Hostname: page-component-74656f9884-j7cmm Total loading time: 0 Render date: 2026-11-10T11:37:39.394Z Has data issue: false hasContentIssue false
  • English
  • Français

Licensing by modification: The case of French de nominals1

Published online by Cambridge University Press:  01 March 2012

ÉRIC MATHIEU
ÉRIC MATHIEU*
Affiliation:
University of Ottawa
*
Author's address: Department of Linguistics, University of Ottawa, 70 Laurier Avenue East, Ottawa, K1N6N5, Canadaemathieu@uottawa.ca
Rights & Permissions [Opens in a new window]

Abstract

The aim of this paper is to provide an analysis of the positive effect that modification has on the distribution of noun phrases in otherwise illicit environments. I focus on de nominals in French. By focusing on these nominals, whose distribution is altered by the addition of modifiers, the paper shows that modifiers can do much more than simply modify: they can change the syntactic and semantic status of a noun phrase. The licensing property of modifiers is an intriguing topic and has not been greatly discussed in the literature. I argue that modifiers can come to play the role of determiners in French as long as they are accompanied by a head de, which is the spell-out of a Cardinal head (see Lyons 1999). My proposal goes back to an old idea put forward by Damourette & Pichon (1911–1940)according to which, in modified contexts, de functions as one half of the article while the adjective functions as the other half. More generally, articles in French are seen as dual entities comprising of a specifier and a head. In the absence of the determiner les, an adjective can raise to the specifier of CardinalP. This is achieved via phrasal rather than head movement.

Information

Type
Research Article
Information
Journal of Linguistics , Volume 48 , Issue 2 , July 2012 , pp. 389 - 426
Copyright
Copyright © Cambridge University Press 2012

1. Introduction

The aim of this paper is to provide an analysis of the positive effect that modification has on the distribution of noun phrases in otherwise illicit environments. To illustrate, while (1a) is completely ungrammatical, the addition of a modifier to the noun romans ‘novels’ in (1b) changes the status of the sentence: suddenly, it is perfectly well-formed.

  1. (1)

    1. (a) * J'ai lu de romans l'été dernier. I. have read DE novels the.summer lastIntended: ‘I read (some) novels last summer.’

    2. (b) J'ai lu de bons romans l'été dernier.I.have read DE good novels the. summer last‘I read good novels last summer.’

It appears that the modifier behaves like a determiner (turning a predicate into an argument), since in lieu of a modifier the addition of les, to give des (I will assume throughout the paper that des=de+les, as is done traditionally), results in an equally grammatical sentence. This is shown in (2).

  1. (2) J'ai lu des romans l'été dernier.I.have read DES novels the.summer last‘I read (some) novels last summer.’

The idea that the modifier acts like a determiner in (1b) is reinforced by the fact, illustrated in (3), that post-nominal modifiers in French do not have a positive effect on the distribution of de nominals. Determiners in French (e.g. le, la, les, un, une, des) are pre-nominal: they can only turn a predicate (noun) into an argument if they c-command their associated nouns in the syntax (des romans ‘novels’ but not *romans des).

  1. (3) *J'ai lu de romans français l'été dernier.I.have read DE novels French the.summer lastIntended: ‘I read French novels last summer.’

By focusing on French de nominals, whose distribution is altered by the addition of modifiers,we will see that modifiers can do much more than simply modify: they can change the syntactic and semantic status of a noun phrase. The licensing property of modifiers is an intriguing topic and has not been greatly discussed in the literature. The present paper is a contribution to this line of research.

My account of the puzzle – why do pre-nominal modifiers improve the grammaticality of anotherwise ill-formed sentence – builds on Androutsopoulou & Español Echevarría's (Reference Androutsopoulou, Echevarría and Griffin2003) insight that the adjective in modified de nominals comes to fill the determiner position (Androutsopoulou & Español Echevarría's Reference Androutsopoulou, Echevarría and Griffin2003 paper will be referred to as A & EE henceforth). I will argue against their idea, however, that the adjective in such contexts raises to D0 (via head movement). Instead, I propose that the adjective raises to the determiner domain by phrasal movement. In addition, whereas A & EE's account underplays the role of de in such environments, I argue that its contribution to the well-formedness of sentences with modified de nominals is in fact crucial.

In viewing both the pre-nominal adjective and the element de as relevant to the licensing of de nominals in subject (and object) positions, my proposal goes back to an old idea put forward by Damourette & Pichon (1911–Reference Damourette and Pichon1940) . According to this proposal, in modified contexts de functions as one half of the article while the adjective functions as the other half.

The proposal is that articles in French are dual entities comprising of a specifier and a head. In indefinite contexts, de comes to occupy the head while les or an adjective in the absence of les occupies the specifier position. This gives support to the general proposal by Lyons (Reference Lyons1999) that articles can be either heads or specifiers, but I show specifically that determiners can be distributed (between a head and a specifier).

My proposal has important theoretical ramifications. First, it shows that, contra Landau (Reference Landau2007), the phonological realization of a given head is often/generally not sufficient to licence a noun phrase in subject position. Visibility of both a head and a specifier is required if a de nominal is to be properly licensed. This proposal is in tune with Giusti's (Reference Giusti and Cinque2002) Principle of economy of lexical insertion (according to which both specifiers and heads need to be visible), showcasing in passing the importance of specifiers, contra Starke (Reference Starke2001).

Second, my proposal shows that head movement is not needed in order to account for the behaviour of modifiers in de nominal contexts. This must be seen as a welcome result, since head movement has been shown to be problematic in current syntactic theory (see Chomsky Reference Chomsky and Kenstowicz2001 for details).

Third, my proposal sheds light on a long-lasting puzzle in French syntax, i.e. the categorical status of de. Some grammarians have claimed that it is a preposition; others, that it is a determiner; or even either a preposition or a determiner depending on the context (see Wilmet Reference Wilmet1986 for a summary of the literature). On my view, de is a determiner but of a special kind: it is not a full partitive or indefinite determiner, as usually claimed, but a semi-determiner (or a deficient determiner, along the lines of what Heyd Reference Heyd2003 has proposed for different facts) that requires another determiner-like element in its specifier for the sentence to be well-formed: either a modifier or les in the more traditional cases.

