Subject Hypothesis, not in others), and then, is moved to [Spec, I″] in order to receive Case in that position. Combining the DP-analysis and the VP-Internal Subject Hypothesis (thus making the latter the ―Predicate-Internal Subject‖ Hypothesis), we have the following completely parallel structures for noun phrases and clauses/sentences (Fukui and Speas 1986) The subjects in both noun phrases and sentences are generated within the projection of the lexical category (N in a noun phrase and V in a sentence), receiving a theta-role in their original positions, and then are raised to the Spec positions of associated non-lexical categories (D in the case of noun phrases, I in sentences) to receive Case (genitive in noun phrases, nominative in sentences). [je ne comprends pas que le genitif plutot que le determinant soit en Spec, le genitif est un argument interne du nucleurs lexical, tout comme l'objet]  The integration of the DP-analysis and the Predicate-Internal Subject Hypothesis was based on the following ideas about the lexicon as it relates to syntactic computation. (See Fukui and Speas 1986, Abney 1987, for some preliminary discussions; see also Fukui 1986 for further discussion on this and related issues.) Items of the lexicon are divided into two major subtypes: lexical categories and ―functional‖ categories. The latter types of category roughly correspond to the traditional non-lexical categories, renamed in consideration of their nature. Lexical categories have substantive content, and include nouns, verbs, adjectives, etc. They typically enter into theta-marking. Functional categories do not have substantive content, and do not enter into theta-marking (although they do have other feature structures, including categorial features, agreement features, etc.).

9

qxls

SN

Lexical categories play an important role in interpretation of linguistic

expressions, and indeed, most of the items in the lexicon belong to this type. Functional categories, on the other hand, do not play a comparable role in interpretation of linguistic expressions; their role is largely restricted to ―grammatical‖ (or ―computational‖) aspects of linguistic structure (although some of the proposed functional categories, e.g. I and D, may sometimes function as operators, bearing some ―semantic‖ import). These categories constitute a small (and often closed) set, which include C, I, D (assuming the DP-analysis), and a few others. Lexical categories bear semantic features, including, in particular, features

having to do with theta-roles (―theta-grids‘‘ in the sense of Stowell 1981). They assign (or ―discharge‖) theta-roles/features associated with them to other phrases, thereby forming larger structures that embed them. Functional categories do not bear theta-roles. Their role is largely restricted to purely formal and computational aspects of linguistic structure such as marking grammatical structures (nominals and clauses) or triggering movement operations. More specifically, some functional categories (functional heads) bear ―agreement features,‖ and these agreement features attract a maximal projection to their neighborhoods (their specifier positions), in order for the latter to agree with the former. numerous ―new‖ functional categories were proposed in the late 1980s, achieving tremendous descriptive success, although from an explanatory point of view, it was clear that the class of possible functional categories has to be severely restricted in a principled way (Fukui 1988, 1995; see also Chomsky 1995a for a ―Minimalist‖ critique of functional elements). The total elimination of X′-theory was proposed and carried out by Chomsky‘s (1994) ―bare phrase structure‖ theory (see also Kayne 1994 for a different approach). The bare theory is couched within the ―Minimalist program‖ (Chomsky 1993), according to which all the principles and entities of grammar must be motivated and justified either by the properties of two ―interface representations,‖ LF and PF, or by considerations of economy bernstein handbook baltin : bernstein The DP Hypothesis: Identifying Clausal Properties in the Nominal Domain Morphological evidence for DP 10

qxls

SN

Abney (1987: 37–53) discusses languages in which the agreement morphology in the clause and the noun phrase match in terms of both the type of agreement expressed and the manner of expression.[exemple maya ci-dessus] In other words, what is observed in these languages is that a possessed noun agrees with its subject in the same way, and with the same agreement morphology, as a verb agrees with its clausal subject. A language falling into this category is Yup‘ik, a Central Alaskan Eskimo language. Consider the examples in (2) (from Abney 1987: (24), 39). In Yup‘ik, both the verb and its subject are marked with matching ergative case, expressed via an identical agreement suffix (-t in this instance), as illustrated in (2a).2 Similarly, a noun and its possessor are marked for agreement and the morpheme involved (i.e., -t), as illustrated in (2b), matches that found in the clause: (2) angute-t kiputa-a-t man-Erg (Pl) buy-OM-SM ―the men bought it‖ (3) angute-t kuiga-t the man-Erg (Pl) river-SM ―the men‘s river‖ Matching nominal and clausal agreement morphology also characterizes Mayan languages. [cf. becquelin & monod ci-dessus] [trop facile avec ces langues = ergatives à genitif] Hungarian, a nominative/accusative language, also exhibits identical agreement affixes on nouns and verbs. The data in (3) (Abney 1987: (36), 44, data drawn from Szabolcsi 1983) illustrate the Hungarian nominal agreement pattern, where case is expressed on the possessor and the head noun agrees with the possessor in person and number. In (4) (Szabolcsi 1983: (4), 90), I have illustrated the parallel subject agreement pattern in the clause, where the sentential subject is marked for case and the verb displays number and person agreement with the subject: (4) az én-ø the I-Nom ―my guest‖