Finally, my proposal illustrates the flexibility of lexical categories (modifiers) to act as functional categories (determiners), showing that a strict distinction between lexical and functional categories is not well-motivated (Déchaine Reference Déchaine1993, Borer Reference Borer2005): an adjective can come to play the role of a determiner.

This is how I shall proceed. I begin in Section 2 by introducing the basic puzzle with which the present paper is concerned. Section 3 gives an analysis of modified de nominals and explains why they manage to appear in all contexts where standard indefinite phrases are licit. Section 4 argues against the allomorphy analysis. Section 5 argues against the idea that modified de nominals behave differently from non-modified de nominals because they are focused. Section 6 argues that Dayal's (Reference Dayal2004) account for modified any cannot be applied to the case of French modified de nominals. Section 7 concludes the paper.

2. A puzzle

The first noteworthy mention in the literature of the relevance and contribution of modification on the distribution and well-formedness of noun phrases comes from Romance languages such as Spanish and Italian. In the normal cases, a bare noun is licit in an object (4a), but not in a subject position (4b).

  1. (4)

    1. (a) Ha invitado escritores. Spanishhas invited writers‘He has invited writers.’

    2. (b) *Escritores han aceptado su invitación.writers have accepted his invitation‘Writers have accepted his invitation.’

However, if we add a modifier to the noun escritores'writers' in (4b), then the sentence becomes grammatical (Delfitto & Schroten Reference Delfitto and Schroten1991, Longobardi Reference Longobardi1994). To illustrate, consider the examples in (5). In (5a) franceses ‘French’ modifies escritores ‘writers’ while in (5b ) the complex modifier de mi generación ‘of my generation' modifies the noun escritores ‘writers’. Both sentences are grammatical in sharp contrast with (4b).

  1. (5)

    1. (a) Escritores franceses han aceptado su invitación.writers French have accepted his invitation‘French writers have accepted his invitation.’

    2. (b) Escritores de mi generación han acceptado su invitación.writers of my generation have accepted his invitation‘Writers of my generation have accepted his invitation.’

The special effect that modification has on the distribution of noun phrases has also been noticed with regard to English any (Dayal Reference Dayal2004). As shown by (6a), any cannot normally appear in a subject position. Traditionally, this is because an any phrase needs to be c-commanded by a proper licensor (such as negation) and can thus only appear in an object position, as in (6b).

  1. (6)

    1. (a) *Any student didn't sign the petition.

    2. (b) The students didn't sign any petition.

However, as (7) testifies, when modified, any is suddenly completely licit in a subject position (the CP who went to the meeting modifies any student).

  1. (7) Any student who went to the meeting didn't sign the petition.

In fact, even in the absence of negation, the subject position is a licit environment for any. This is shown in (8a). In the absence of a modifier, any cannot of course appear without a proper licensor such as negation, as shown by (8b).

  1. (8)

    1. (a) Any student who went to the meeting signed the petition.

    2. (b) *The students signed any petition.

A third set of facts illustrates further the positive effect that modification has on the distribution of noun phrases. As shown by Wolter (Reference Wolter and Puig-Waldmüller2007), normally a demonstrative phrase such as those employees in (9a) is scopally inert: it takes widest scope only. On the other hand, a definite description such as the employees in (9b) is ambiguous between a transparent and an opaque reading. On the opaque reading, the referent of the employees varies across John's belief worlds, but on the transparent reading it does not.

  1. (9)

    1. (a) John believes that those employees were fired.

    2. (b) John believes that the employees were fired.

However, if we modify the phrase those employees as in (10), then suddenly it is possible to obtain both a transparent and an opaque reading: the referent of those employees responsible varies across John's belief worlds. The demonstrative description is no longer scopally inert.

  1. (10) John believes that those employees responsible were fired.

The present paper focuses on a fourth set of facts: the case of French de nominals. As seen in the introduction, the bare noun in (1a) becomes licit in the sentence if modified, as in (1b). The generalization also holds for the subject position, as shown in (11).

  1. (11)

    1. (a) *De romans ont été publiés l'été dernier.DE novels have been published the.summer lastIntended: ‘(Some) novels were published last summer.’

    2. (b) De bons romans ont été publiés l'été dernier.DE good novels have been published the.summer last‘Good novels were published last summer.’

Finally, as mentioned in the introduction, while it is possible for a pre-nominal modifier to save the sentence from ungrammaticality, post-nominal modifiers cannot undo the ill-formedness of the sentence (3).

In this section, we have seen that the positive effect that modification has on the licensing of bare nominals is a general phenomenon. It is a new topic worthy of attention that will help us understand better the flexible nature of syntactic categories. The case of French is particularly interesting because it sheds light not only on the nature of adjectives and the role they can play beyond the simple case of modification, but also on the nature of determiners and their syntactic structure.

3. The account

In this section, I provide a syntactic solution to the French puzzle. The solution that I am proposing builds on A & EE's analysis of modified de nominals. My analysis is similar in that I propose that pre-nominal adjectives in the contexts under review behave like determiners, but my analysis is nevertheless different and has, I believe, many advantages over A & EE's account.

3.1 De as a semi-determiner

My proposal is as follows: de is a semi determiner that must work in tandem with another element in the specifier of CardP in order to yield a grammatical sentence (see Heyd Reference Heyd2003 for the idea that de in negative contexts is a deficient determiner). Since double determination is required in the case of des (on my view, des is the combination of de+les) de is not capable of saturating the nominal variable on its own (12a). This is why either les, a free form, needs to be added, as in (12b) or – and this is the interesting part – an adjective needs to be added, as in (12c).

  1. (12)

    1. (a) *J'ai lu de romans.I.have read DE novels*J'ai lu [CardP de [NP romans]]. Intended: ‘I read novels.’

    2. (b) J'ai lu des romans.I.have read the+DE novels J'ai lu [CardP les+DE [NP romans]].‘I read novels.’

    3. (c) J'ai lu de bons romans.I.have read DE good novelsJ'ai lu [CardP bons+DE [NP romans]]. ‘I read good novels.’

The facts presented in this paper and the essential ingredients of my analysis tally with the view according to which determiners are the combination of both a specifier and a head. Following Lyons (Reference Lyons1999), only the specifier or the head of definite determiners is filled with phonological material in some languages, while in others double determination is possible. For example, in English, there is only one definite determiner: the. In Danish and Icelandic, however, both a free form and a suffixal form coexist in the language. The choice between the two forms is grammatically conditioned: the free form article is used whenever the noun is preceded by modifying material; the suffix otherwise. Relevant examples appear in (13) and (14).

  1. (13)

    1. (a) huset ‘the house’ Danish

    2. (b) det gamle hus ‘the old house’

    3. (c) de tre huse ‘the three houses

  1. (14)

    1. (a) báturinn‘the boat’ Icelandic

    2. (b) hinn fallegi bátur ‘the beautiful boat’

    3. (c) hinn eini bátur ‘the one boat’ (Lyons Reference Lyons1999: 77)

As shown in (15), Swedish also has two definite articles, a free form and a suffix, cognate with those of Danish, but differs in that the two can co-occur (Holmes & Hinchliffe Reference Holmes and Hinchliffe1994). The free form is used, as in Danish, in the presence of a pre-nominal modifier, but the affix also appears in this case.

  1. (15)

    1. (a) resan‘the journey’ Swedish

    2. (b) den långa resan ‘the long journey’

    3. (c) de fyra resorna ‘the four journeys’ (Lyons Reference Lyons1999: 78)

Here both a free form and an affix are phonologically spelled-out. I assume like Lyons (Reference Lyons1999) that the affix sits in D0 while the free form appears in the specifier of DP, where D stands for ‘definiteness’ rather than ‘determiner’. On his view, because the noun raises to D0, the suffix ends up attached to the noun.The tree diagram in (16a) represents the structure before the movement while the diagram in (16b) represents the structure after movement.

  1. (16)

    1. (a)

    2. (b)

An indefinite article on the other hand is the equivalent of the cardinal ‘one’. For example, in French, as in many languages, it is clear that the indefinite determiner is a cardinal element: un ‘a’ is the same word as the cardinal un ‘one’ (most, if not all, indefinite determiners are derived from adjectives historically, see Greenberg Reference Greenberg, Greenberg, Ferguson and Moravcsik1978, Haspelmath Reference Haspelmath, Bach, Jelinek, Kratzer and Partee1995). The projection headed by the cardinal ‘one’ or un is CardP.

Just like the case of the definite article discussed above, depending on the language, the indefinite article is a free form (e.g. French un/une) or an affix (e.g. Romanian -ul, as in pom-ul ‘the tree’). In some languages double determination is possible. As discussed by Lyons (Reference Lyons1999), in Persian, yek is a free-standing indefinite article that can appear together with an affix -i. The addition of the latter makes the NP non-specific, arbitrary (the noun on its own can mean either ‘the man’ or ‘a man’), thus there is a clear semantic effect, but syntactically we have a clear case of both a free form and a suffix denoting indefiniteness. When modified, the noun does not carry the suffix, but the modifier does instead, see (17a). When not modified, the noun carries the suffix, see (17b).

  1. (17)

    1. (a) yek mard-ī khub ‘a good man’ Persian

    2. (b) yek ketābhā-ī ‘some books’ (Lyons Reference Lyons1999: 90)

This is not unlike the French cases under review. When a modifier appears post-nominally in French, des must be used. This is the free form. However, when a modifier appears pre-nominally de is preferred, if not obligatory in the majority of cases. I will argue below that de is in fact a clitic. It must attach to the modifier.

My account explains why an adjective cannot act as a determiner on its own: the contribution of de is not negligible. It is precisely the combination de+adjective that acts as a determiner. The tree diagram in (18a) below gives the representation for the string de très bons romans ‘very good novels’. The adjectival phrase in (18a) occupies the same position as a determiner would occupy had the adjective not surfaced. The tree diagram in (18b) represents the case where it is the free form les that surfaces in the specifier of DP. Of course, de appears before the adjective or les in the morphology: we come back to this point below. The syntactic structure spells out two different positions: one is a specifier, the other a head (I choose the adjunction theory of modifiers rather than Cinque's (Reference Cinque1993, Reference Cinque, Cinque, Koster, Pollock, Rizzi and Zanuttini1994) approach, i.e. APs as specifiers of Agreement projections, but nothing hinges on this matter). Note that I assume that les in (18b) is a maximal projection (a DP). Angled brackets, as around AP in (18a), indicate that the item is a copy of the raised element.

  1. (18)

    1. (a) de très bons romans

    2. (b) des très bons romans

The question that now arises is why de always surfaces to the left of the AP or to the left of les. The diagrams in (18a–b) are only the syntactic structures, not the phonological output.

I want to argue that de is a phrasal proclitic: it is phonologically weak. For example, it is often reduced, as in d'excellents romans ‘excellent novels’. It also forms a prosodic word with the adjective: it needs to attach to the overall XP with which it is associated, namely CardP, as in (19a) below. As is well-known, phrasal clitics have special properties (see Zwicky Reference Zwicky1977): they are not adjacent to a word that they select but attach instead to a specific syntactic phrase (on phrasal clitics, see Klavans Reference Klavans1985, Anderson Reference Anderson1992, Legendre Reference Legendre, Gerlach and Grijzenhout2000, Spencer Reference Spencer, Gerlach and Grijzenhout2000 and Himmelmann Reference Himmelmann, Haspelmath, König, Oesterreicher and Raible2001 especially for phrasal proclitic articles). I assume that the leftward operation is achieved at PF. In (19b) les has to further merge morphologically with de yielding des since we end up with a portmanteau word.

  1. (19)

In summary, my analysis shows that in some languages, i.e. French, it is necessary for both the head and the specifier of a given nominal to be visible for that nominal to be licensed. This is because the function of the French indefinite article is distributed along the specifier and the head. This result is very different from Landau's (Reference Landau2007) proposal about visibility in syntax, since on his view it is the head of a given phrase that must be filled (the specifier may remain empty). On my view, both ε and ∂ in (20) must be filled (see Section 5 below, where I give a proper introduction of Landau's Reference Landau2007 account; in (20), ε=specifier position and ∂=head position).