vendég-e-m guest-Poss-1Sg

(5) a te-ø vendég-e-d the you-Nom guest-Poss-2Sg 11

qxls

SN

―your guest‖ (6) (a) Mari-ø (the) Mary-Nom ―Mary‘s guest‖

vendég-e-ø guest-Poss-3Sg

(7) Mari-ø alud-t-ø Mary-Nom sleep-Past-3Sg ―Mary slept‖ [finno ougrien est suspect d'ergativophilie, perrot] As Abney discusses and illustrates, Turkish also displays DP-internal agreement (patterns and data Abney examines are from Underhill 1976, Kornfilt 1984). In this language, the possessor displays genitive case and the head noun agrees in number and gender with the possessor. Although Turkish nominal agreement morphology is not identical in form to the corresponding verbal agreement morphology, Kornfilt has shown that both nominal and verbal agreement morphology licenses pro-drop, a property which is apparently also found in the other languages discussed above [je ne vois pas le rapport] [c'est moins clair déjànon ergatif] D-position = the locus of referentiality. [interessant pour DP dans omnipredicativité] proper names raise to the D-position [c'est pour ça que souvent pas de determinant avec les noms propres] 2.1.2 classe (lexicale) du nucleus 2.1.2.1 nucleus PRN a more accurate characterization of most pronouns is that they take the place of noun phrases. In many languages, it may be difficult to distinguish pronouns from nouns except on a semantic basis. 2.1.2.2 nucleus nom propre Names, and to a lesser extent pronouns denote unique referents, no further narrowing of their referential domain is necessary. Consequently, they can only take non-restrictive modifiers e.g. relatives) 12

qxls

SN

English permits noun phrases consisting of just a noun when that noun is a proper noun (I like Pat), 2.1.2.3 nucleus deverbal nominalisations : différence entre les dependants dans les SN à noyau nominal primitif et idem à déverbal; mais pas recuperation des participants 2.1.3 nucleus lexical zero 2.1.3.1 uniquement dépendants même forme que SN mais pas de nom (8) Your car is nice, but John’s is nicer One approach to such noun phrases lacking nouns is to analyse them as involving ellipsis of a head noun, that is, as involving a noun that is present at some level of structure but which is not expressed overtly. One argument that is given for such an approach is that when a speaker uses noun phrases of this sort, it is normally the case that it is clear in the context what noun could have been used. certains SN sans nucleus ne sont pas un SN avec nom moins le nom : noms abstraits en espagnol the article lo cannot be used with a noun, suggesting that any analysis in terms of ellipsis is problematic lo grande art large ‗the large [thing]‘ approche pronominale : Another possible approach to such cases with adjectives but no nouns is to say that the adjective is functioning as a noun in such cases. However, such an approach confuses word class with grammatical function. Such an approach is motivated if the phenomenon is lexically constrained, but not if it is productive for all members of a class. Treating these adjectives as nouns is analogous to saying that music in English music teacher is an adjective because it is modifying a noun, rather than simply saying that English allows nouns to modify nouns. 13

qxls

SN

2.1.3.2 relative sans tête même forme qu'une relative mais pas de nom relativisé : figées (9)

qui m'aime me suive

(10)

honni soit qui mal y pense

actives (11)

qui l'a fait le fera encore

(12)

elle n'est pas qui vous croyez

faire la difference avec nominalisationcriteres de finitude; exemple de confusion : las cláusulas relativas en tapiete se construyen por medio de sufijos nominalizadores que, además, cumplen la función de marcar un grupo de construcciones subordinadas como cláusulas de complemento (CC)  la forme est celle d'une nominalisation, la fonction est celle d'une relative ou d'une completive  les appeler comment? pour qu'il y ait relative sans tete, il faut montrer qu'il existe des syntagmes nominaux à noyau lexical zero (le bleu), vs. nominalisation; *le de Perpignan / ok el de Perpiñan mexico 2010 : (13) le voy a tomar [lo [que es [una copia de su pasaporte]]]  lo prouve qu'il n'y a pas de NOM Ø dans ce syntagme, et que c'est l'article qui est nominalisé 2.2 dependants arguments / modificateurs

14

qxls

SN

la distinction {restriction de la reference / modelage de l'entité} ne se superpose pas à {modificateurs grammaticaux / modificateurs lexicaux} restriction : des constituants nominaux dans le rôle de génitif [argument ou adjoint], dont la signification de base est que le référent du nom dont ils dépendent se distingue des autres référents potentiels du même nom par son appartenance à la sphère personnelle de l'individu auquel réfère le constituant nominal dans le rôle de génitif ; par exemple, dans J'ai été très intéressé par [le livre de Jean] , le constituant [le livre de Jean] peut se référer à un livre que m'a prêté Jean, à un livre que Jean a écrit, à un livre dont Jean parle tout le temps, et de manière générale à tout livre susceptible d'être considéré comme ayant à un titre ou à un autre une relation privilégiée avec l'individu Jean. taille du Sn en fonction des dépendants : anglais : (14) Richard Burns scored [his [first [[FIA [World [Rally]]] championship] victory]] in the… (15) phonemic diversity supports [a [serial [founder [effect]]] model [of [[[language] expansion] [from [Africa]]]]] venezuela (16) Anticipo a Usted [ mis gracias [ por la atención [3que conozco le asignará a la solicitud [que por este intermedio le formulo]]]] 2.2.1 arguments les arguments à nucleus lexical sont eux-même des SN (17) [the king of England]'s will genitif distinguer deux types de genitif : argument / adjoint - localisation marcas de dependencia - sobre el núcleo (Iquitos: la mesa su-pata) - sobre el dependiente (Jean's hair)