  1. (20) Visibility

The notion of economy adopted herein is closer to Giusti's (Reference Giusti and Cinque2002) proposal. According to Giusti (Reference Giusti and Cinque2002: 70), a functional head is realized only if it is absolutely necessary, but when it is inserted, a functional projection has to be licensed, either by making the specifier visible or making the head visible or both.

  1. (21) Principle of economy of lexical insertion

    A functional projection must be licensed at all levels of representation by

    1. (a) making the specifier visible

    2. (b) making the head visible

      (Giusti Reference Giusti and Cinque2002: 70; see also Dimitrova-Vulchanova & Giusti Reference Dimitrova-Vulchanova, Giusti, Alexiadou and Wilder1998)

Clauses (a) and (b) may operate either disjointly or conjointly (depending on the language or the construction). In the case of indefinite phrases in French, they apply conjointly.

3.2 XP movement versus head movement

Androutsopoulou & Español Echevarría (Reference Androutsopoulou, Echevarría and Griffin2003) argue that examples such as (1b) above are grammatical because the adjective raises to D0. Although they do not spell out the details and consequences of such an analysis, the idea is that phrases with pre-nominal modifiers are the equivalent of nominals headed by D0.

The account in A & EE is similar to what Delfitto & Schroten (Reference Delfitto and Schroten1991) – D & S henceforth – propose for the Italian facts pertaining to bare nouns with pre-nominal adjectives (see Section 2 above). These authors claim that in such contexts the noun raises to D0 at LF. In both accounts (A & EE's and D & S's), the D0 position comes to be filled: by the adjective itself (overtly) in A & EE, in which case the adjective acts like a saturator for the variable introduced with the noun (in lieu of a determiner), and by the noun (covertly) in D & S. On the latter view, adjectives that can appear pre-nominally are heads while those that cannot are phrasal (they cannot incorporate into D0).

Regarding the D & S account, the first problem of note concerns the architecture of the grammar, and the need to update it. It is no longer possible to appeal to differences between pre-Surface Structure and post-Surface Structure when talking about movement, since such a level of representation is no longer postulated in minimalism (Spell-Out is not a level of representation, see Chomsky Reference Chomsky1995). This means that on D & S's account, the noun will have to move to D0 overtly. This is certainly not an impossible operation and has been proposed independently for proper names (Longobardi Reference Longobardi1994). Proper names are licensed in both object and subject positions. Although they contain an empty D0, the proper name raises to it (at least in languages like English; Greek has obligatory determiners with proper names). But the main problem with D & S's account is that it is not clear why bare nouns would move to D0only when modified. The movement is not motivated – it is purely ad hoc. The question is: What stops bare nouns to move to D0 in Italian even in the absence of a modifier?

The A & EE analysis of the French facts does not face the same problems as the D & S analysis, since on the latter view raising of the adjective to D0 is triggered only when les (treated as an expletive determiner, since des=de+les, where les has no definite or specific import), is absent.Footnote 2 In des bons livres ‘good books’, the adjective bons is a specifier just as in the case of post-nominal adjectives, e.g. des livres intéressants ‘interesting books’, les is not absent (it is in D0).

The structure in (22) is a simplified version of what A & EE proposes, but is sufficient for our purposes. Note that the adjective raises to D0 via head movement (form inside the AP) and that the element de plays no role: in their analysis, it is treated as an external element to the DP.

  1. (22)

On this view, post-nominal adjectives are different and are not capable of saving the sentence from ungrammaticality because a determiner is necessary: the determiner is the head of the whole structure rather than the adjective. In other words, post-nominal adjectives are specifiers throughout the whole syntactic derivation and they can never incorporate into D0 because of their categorical status. In contrast, pre-nominal adjectives are heads (see Lamarche Reference Lamarche1991, Valois Reference Valois1991a, Bouchard Reference Bouchard1998 for French and Sadler & Arnold Reference Sadler and Arnold1994 for English) and as heads they can incorporate into D0.

According to A & EE, this squares well with independent properties distinguishing pre-nominal from post-nominal adjectives. First, Romance pre-nominal adjectives, in contrast to post-nominal ones, cannot take complements. Compare (23a) with (23b).

  1. (23)

    1. (a) le professeur respectueux de ses édiantsthe professor respectful of his students‘the professor respectful of his students’

    2. (b) *le respectueux de ses étudiants professeur(le respectueux professeuris marginal but not impossible)

Second, while liaison is obligatory with pre-nominal adjectives, as shown in (24a), it is optional with post-nominal adjectives, as illustrated in (24b). Because of these facts, Lamarche (Reference Lamarche1991) argues that pre-nominal adjectives incorporate into the noun.

  1. (24)

    1. (a) les bonnes [z] idées de Paulthe good ideas of Paul‘Paul's good ideas’

    2. (b) (i) les livres anglais the books English(ii) les livres [z] anglaisthe books English

However, my account is superior to A & EE's. First, it explains why in the cases under review an adjective cannot act as a determiner on its own. If it could, then the sentence in (25) would be grammatical. Given A&EE's account, this sentence is expected to be grammatical, since de plays absolutely no role in the analysis.

  1. (25) *J'ai lu bons romans.I.have read good novels‘I read good novels.’

On my view, the contribution of de is crucial. I argue that it is precisely the combination de+adjective that acts as a determiner. My analysis goes back to an old idea first mentioned by Damourette & Pichon (1911–Reference Damourette and Pichon1940) according to which de functions as the second half of the article. My account is thus different from most traditional accounts. For many linguists, de is an article (Frei Reference Frei1960) while for others it is a preposition (Clédat Reference Clédat1901, Wagner & Pinchon Reference Wagner and Jacqueline1962, Gross Reference Gross1967). Yet others (Kupferman Reference Kupferman2004) have proposed that de has a dual nature, either prepositional (when interpreted as partitive) or a quantificational head. On this view, a nominal phrase such as des épinards ‘spinach’ looks as follows (Kupferman Reference Kupferman2004: 67):

  1. (26)

My analysis shares with that of Kupferman's (Reference Kupferman2004) the idea that des is the combination of de and les (one argument in favour of this view is that des alternates with de ces ‘of these’, de mes ‘of my.pl’, ‘de ceux'of those', de ce que ‘of which’, etc.). The difference between Kupferman's proposal and mine is that, although I treat de as an element separate from les, I nevertheless advocate a strong relationship between the two elements, since for me both de and les are two parts of a single determiner entity. This allows me to explain facts pertaining to de+adjective strings, facts that are not studied or explained in Kupferman's (Reference Kupferman2004) analysis.

The second point of comparison with A&EE concerns evidence that in the cases under review the adjective raises to the D domain, not via head movement, but via phrasal movement instead. Evidence for such a claim comes from three sources. First, modifiers that modify de nominals can themselves be modified. The examples in (27) show that the pre-nominal adjective bons ‘good' can surface with an intensifier such as très ‘very’ (27a) or assez ‘quite’ (27b).

  1. (27)

    1. (a) J'ai lu de très bons romans l'été dernier.I. have read DE very good novels the.summer last‘I read very good novels last summer.’

    2. (b) Il a obtenu d'assez bons résultats.he has obtained DE.quite good results‘He obtained quite good results.’

Second, as illustrated by (28), de nominals can appear in comparative constructions. In this example, plus ‘more’ modifies beaux musées ‘beautiful museums’.

  1. (28) J'ai visité de plus beaux musées cette année que l'année dernière.I.have visited DE more beautiful museums this year than the.year last‘I visited better museums this year than last year.’

Third, a modified de nominal can be coordinated while being modified by a single intensifier, as shown in (29).

  1. (29) J'ai lu de très bons et charmants romans.I.have read DE very good and charming novels‘I read very good and (very) charming novels.’

On the basis of such facts – and also of the fact that pre-nominal adjectives can be focalized (une SUPERBE voiture ‘a superb car’) – Laenzlinger (Reference Laenzlinger2005) argues that, if we follow Kayne's (Reference Kayne1975) tests for clitichood of pronouns (akin to incorporation, see Lamarche Reference Lamarche1991, Valois Reference Valois1991a, Bouchard Reference Bouchard1998), it turns out that pre-nominal adjectives do not qualify as clitics incorporated into the noun (incorporated elements/clitics cannot be focalized, they form a prosodic word with their hosts). We should also add that clitics in French are limited in number (they are closed class items). Pre-nominal adjectives, on the other hand, are members of an open class and do not become part of the phonological word with which they surface (this is a point made by Lieber (Reference Lieber1992: 52–53), who argues against Stowell's (Reference Stowell1981) original idea according to which modifiers are positioned after the head noun, but raise as clitics when they surface pre-nominally).

We appear to have contradictory evidence, however, because as mentioned above, in French (and in English) it is not possible for pre-nominal adjectives to surface with complements whereas post-nominal adjectives are free to appear with complements. If, as I propose, pre-nominal adjectives raise to the D domain via phrasal rather than head movement when de does not surface with les, we might expect pre-nominal adjectives to appear with complements. However, this is not possible, as shown by (30).

  1. (30) *De respectueux de leurs étudiants professeurs ont été félicités.DE respectful of their students professors have been congratulatedIntended: ‘Professors respectful of their students were congratulated.’

Having said that, as pointed out by Alexiadou, Haegeman & Stavrou (Reference Alexiadou, Haegeman and Stavrou2007:389), if not having a complement necessarily leads us to postulate a head analysis, then all pre-nominal English adjectives would have to be derived by the head analysis, since they cannot surface with a complement either. Yet, we have much independent evidence that many adjectives in English have a predicative source and thus originate in a post-nominal position (see Kayne Reference Kayne1994). In addition, pre-nominal adjectives in English can be conjoined and can appear with a full array of adjectival modifiers. This is shown in (31) – examples from Lieber (Reference Lieber1992: 53).

  1. (31)

    1. (a) a very big and unbelievably dirty dog

    2. (b) a not too surprisingly tall woman

It should also be noted that complements with pre-nominal adjectives are not ruled out universally: it is possible in languages such as Greek, German, Albanian, Romanian and Bulgarian (see Alexiadou et al. (Reference Alexiadou, Haegeman and Stavrou2007: 388) for Greek and German examples). It must therefore be the case that French pre-nominal adjectives cannot surface with complements for independent reasons.

Finally, there are problems with head movement in minimalist grammars. According to Chomsky (Reference Chomsky, Martin, Michaels and Uriagereka2000, Reference Chomsky and Kenstowicz2001), ‘a substantial core of head-raising processes, excluding incorporation in the sense of Baker (Reference Baker1988), may fall within the phonological component’ (Chomsky Reference Chomsky and Kenstowicz2001: 37). In recent years, it has been popular in the literature to take the operation out of the narrow syntax, relegate it to PF or eliminate it completely and replace it with XP movement/remnant movement. Head movement is problematic for minimalist syntax because (i) it violates the Extension Condition; (ii) it also violates the c-command condition on traces; (iii) unlike specifier–head checking relations, head-to-head relations do not seem to involve ‘complementary’ (‘matching’) features (say, interpretable φ-features of nouns and uninterpretable φ-features of verbs); and (iv) head movement seems to have no semantic effect.

That we do not need head movement in order to account for the facts under review must be seen as a welcome result. In sum, although the current proposal owes much to A & EE's original insight, my conclusions are very different from theirs.

3.3 Scope

If the idea according to which pre-nominal modifiers in de nominal contexts come to act as determiners or saturators of nominal's variables is correct, then it is predicted that such phrases should behave like canonical indefinite phrases with respect to scope. This prediction is borne out.

As is well-known, object de nominals that are not modified cannot receive wide scope: they are only interpreted as taking narrow scope. To illustrate, (32) below can only mean ‘It is not the case that I have read novels this summer’ and not ‘There are some novels and it is not the case that I have read them this summer.’ According to Mathieu (Reference Mathieu2004), the reason why they receive only narrow scope is because they are semantically incorporated (they denote properties). This is a property they share with bare nouns in other languages that have overt articles, e.g. Spanish and Italian: these denote properties and receive only narrow scope (Longobardi Reference Longobardi1994, Chierchia Reference Chierchia1998, Dobrovie-Sorin & Laca Reference Dobrovie-Sorin, Laca and Godard2003, among many others). The sentences in italics after the examples give the different scope possibilities/semantic readings.