15

qxls

SN

sur head et sur dependant : quechua1 (18) Alex-manta bolsa-n Alex-ablativo bolsa-3singulier ' a bolsa do Alex ' dans marcas de dependencia sobre el núcleo, the possessor noun phrase can generally be left out  le SN 'possesseur' est un argument du nom nucleus (l'affixe possessif = accord) ou un adjoint appositionnel (affixe possessif = argument)?  problematique de l'omnipredicativité et le statut des SN; aussi : = ou  personne sur verbe common type of genitive construction is one with no marking of the relationship at all, where the possessor and possessed noun are simply juxtaposed valence du nucleus : alienable / inalienable : constructions simples / complexes, prinaires / derivées, nature des 'classificateurs genitivaux'  nomes genéricos relacionais; structure du SN avec NGR : apposition / hypotaxe; classe des NGR (1 kat, jê, à une vingtaine panare)2 ; wayampi 2 animal domestique, animalgibier (19) paje lima pajé

malakaja e-kalej

AnimalDeEstimação

gato

1SINGULARII-arranhar

' o gato do pajé me arranhou '

question des classificateurs numeraux chinois & maya : on dit souvent que classificateurs s'affixent à numeral pour donner un modificateur du nom, genre [trois-ObjetRond + oranges]; Gela (Crowley (2002)), an Austronesian language of the Solomon Islands (20) e tolu num three ‗three pigs‘

na bolo art pig

idem wayampi (21) ilute quatro

ena

nuli

a-pyy-ta

recipiente

arroz

1SINGULAR-comprar-FUTURO

' vou comprar quatro pacotes de arroz '

1

2010 Pfänder, S.: Gramática mestiza, La Paz: Academia Boliviana de la Lengua/Editorial Signo

2

voir copin dans 16

qxls

SN

Relational nouns, such as uncle or sister, are known to possess an internal argument slot, which may be left unsaturated in plurals, leading to a reciprocal interpretation (Eschenbach 1993, Hackl 2002, Staroverov 2007): (22) three sisters walked in complétives : avec ou sans genitif (23) la sospecha de que nunca volvería inglês: só o subordinador that, sem o genitivo structure de syntagme à NRG deux notions de prédicat : logique classique (aristote) sujet / predicat, qui se réduit à thème - rhème : proposition = poser une entité, y rapporter de l'information nouvelle predicatrheme calcul des predicats (partie de la logique mathématique de frege3) : expression f(x) = fonction f prend une valeur particuliere selon la valeur de la variable x, son argument (on dit que f s'applique à x)fonction doit être saturée = l'argument prend une valeur, e.g. jean; predicat dans ce sens = expression de vant être saturée (à n places de variables, qui reçoivent chacune une valeur particulière)  predicat-valence copin : NGR est le nucleus parce qu'il est un prédicat saturéchangement ordre; moi : non, le N à droite est le predicat, structure de proposition avec sujet et predicat : notion de modifier's argument (ça vient d'informatique; on le trouve dans marantz qui a dû le prendre dans grammaire catégorielle4, Lexical Meaning in Context: A Web of Words Nicholas Asher; exemple de red 3

reduire la methematique a la logique

4

The basic ideas of categorial grammar date from work by Kazimierz Ajdukiewicz (in 1935) and Yehoshua Bar-Hillel (in 1953). In 1958, Joachim Lambek introduced a syntactic calculus that formalized the function type constructors along with various rules for the combination of functions. Montague grammar uses an ad hoc syntactic system for English that is based on the principles of categorial grammar. Although Montague's work is sometimes regarded as syntactically uninteresting, it helped to bolster interest in categorial grammar by associating it with a highly successful formal treatment of natural language semantics. 17

qxls

SN

red est un predicat-valence, il a une structure argumentale, comme un VRB il a un argument : red(apple); red peut fonctionner comme prédicat-rheme, il décrit une manière d'exister comme information nouvelle rapportée à une entité (eventuellement pourvue d'un referent) : (24) the apple is red red peut fonctionner comme modificateur dans un SN : red apple, il continue à avoir un argument, apple, qui est le nucleus du SN; mais dans le SN red n'est pas exactement la description d'une manière d'exister comme information nouvelle rapportée à une entité apple il délimite le sous-type d'entité apple, c'es en ça qu'il est un modifieur un argument de modificateur ressemble au nom dominant d'une relative; (25) [the colleague [you invited]] is calling colleague est le nucleus du SN, [you invited] est le modificateur; pourtant colleague est un argument de [you invited] = objet; essayons anglais bis sans that/who(m) (26) [the colleague [invited me]] is calling il y a plein de langues comme ça (sans maraque de subordination), pas seulement en australie = colleague est le nucleus du SN, [you invited] est le modificateur; pourtant colleague est un argument de [you invited] = sujet avec predicats nominaux (27) [the colleague [(is) an astronomer]] is calling (28) [ao gato [b(é) AnimalDeEstimaçãoDe [cpajé]c]b]a me arranhou