  1. (32) Je n'ai pas lu de roman(s) l'été dernier.I NE.have not read DE novel(s) the.summer last‘I didn't read any novel(s) last summer.’‘It is not the case that I read a novel/novels last summer’ *‘There is/are a novel/some novels and it is not the case that I read it/them last summer.’

However, a modified de nominal can take wide scope in relation to negation. This is shown in (33). The sentence can mean either ‘It is not the case that I read good novels last summer’ (in which case the de nominal takes narrow scope) or ‘There are some good novels and it is not the case that I read them last summer’ (in which case the de nominal takes wide scope). The judgement about the wide scope is subtle, but there is a clear contrast between (32) and (33): (33) is ambiguous, (32) is not.

  1. (33) Je n'ai pas lu de bons romans l'été dernier.I NE.have not read DE good novels the.summer last ‘It is not the case that I read good novels last summer.’ ‘There are some good novels and it is not the case that I read them last summer.’

In other contexts, the similarity between de and des nominals in terms of scope is even clearer. As indefinite phrases, des nominals can take either narrow or wide scope in relation to a predicate like chercher ‘to look for’. This is shown in (34): des enveloppes ‘envelopes’ can be interpreted either specifically (I am looking for specific envelopes) or non-specifically (I am looking for (any) envelopes – I have none in mind).

  1. (34) Je cherche des enveloppes.I look.for DE+LES envelopes‘I am looking for envelopes.’ ‘There are envelopes (somewhere) and I am looking for them.’

The example in (35) is similar to the one in (34) in that the modified des nominal can yield either a wide or a narrow scope reading. I may be looking for specific big envelopes that I put away somewhere or I am simply in search of big envelopes.

  1. (35) Je cherche des grandes enveloppes.I look.for DE+LES big envelopes‘I am looking for big envelopes.’ ‘There are some big envelopes (somewhere) and I am looking for them.’

When we use a modified de nominal instead of a modified des nominal, we obtain the same readings. Although, admittedly the most natural reading for (36) is one according to which the nominal phrase takes narrow scope in relation to chercher ‘look for’, the interpretation according to which the nominal takes wide scope is not impossible in the right context.Footnote 3

  1. (36) Je cherche de grandes enveloppes.I look.for DE big envelopes‘I am looking for big envelopes.’ ‘There are big envelopes (somewhere) and I am looking for them.’

In the same vein, whether we use de or des, the example in (37a) below has a reading according to which there are close relatives of mine such that, if they die, I will be rich. Both modified de and des nominals are thus able to take scope over an if clause to which they belong, i.e. escape a scopal island. This is a property that indefinites have generally (see Reinhart Reference Reinhart1997), and Spector (Reference Spector, Quer, Schroten, Scorretti, Sleeman and Verheugd2003) has shown this for French des nominals. In turn, I show this for modified de nominals to make the point that their scopal properties are identical to des nominals.Footnote 4 A non-modified des nominal is exemplified in (37b).

  1. (37)

    1. (a) Si de proches parents à moi meurent, je serai riche.if DE close relatives to me die I be.futrich‘If close relatives of mine die, I will be rich.’‘If close relatives of mine die, I will inherit a house.’ ‘There are close relatives of mine, and if they die, I can inherit a house.’

    2. (b) Si des parents à moi meurent, je serai riche.‘If relatives of mine die, I will inherit a house.’ ‘There are relatives of mine, and if they die, I can inherit a house.’

I assume that the modified de N achieves wide scope via a Choice Function mechanism à la Reinhart (Reference Reinhart1997).

3.4 The case of post-nominal modifiers

This section addresses the question as to why post-nominal modifiers cannot act as determiners in French de nominal constructions. Because on my analysis pre-nominal modifiers in such contexts are phrasal (just like post-nominal modifiers), I lose the contrast introduced by the A&EE account that allowed the authors to explain why post-nominal modifiers play no role in the enhanced distribution of de nominals.

The solution to this problem is simple: post-nominal modifiers are in the wrong position to begin with. In order for a modifier to act as a determiner together with de, it must be able to appear pre-nominally independently.

All in all, it seems to me, that the position of adjectives in French is fairly free: many adjectives in French can appear pre- or post-nominally.Footnote 5 This does not mean that the interpretation is constant. Adjectives that are freer in their distribution appear to obtain a different interpretation depending on their position. Post-nominal adjectives, for example, have an emphatic interpretation or a strong subjective reading or are interpreted as restrictive/intersective (see Bouchard Reference Bouchard2002 for a detailed description). Two examples appear in (38). Whereas (38a) singles out a part of the writers as being mediocre (restrictive, intersective reading), (38b) qualifies all writers as mediocre.

  1. (38)

    1. (a) nos écrivains médiocresour writers mediocre‘our writers who are mediocre’

    2. (b) nos médiocres écrivainsour mediocre writers‘our mediocre writers’(Bouchard Reference Bouchard2002: 98, citing Blinkenberg Reference Blinkenberg1933: 108)

What is interesting is that when pre-nominal, the adjective surfaces with de (des is impossible) but when post-nominal, the adjective surfaces with des (de is impossible). This is shown in (39).

  1. (39)

    1. (a) Nous avons interrogé des/*d'écrivains médiocres.we have interviewed DE+LES/DE.writers mediocre‘We interviewed mediocre writers (i.e. writers who are mediocre).’

    2. (b) Nous avons interrogé de/*des médiocres écrivains.we have interviewed DE/DE+LES mediocre writers‘We interviewed mediocre writers.’

The difference in distribution between de and des nominals also surfaces with quality adjectives that tend to appear more often than not post-nominally, e.g. fier ‘proud’ (quality adjectives such as magnifique ‘magnificent', superbe ‘superb’ tend to be freer in their distribution: they can appear pre- or post-nominally). When the quality adjective is pre-nominal, de is preferred, as shown in (40a). When it is post-nominal, des is preferred, as illustrated in (40b).

  1. (40)

    1. (a) J'ai rencontré des/*de parents fiers à la réunion.I.have met DE+LES/DE parents proud at the meeting‘I met proud parents at the meeting.’ (i.e. ‘I met parents who were proud.’)

    2. (b) J'ai rencontré de/*des fiers parents à la réunion.I.have met DE/DE+LES proud parents at the meeting‘I met proud parents at the meeting.’