18

qxls

SN

gato est le nucleus du SN [a...]a, et [b...]b est le modificateur; pourtant gato est un argument de [b...]b = sujet; en outre, AnimalDeEstimaçãoDe est le nucleus du SN [b...]b, et pajé est l'argument interne de [b...]b5 ordre : contrairement à ce que je dis en 2005 sur base kat et panara, il n'y a pas isomorphisme d'ordre entre sujet - predicat proposition  [nucleus [NGR...] = wayampi; limite au parallèle 2.2.2 modificateurs Restrictive modifiers are used to narrow down the domain of reference, in this way resembling definite determiners. the tall man opened his mouth Non-restrictive modifiers are used to enrich the description of the referent ça traverse la distinction grammatical / lexical 2.2.2.1 determinants tous des Restrictive modifiers Many linguists use the term determiner for definite and indefinite articles, as well as other words like demonstratives & all languages appear to have words that we can call ‗demonstratives‘. two-way contrast in terms of distance from the speaker [3 espagnol] un ensemble de‘déterminants’définis de façon strictement syntaxique par la propriété de permettre au nom d’accéder sans restriction au statut de constituant syntaxique Many traditions categorize articles as a type of determiner, this class including as well such words as demonstrative modifiers of nouns (as in English this book). However, languages differ in whether articles and demonstratives belong to the same word class. In English, they do, appearing in the same determiner position at the beginning of noun phrases; in English, one cannot have both an article and a demonstrative (*the this book). But in many other languages, articles and demonstratives are separate word classes. In Fijian, for example, the article precedes the noun while the demonstrative follows

5

noter que copin appelle [b...]b relative nominale

19

qxls

SN

As noted above, English determiners follow the definite quantifier slot. Various occupants of the determiner slot in English are: (14) a. Deictic that horse, this horse b. Definite the woman c. Indefinite a girl, some children e. Possessive [je doute] démonstratif / anaphoriques / interrogatif (quel ? )sélection parmi plusieurs référents potentiels renforceurs de demonstratifs : distal / proximal démonstratifs = origine des articles et des pronoms de 3ª surtout pour non animés origine articles : definis : they start as demonstratives, get extended to anaphoric usage and then finally (in some cases) can be used nonanaphorically as well. Demonstratives that are used as definite markers are often restricted to anaphoric usage; the most common source of definite marking involves the re-grammaticalization of a distal demonstrative indefinis : In some languages, the form of the indefinite article is the same as that of the numeral for ‗one‘, but the syntax is different [position] Haspelmath, Martin. 1997. Indefinite Pronouns. Oxford: Clarendon Press. : Il est fréquent dans les langues qui ont un article indéfini dérivé du numéral Un que ce numéral soit également à la base d'un pronom indéfini (Haspelmath 1997, Heine 1997, Heine & Kuteva 2002 : 221). Le plus souvent le pronom n'est pas identique au numéral. Celui-ci s'associe à un autre élément (Haspelmath 1997: 183-84, Heine & Kuteva 2002: 221), un quantifieur, comme en français ―quelqu'un― et en anglais ―someone‖, ou un pronom interrogatif, comme en pashto yaw (Un) cok (qui) ‗somebody‘ par exemple, pour la formation du pronominal. 20

qxls

SN

2.2.2.2 adjectifs identifier existence : absurde, ethnocentrique = 'semantic adjectives' qui sont selon des langues des adjectifs ou des noms ou des verbes  difference adjectif / participe (même fonction, origine synchronique différente : lexique / verbe)  extension de la classe : sik ~20; wayampi 1 affinités synchroniques / diachroniques avec classes verbes / noms (langues où aucun rapport = plutôt rares); sik : verboides + pe2.2.2.3 adjoints distinguer d'arguments, même pour genitif : interiorvalência2 do corpoargumento / estradavalência1 de terraadjunto il peut y avoir une structure hierarchique (29) [anomnucleus [bargument]b]a [adjoint] sans marque adpositionnelle : Nouns as modifiers (sous-type d'entité) jazz band, Dixieland rock the university president selection committee each modifying noun carries its own primary lexical stress, and thus retains its independence as a lexical word. But modifying nouns may eventually fuse with their head to yield noun compounds, in English have a characteristic stress pattern: The primary wordstress is invariably placed on the first noun—the modifier mail-man [ patrons de lexicalisation] différents types de sequences NOM NOM en anglais (image ici)

21

qxls

SN

that modifying nouns, even when not compounded, must be adjacent to the head noun probably reflects the fact that nouns typically code more inherent, generic, durable qualities avec marque adpositionnelle : tea for two Languages differ as to whether they allow adpositional phrases or noun phrases with oblique cases to modify nouns  in Lezgian (Haspelmath 1993) To express what English would express by means of a prepositional phrase modifying a noun, Lezgian must place the modifying phrase in a relative clause with an appropriate verb, such as the verb meaning ‗be‘ leque de marcas oblíquas possíveis (comparar com adjuntos oracionais) adjetivos regem caso às vezes: indiferente ao sentimento, limpo de obstáculos 2.2.2.4 head shift le nucleus est un nom sans referent, abstrait, qui explicite une propriété, qualité, d'un nom qui est formellement son dépendant  en diachronie, le dépendant formel reprend des proprietés de nucleus, et le nucleus formel perd des propriétés de nucleus; (30) beaucoup de personnes l'ignorent étapes initiales ou intermédiaires (voir dixon sur jarawara ci-dessus, mais avec des noms concrets) (31) um belo de um truque (32) uma beleza de carro está parado em frente da casa (ok?) (33) un montón de casas están desocupadas 22

qxls

SN

(34) un tas de livres sont tombés (35) a lot of books are sold probablement origine des classificateurs 2.2.3 dependants zero noms nus uniquement, ie. avec nucleus qui admettent dependants (ça ecarte pronoms, noms propres) English permits noun phrases consisting of a mass noun (I like milk) or the plural of a count noun (I like flowers), this is not possible with the singular of count nouns (*I like flower); in these cases, English requires some sort of determiner (I like this flower, I have brought you a flower). 2.3 cohésion indiquent la cohérence interne du SN : hierarchies syntaxiques, accord, ordre, compatibilités, syncrétismes, head attraction, ligateurs 2.3.1 accord grammatical agreement within the noun phrase — in gender and number, case, or definiteness — can serve as a means forbinding NP constituents together; that is, of signalling the noun-modifier relation gender (ou classe) may be expressed on nouns, determiner elements, adjectives, and/or other modifiers. e.g. adjectifs latins genre, nombre et cas Of these, only the first — class/gender — is an inherent feature of the head noun. The others are grammar-coded semantic or pragmatic further specification of the particular NP. Hebrew may also serve to illustrate definite agreement— a feature of the NP rather than of the noun— across many modifiers in the NP

23

qxls

SN

Case-role, again an NP feature rather than a noun feature, can also form the basis for ‘agreement’, as in Latin (Palmer 1954): nombre dans beaucoup de langues n'est pas marqué sur le SN lui-même mais ailleurs (predicat, modifieurs [maya, portugais bresil] 2.3.2 ordre ordre dans SN doit etre plus rigide que dans phrasemoins de pragmatique order is quite rigid in English, more flexible in Japanese. attention a projection de la sémantique pour identification de l'ordre. une merveille de film male chimpanzee  female chimpanzee

chimpanzee male  human male

one day

air force one

2.3.2.1 nucleus & arguments (+ alignement) : anglais genitif saxon pré-N (germanique), genitif normand post-N ( roman) (36) the ennemy's destruction of the city ordre genitif-nom peut changer selon que le genitif est pronominal ou nominal Daly, language of north Australia (Tryon (1970b)), exhibits the opposite pattern, with the lexical genitive preceding the noun (at least for inalienable possession), as in (62a), and the possessive pronoun following the noun, as in (62b). (37) Micky piyamerr Micky hair G N ‗Micky‘s hair‘

piya ngany head my N Poss ‗my head‘

(38) ma photo de john lennon 24

qxls

SN

2.3.2.2 nucleus & modifieurs Simple modifiers—bound morphemes or lexical words—precede the head noun in English. Complex modifiers—phrasal or clausal ones—follow. Post-nominal modifiers in English are large and syntactically more complex pre-nominal: quantifiers, partitive-definite (some of the people) and indefinite (some women) vs. Modifiers that follow the head noun are the plural morpheme (Pl), relative clauses (Rel), prepositional phrases (PP) or noun complements (N-Comp). In Fijian, for example, the article precedes the noun while the demonstrative follows interrogative modifiers with meanings such as ‗which‘, ‗what sort of‘, ‗how many‘, and ‗whose‘. In most languages, such elements occur in the same place in the noun phrase as corresponding noninterrogative words, but there are exceptions. For example in Ocotepec Mixtec, most modifiers, including demonstratives and adjectives, follow the noun, as in (128a), but the interrogative modifier meaning ‗which‘ or ‗what‘ precedes the noun paralleles entre adverbe et adjectif (position) dans X-barre & DP : langues romanes : the position of the noun relative to the adjective(s) cannot be due to variation in the position of the adjective, but rather, it must be a result of variation in the position of the noun.[+correlats phonologiques, comme accent] (le nom monte a la position de specificateur de D) adjectifs non restrictive adjectives precede the head noun and restrictive ones follow : un hombre pobre ‗a poorman ‘ (R) un pobre hombre ‗an unfortunate man‘ (NR) un hombre muy pobre ‗a very poor man‘ (R) (39) *un muy pobre hombre (40) une belle femme 25

qxls

SN

une grande femme *une noire femme In some languages, the position of cardinal numeral and ordinal numeral are different Article and noun The order of article and noun exhibits a correlation with the order of verb and object, although the correlation is weaker than most of the other correlations discussed in this chapter. In particular, it is more common for the article to precede the noun inVOlanguages, as in English (the dog) and the Fijian example in (83a), but to follow the noun in OV languages, as illustrated in (83b) by the indefinite article in Kobon, an East New Guinea Highlands language Noun and relative clause Almost all VO languages place the relative clause after the noun, among OV languages, both orders are about equally common ; It is often mistakenly thought that the order of adjective and noun correlates with the order of object and verb, but it is now known that this is not the case (see Dryer (1988, 1992)). It is often thought that OV languages tend to be AdjN and that VO languages tend to be NAdj, but it turns out in fact that this is not so, that NAdj is somewhat more common than AdjN among both OV and VO languages. Part of the source of this problem is that the languages in the sample used by Greenberg (1963) Demonstrative and noun Demonstrative modifiers of nouns, like adjectives, are common either before the noun or after the noun among both OV and VO languages, though in both types of languages DemN order is slightly more common; demonstratives do not exhibit a correlation in their position with the order of object and verb, Numeral (numeral) and noun Both NumN and NNum order are common among OV and VO languages nom nom nom comme modificateur : penchant pour l'ethnocentrisme : If a language is OV, then it will be [NOUN] + [PLURAL NOUN] [parce que plural noun est juste 'groupe, ensemble, quartète, turma, gang, galera' donc on a un ordre normal [modifieur + nom], congru avec OV] 2.3.2.3 dependants & dependants bhatt internet : en raison de structure (29)