In addition, although it is often said that colour terms in French have to appear post-nominally, it turns out that they can appear pre-nominally as long as they surface with de. When they appear post-nominally, on the other hand, they must surface with des. This is shown in (41) and (42).

  1. (41)

    1. (a) Des/*de parures noires ornaient les murs.DE+LES/DE fineries black decorated the walls‘Black fineries decorated the walls.’ (i.e. ‘Fineries that were black decorated the walls.’)

    2. (b) De/*des noires parures ornaient les murs.DE/DE+LES black fineries decorated the walls‘Black fineries decorated the walls.’

  1. (42)

    1. (a) Des/*de prairies vertes s'étendaient au loin.DE+LES/DE prairies green self-stretched at.the far‘Green fields stretched in the distance.’ (i.e. ‘Fields that were green stretched in the distance.’)

    2. (b) De/*des vertes prairies s'étendaient au loin.DE/DE+LES green prairies self-stretched at.the far‘Green fields stretched in the distance.’

Of course, examples with predicative adjectives in pre-nominal position have a literary flavour to them.Footnote 6 However, the important point is that they are grammatical: a native speaker of French would have no trouble understanding what they meant if he/she came across them in, say, a novel and speakers have access to a subtle difference in meaning between the two variants. The post-nominal adjective is intersective while the pre-nominal adjective is not. What is noteworthy, however, is that when the colour term appears pre-nominally, de is possible, but des is ruled out.

The tree diagram in (43a) shows the structure for des prairies vertes in (42a) above, and (43b) is the representation for de vertes prairies in (42b). In this case, the adjective is in the specifier of CardP (see Abeillé & Godard Reference Abeillé and Godard1999, Demonte Reference Demonte, Franco, Landa and Martin1999, Androutsopoulou Reference Androutsopoulou2000, Laenzlinger Reference Laenzlinger2005, among many others, for the idea that pre-nominal adjectives attain their position via movement from a post-nominal position).

  1. (43)

    1. (a) des prairies vertes

    2. (b) de vertes prairies

I propose that movement of a modifier to the pre-nominal position is triggered by a feature on D0, call it F. The exact nature of this feature remains to be determined but for some cases we have reviewed it is clear that the feature has something to do with non-restrictiveness/non-intersectiveness. It is as if the post-nominal modifier is focused (des parures noires ‘fineries that are black (rather than white)’) whereas when the modifier is pre-nominal it is the noun together with the adjective that are focalized. For some authors, it is a Degree feature that triggers the raising of pre-nominal modifiers (Demonte Reference Demonte, Franco, Landa and Martin1999) or a Strong subjectivity (emphasis) feature (Laezlinger 2004).

There are only two kinds of adjectives that cannot appear pre-nominally: nationality terms and pure predicative adjectives (what I mean by ‘pure’ is that they cannot be used attributively – see footnote 6 above). If we follow Laenzlinger's (Reference Laenzlinger2005) hierarchy for modifiers, given in (44) below, we notice that the higher in the hierarchy the modifier is the easier it is for that modifier to appear in a pre-nominal position provided that de rather than des is used, and the lower the modifier is on the hierarchy the more difficult it is for it to appear pre-nominally if des is used (de may surface with pre-nominal colour terms).

  1. (44) Aquant > Aepistemic > Aquality > Asize > Aform > Acolour > Anationality > Aprede.g. nombreux possible talentueux immense long vert français nunumerous possible talented huge long green French naked

Here is a list of examples with adjectives in a pre-nominal position using des or de. Apart from the lowest two terms, which as I have mentioned, are purely predicative, it is always possible to use de.

  1. (45)

    1. (a) ??des nombreux problèmes/de nombreux problèmes

      ‘many problems’

    2. (b) ?des possibles erreurs/de possibles erreurs

      ‘possible errors’

    3. (c) ???des talentueux étudiants/de talentueux étudiants

      ‘talented students’

    4. (d) des immenses tours/d'immenses tours

      ‘huge towers’

    5. (e) des longues heures/de longues heures

      ‘long hours’

    6. (f) *des vertes vallées/de vertes vallées

      ‘green valleys’

    7. (g) *des français romans/*de français romans

      ‘French novels’

    8. (h) *des nus corps/*de nus corps

      ‘naked bodies’

When adjectives surface post-nominally, the pattern is the reverse: de can simply never be used and des is acceptable. The correlation/generalization is thus that de is used with pre-nominal adjectives but des is used with post-nominal adjectives.

  1. (46)

    1. (a) ??des problèmes nombreux/*de problèmes nombreux

      ‘many problems’

    2. (b) des erreurs possibles/*d'erreurs possibles

      ‘possible errors’

    3. (c) des étudiants talentueux/*d'étudiants talentueux

      ‘talented students’

    4. (d) des tours immenses/*de tours immenses

      ‘huge towers’

    5. (e) ?des heures longues/d'heures longues

      ‘long hours’

    6. (f) des vallées vertes/*de vallées vertes

      ‘green valleys’

    7. (g) des romans français/*de romans français

      ‘French novels’

    8. (h) des corps nus/*de corps nus

      ‘naked bodies’

While Laenzlinger (Reference Laenzlinger2005) argues that the noun raises around several modifiers along a skeleton that contains several functional projections (e.g. QuantP, SubjP, WeakP, etc.), I argue that it is the adjective that raises to the D domain (for a summary of the problems that are associated with the postulation of N0 movement in the DP, see Valois Reference Valois1991b; Bernstein Reference Bernstein1993; Cinque Reference Cinque1993, Reference Cinque, Cinque, Koster, Pollock, Rizzi and Zanuttini1994, Reference Cinque1999; Longobardi Reference Longobardi1994, among many others, see also Alexiadou et al. Reference Alexiadou, Haegeman and Stavrou2007). My account nevertheless shares with Laenzlinger's proposal the idea that what raises higher in the tree is an instance of XP movement rather rather than head movement.Footnote 7 The problem with Laenzlinger's (Reference Laenzlinger2005) analysis is that it simply glosses over the fact that for adjectives to appear pre-nominally in French, de rather than des must be used. Although, many of the examples he introduces contains de, it appears that on his view de is simply an allomorph of des. As should be clear by now, we know that this cannot be the case. While the distribution of adjectives within the noun phrase has been extensively discussed in recent years (Bouchard Reference Bouchard2002, Knittel Reference Knittel2005, Laenzlinger Reference Laenzlinger2005, and others), it is never emphasized that when adjectives are pre-nominal, de rather than des must be used: this fact is simply glossed over. It must, however, be a significant fact because there appears to be a deep connection between pre-nominal adjectives and de.