26

qxls

SN

if an NP contains both a complement PP and an adjunct PP, the complement PP should precede the adjunct PP. This prediction turns out to be true. (41) the student [of Physics] [with long hair] (42) *the student [with long hair] [of Physics] formulation generale doit être : si argument et adjoint du même côté, argument plus près du nucléus adjectifs entre eux : An adjective will be placed closer to the noun stem if it is: •more central to the meaning of the noun; •more inherent, durable quality of the noun; •more generic (rather than specific) information; •more given (rather than new) information; •non-restrictive (rather than restrictive). dryer word order shopen I One cross-linguistic generalization governing the order of Dem, Num, and Adj is that when all three appear on the same side of a noun and one order is preferred, the demonstrative typically is furthest from the noun and the adjective closest, (construction du sous-type d'entité plus intime, construction de la reference mois intime) if a language places both the demonstrative and the numeral before the noun and the adjective after the noun and if there is a preferred order for the Dem and Num, that order will typically be Dem before Num. correlations entre [XY]SN et verbe/objet : oui OV genitive–noun standard–marker standard–adjective noun–article relative clause – noun noun – plural word

VO noun–genitive marker – standard adjective – standard article–noun noun – relative clause plural word – noun 27

qxls

SN

non adjective / noun demonstrative / noun numeral / noun degree word / adjective nom & nom modifieur problème: l'ordre donne le sens dans (43) end game / game end (44) machine sex / sex machine il doit le donner aussi dans (45) planet earth / earth planet (46) human male / male human6 (47) awiri pebito / awiri pesorowato  l'ordre ne varie pas, c'est toujours modifieur + nucleus 2.3.3 compatibilités The best evidence that all these determiners belong to the same syntactic slot in English is theirm mutual exclusivity [critere de dryer pour classe de determinants] other languages the definite article may coexist with otherdefinit e determiners, as in Hebrew adjective may be preceded by an adverb or intensifier, the slot itself is technically that of an adjectival phrase article et possessif, article & demonstratif (in)compatibilités peuvent conditionner ordre : dryer dans shopen 6

'battle scenario: male human vs. male chimpanzee'

28

qxls

SN

Aari (Hayward (1990)), an Omotic language spoken in Ethiopia, demonstratives more commonly follow the noun, as in (177a), but they only precede the noun if the noun is followed by a numeral, Himmelmann (2001)7 observe que les langues du monde ont tendance à éviter toute cooccurrence de l'article indéfini et du numéral un même lorsqu'ils ne sont pas identiques, à l'exception du Sinhala et du Turc. En Sinhala, le numéral et l'article ne sont pas identiques et peuvent figurer librement dans le même contexte (48). En turc l'article et le numéral sont formellement identiques ; ils peuvent être combinés, mais ils occupent une place différente, l'article se place entre le nom et l'adjectif (49). (48) pota-k pot eka-k book.SG-INDÉF book.pl. one-INDÉF ―A book‖ "One book" (Sinhala) (49) büjük bir tarla big one field ―A large field‖

~ ~ ~

bir büjük tarla one big field ―One large field‖ (Turc)

2.3.4 syncretismes e.g article roman (definitude, genre, nombre, and, in some languages, case) Ilocano (Rubin (2000, p.c.)), an Austronesian language of the Philippines, has a set of eight words that vary for number, for case (core versus oblique) and for whether the noun they occur with is a common noun or a proper noun. 2.3.5 head attraction kat, movima {Kopa-nagenitif bakon}  /Kopa # nabakon/ (50) {kay-a-poj comer-DIRETO-CAUSATIVO

dedo de Kopa

us

itila:kwa}

as

pa:ko

ARTIGOMASCULINO

homem

ARTIGONEUTRO

cachorro

' the man feeds the dog '  /kayapojaus # itila:kwa/8 7

Himmelmann, Nikolaus P. 2001. ‗Articles‘. M. Haspelmath et als (eds), Language Typology and Language Universals: An International Handbook. Berlin / New York: Mouton de Gruyter. 831-841. 8

a avant us : epenthétique

29

qxls

SN

2.3.6 linkers morphèmes dont la seule fonction est d'indiquer l'adjacence immediate : mandarin, TG 2.3.7 SN discontinus 1) SN discontinus : Baker mohawk, hale warlpiri, vidal pilaga 9 et 2) blake 1983 145 dans rijkhoff 19 : en kalkatungu, 'there are in fact no noun phrases [...] where an argument is represented by more than one word we have nominals in parallel or in apposition [...] each word is a constituent of the clause' some scattering of members of the noun phrase is allowed even in a language as rigidly ordered as English extraposition of restrictive REL-clauses (givon) (51) A woman came in next who spoke no English Walpiri (Hale 1976): Adjacent, optional case agreement (52) tjantu wiri-nki tji yalku-nu dog big-erg me bite-past ‗The big dog bit me‘ Non-adjacent, obligatory case agreement (53) tjantu-nku tju yalku-nu dog-erg me bite-past ‗The dog bit me, the big one‘

wiri-nki big-erg

distinguer deux interpretations : SN discontinus = 1) vs. déterminant utilisé pronominalement et instituant un SN par lui -même = 2) apposition Fox (Algonquian), a modifier may be expressed first, while its head noun is post-posed (Dahlstrom 1987): 9

Constituyen ejemplos de constituyentes escindidos, donde el demostrativo en (1a) está separado del núcleo y en (1b) [exemple] ‗las uñas de esa mujer‘, la parte de la frase nominal que designa al poseedor (la mujer) dislocada con respecto al poseído (sus uñas)

30

qxls

SN

(54) neswi e.h=ayo.wa-chi three use-3pl/asp ‗They use three songs‘

nakamo.nani songs

a modifier is used as an anaphoric pronoun, and the post-posed head noun is then right dislocated [ ce n'est pas de la discontinuité] = 2) apposition quantifier float in French, illustrated in the examples in (6) (examples adapted from Sportiche 1988): (55) a. Toutes les filles ont reçu les notes b. Les filles ont toutes reçu les notes. [on peut dire avec une seule unité prosodique ou non : c. Les filles ont, toutes, reçu les notes. alors que d. * Toutes, les filles ont reçu les notes  = 2) apposition idem foley sur yimas dans rojkhoff 20 : SN limité à 2 elements, nucleus et 'modificateur'; s'il faut un modificateur de plus, ce dernier est nominalisé par une marque d'accord et institue un nouveau SN, appositif = 2) apposition nama hottentot choisit entre mettre modifieurs dans SN ou les nominaliser et en faire des SN séparés (56) [kinii-di book-3PLURIELFEMININ

ne] [!nona-di]

[Gombate di-di]

ces

Gombate

trois-3PLURIELFEMININ

CeluiDe-3PLURIELFEMININ

'ces trois livres de Gombate'

[à partir de là, la contiguité ou non contiguité de ces trois SN doit être une question de pragmatique] = 2) apposition 2.3.8 apposition

31

qxls

SN

 la coreference n'est pas un critere decisif pour l'apposition : wayampi copin; coreference avec apposition  chaque element pourvu de posposition (=warlpiri deuxième exemple) (57) [mote-ku] motor-COLETIVO

le

[o-jea

mae-ku]

le

o-jmona

POSPOSIÇÃO

3II-quebrar

NOMINALIZADOR-COLETIVO

POSPOSIÇÃO

3I-roubar

' ele roubou os motores, os que tinham quebrado '

coreference dans un même syntagme  une seule postposition (=warlpiri premier exemple)  hierarchisation syntaxique (58) [[mote-ku] motor-COLETIVO

o-jea

mae-ku]

le

o-jmona

3II-quebrar

NOMINALIZADOR-COLETIVO

POSPOSIÇÃO

3I-roubar

' ele roubou os motores que tinham quebrado '

= 'ele roubou [os que tinham quebrado [que eram motores]]' (copin donne une hierarchie differente : nucleus = moteur) 3 fonction 3.1 sémantique modelage de type d'entité 3.1.1 qualités adjectif (discours frivole) : semantique des adjectifs nom (hombre araña, femme enfant, papier toilette)  tendance à la lexicalisation 3.1.2 possession semantique du genitif (59)

ma photo de john lennon

(60)

tu veux ma photo?

alienable / inalienable comme correlats semantiques de monovalent / divalent domaine de l'inalienable : notions relationnelles; 32

qxls

SN

frontière semantique alienable / inalienable varie de langue à langue (certaines n'y mettent pas anatomie, d'autres n'y mettent pas parenté): Amazonie riche : artefact fabriqué / acquis; samoa : entités qui determinent le possessseurinalienable (roi qui gouverne X) / entités déterminées par le possesseur alienable (roi que X met sur le trone) sémantique des NGR (interaction matérielle avec possesseur colette) 3.1.3 quantification nombre nombre lexical = inherent au nucleus (c'est sa combinatoire avec sa morphologie ou avec les autres elements du SN qui détermine sa classe) individuel generique collectif dense discret exemple generiques / individuels (61) Na semana passada levei ração para porco e um radiador de trator na esteira. Tudo que precisa levar eu ponho nela e vou. (hercules) nombre grammatical distributif10 collectif singulier duel (parfois dans singulier, parfois dans pluriel) triel paucal (parfois sous produit du duel) pluriel (implication trielduel pluriel) difference entre collectif et pluriel : pluriel réplique les entités, collectif regroupe les entité en un tout doté de structure interne ('orquestre', 'file d'attente') 10

toba : sistema de número compuesto por los valores: singular y plural, junto a otras dos categorías afines: distributivo y colectivo, las cuales pueden ocurrir en un mismo nombre y junto a los morfemas de número. El número dual sólo es marcado sintácticamente

33

qxls

SN

le singulier peut être construit ( singulatif, classificateur) numeraux cardinal / ordinal Cardinal and ordinal numerals often differ in their syntax [ordre] [penser à numeraux qui sont des noms : paire, trio] Ordinal numerals are most commonly derived from cardinal numerals ; langues : cardinal numerals cannot directly modify nouns but must be accompanied by a numeral classifier c'est un modificateur ou c'est un nom?  There is a question in some languages whether the numeral modifies the noun or whether it is better to view the numeral as the head and the noun as modifier. This latter view is often suggested for languages with numeral classifiers; under this view the numeral modifies the classifier and the numeral plus classifier serves as head, with the noun as modifier. Such an analysis also suggests itself for some languages without numeral classifiers. In Rif Berber (Kossmann (2000)), spoken in Morocco, most modifiers follow the noun, but numerals precede, as in (42a); but the construction they occur in is the same as the genitive construction, illustrated in (42b), suggesting that the numeral is the head. numerals must occur with an invariant word, which is strictly speaking not a classifier (since there is only one of them), but which otherwise functions like a classifier in that its presence is required if the numeral is modifying a noun. An example of this is the word e (glossed ‗num‘ for ‗number marker‘) in Gela (Crowley (2002)), an Austronesian language of the Solomon Islands, as in (41). [ce n'est pas un number marker, c‘est l'equivalent du NGR!] (62) e tolu num three ‗three pigs‘

na bolo art pig

34

qxls

SN

quantifying words, which behave like cardinal numerals in most languages, but which in some languages do not. This includes words meaning ‗many‘, ‗much‘, ‗few‘, ‗all‘, ‗every‘, and ‗some‘parfois distribution speciale Degree word intensifiers, or adverbs, words modifying adjectives that are analogous in meaning to English words like very, more, rather, somewhat, and slightly in many languages, degree words do not behave as a grammatically well-defined class, and often some degree words precede the adjective while others follow, within the same language intensif (augmentatif / diminutif) quantification d'adjectifs (63)

la voz más baja del coro

entité comparée (voz) / base de comparaison ([voces] del coro) / operador de comparacion (bajo) / dispositif formel de comparaison (eg. morphologie cas spatiaux : ablatif ou allatif sur base de comparaison, autres) (ici adjectif + cas partitif) a nicer movie 3.1.4 genre sur les personnes, implication dans l'existence des distinctions de genre : 1  2  3 (faux en basque) genre : ne pas confondre avec les noms male / femelle fondé sur sexe vs. arbitraire  exhaustivité inherent (pez / ave) vs. attribué grammaticalement (epiceno) un / una gorila opposition equipollente ou privative (marqué / non marqué : roman (M) /arawak, arawa (non marqué F) que signifie neutre? (non animé, ou abstrait (espagnol lo)) distinction genre / classes; compatibilité dans même langue (reputés incompatibles)

35

qxls

SN

3.1.5 classes classificateurs (combinent avec noms) vs. noun classes (=accord) la nature parasitique de la classification (idem genre) : on classe pour · cohésion du SN · pistage reference par l'accord, (anaphoriques & defini en Jacaltec) · quantification (par les numeraux, la singulation), · valence-possession ( interaction possesseur-possédé : domestique, comestible) 3.1.6 saillance plus de distinctions sur plus saillants (manifestables sur modificateurs) (mais sur personne 1 & 2 moins de distinctions de genre, plus de nombre) nombre : saillance basseneutralise distinctions11 3.1.7 aspect-temps TG, Kwakw'ala, Marshalais 3.2 referentialité restriction sur la classe des individus possibles nom + nom adjectif : words meaning ‗another‘, ‗different‘, ‗same‘, ‗only‘ 'mesmo' dans ele mesmo; aussi certains adjectifs (couleur, taille) mais pas d'autres (interessant) argument : un genitif defini-référentiel restreint plus facilement la référence du nucleus : Many languages distinguish a genitive construction with a referential genitive from one with a nonreferential genitive. In Roviana (Corston-Oliver 2002), the construction with a referential genitive involves a possessive suffix on the head noun, as in (132a), while a nonreferential genitive involves juxtaposition, as in

11

jarawara pas de distinction de nombre sur non animés

36

qxls

(64) mamalaengi-na [barikaleqe voice-3sg.poss [woman ‗that woman‘s voice‘

SN

hoi] that]

(65) mamalaengi barikaleqe voice woman ‗a woman‘s voice‘ / ‗a female voice‘ adjoint locatif : el paisano aquí deixis personne inclusif/exclusif personne du SN : generalement 3º, mais we linguists, sik N-mü, nahuatl espace (source de reference) essa aqui (cantuaria) (portugais bresil perd esta) demonstratifs et leur reinforcers (distal/proximal) (espace + contexte) direccionalidad12 localisation13 referenciant hypothese omnipredicativité-non configurationalité : noms = predicats; pour occuper une position d'expression referenciante, besoin de materiel explicite; souvent article defini, parfois morphème specialisé, referentiant; souvent on n'identifie pas le referenciant (pas considération del'omnnip´redicativité) et on le prend pour un article au motif qu'il est plus ou moins onligatoire, eg. dryer dans shopen in Koromfe (Rennison (1997)), a Niger-Congo language of Burkina Faso and Mali, in which there is an article a [comme en nama], illustrated in (21), which is used with all common nouns except ones already modified by certain other modifiers, like demonstratives and numerals

12

al sintagma nominal, el toba posee un paradigma de seis deícticos que indican posición y direccionalidad, obligatorios para cada nombre 13 al sintagma nominal, el toba posee un paradigma de seis deícticos que indican posición y direccionalidad, obligatorios para cada nombre

37

qxls

SN

definitude : dryer dans shopen : Many languages have a definite article or an indefinite article but not both ; Kutenai, a language isolate spoken in western Canada and the United States, for example, there is a definite article, Less common are languages in which there is an indefinite article but no definite article [je crains qu'il ne fasse un seul paquet avec article et referentiant]; sik et kat + wayamapi (copin) seulement un indefini, pas d'indefini definite articles Most commonly, are restricted to anaphoric uses and are sometime glossed as ‗previously mentioned‘. The definite article nd-i in Ngiti (Kutsch Lojenga (1994)), a Central Sudanic language of the Democratic Republic of the Congo, illustrated in (6), is an instance of a definite article with such a restriction. (66) y`a nd-i dza this def house ‗this house (mentioned before)‘ emphase sur identité (X mesmo, X himself) negation identité (tapirapé) [non X] est arrivé 3.3 pragmatique jeu sur sens vs. reference: le pere de Marie parlant à la mère de Marie (=camera de Kuno) (67)

Marie est rentrée à trois heures

(68)

ta fille est rentrée à trois heures

defini (voir dans referentialité) source information : visible / non visible (avec demonstratifs); autres : (69) entró el dizque gran profesor

38

qxls

SN

proximal / obviatif (algonquien) : deux troisiemes personnes, hierarchisées pragmatiquement :, (potawatomi : trois troisiemes personnes)saillance pragmatique (interferent avec saillance semantique : possesseur > possédé) modalité : honorifiques, hypocoristiques (interlocuteur, troisieme personne)

39