To conclude on this topic: nationality terms and pure predicative adjectives cannot raise to a pre-nominal position, probably because as strong predicative adjectives they are very embedded in the complement position, much more so, for instance than colour terms.

3.5 The case of du

As pointed out by a JL referee, contrary to the indefinite article des, the partitive article du/de la, is not reduced to de when co-ocurring with a pre-nominal adjective. Compare (47a) with (47b) (these are the referees' examples and judgments).

  1. (47)

    1. (a) Il a bu du (bon) vin.he has drunk DU good wine‘He drank good wine.’

    2. (b) *Il a bu de bon vin.he has drunk DE good wineIntended: ‘He drank good wine.’

Examples such as (47b) are supposed to show that de cannot systematically save a sentence from ungrammaticality when it surfaces with an adjective. One might argue that on my analysis it is predicted that de can systematically act as one half of an article, since I propose a purely structural constraint (i.e. both Card0 and Spec-CardP must be filled or, in other words, determiners are a combination of two elements). However, as pointed out by Jones (Reference Jones1996: 228), examples such as (48) are possible in very high/literary registers. For me, they are archaic, but grammatical.

  1. (48) Ils vendent de bon vin.they sell DE good wine‘They sell good wine.’

Why does (48) sound archaic? Although there is a tradition where du and des form a homogenous group (Galmiche Reference Galmiche1986b), I want to argue, following Anscombre (Reference Anscombre1996), that they are in fact very different. First, there is a semantic difference between the two elements. While the partitive interpretation is clear in the case of the singular, it is not natural in the case of the plural. The interpretation obtained in (49b) for des vins fins ‘fine wines’ is more like ‘several fine wines’ rather than partitive.

  1. (49)

    1. (a) Au cours du repas, Max a servi du vin fin.during of.the meal Max has served DU wine fine‘During the meal, Max served fine wine.’

    2. (b) Au cours du repas, Max a servi des vins fins.during of.the meal Max has served DE+LES wines fines‘During the meal, Max served fine wines.’ (Anscombre Reference Anscombre1996: 84)

As pointed out by Anscombre (Reference Anscombre1996), there is in fact a preferred use of du versus des depending on the accompanying noun. Compare (50a) with (50b).

  1. (50)

    1. (a) Dans les boulangeries, on vend du pain/?des pains.in the bakeries one sells DU bread/DE+LES breads‘In bakeries, bread/breads is/are sold.’

    2. (b) Dans les boulangeries, on vend ??du gateau/des gateaux.in the bakeries one sells DU cake/DE+LES cakes‘In bakeries, cake/cake is/are sold.’ (Anscombre Reference Anscombre1996: 84)

In addition, while des can be used in generic sentences (at least in subject position), it is not possible for du to be used in such contexts. Compare (51a) with (51b).

  1. (51)

    1. (a) Des chats ont besoin d'affection.DE+LES cats have need of.affection‘Cats need affection.’

    2. (b) *Du ketchup est à base de sauce tomate.DU ketchup is at base of sauce tomato‘Ketchup is made from tomato sauce.’ (Anscombre Reference Anscombre1996: 86)

Finally, while recognizing that examples such as (48) were common until recently (they are found with certain 20th-century authors), Anscombre (Reference Anscombre1996) points out correctly that today their use is marginal and that therefore this must indicate that du is different from des, since examples with de+A (those that are of interest to us in this paper) are perfectly grammatical. In other words, the distribution of du versus des (and its variant de in the de+A complex) is completely different. While this may not always have been the case, ‘on assiste donc en français contemporain à un divorce entre le role de de dans des et dans du/de la' (Anscombre Reference Anscombre1996: 85).Footnote 8

In conclusion, we can take the fact that J'ai lu de bons livres ‘I read good books’ is grammatical while *J'ai bu de bon vin ‘I drank good wine' is not for many people as direct evidence that du is not the singular form of des and that the two elements are distinct. This means that de in du requires a different analysis from the one I propose in this paper. I leave this interesting topic for further research.

4. The allomorphy hypothesis

An alternative and simple explanation for the grammaticality of (1b) above is that de might simply be an allomorph of des. This is what Kayne (Reference Kayne1984) proposes in passing for data such as these. On this view, (1b) is the equivalent of (52), the version with a full determiner.

  1. (52) J'ai lu des bons romans l’été dernier.I.have read DE+LES good novels the.summer last‘I read good novels last summer.’

Jones (Reference Jones1996) proposes a similar idea. Although he concedes that ‘it is not clear why the presence of a prenominal adjective should favour or allow the use of de without a definite article to indicate indefiniteness’ (Jones Reference Jones1996: 228), he nevertheless tentatively proposes that in such contexts les is simply phonologically dropped. On this view, de bons romans is thus not a bare noun, but a full (indefinite) noun phrase.

This solution cannot, unfortunately, be on the right track. Although my analysis shares with the allomorphy view the idea that modified de nominals are not bare nouns but full indefinite phrases, I show in the present section that we cannot claim that les is simply deleted in modified environments. The contribution of the pre-nominal modifier is essential for the proper licensing of de nominals in the contexts under review.

First, as (53a) illustrates, des nominals, whether they are modified or not, are not really tolerated under the scope of negation. They behave like Positive Polarity Items (PPIs). If they are judged grammatical, they are interpreted with wide scope only. However, (53b) is fully grammatical and in this case the nominal can achieve either wide or narrow scope with respect to negation.

  1. (53)

    1. (a) ??Des (bons) romans n'ont pas été publiés l'été dernier.DE+LES good novels NE.have not been published the.summer lastIntended: ‘(Good) novels were not published last summer.’‘There are (good) novels and it is not the case that I read them last summer.’ *'It is not the case that I read (good) novels last summer.'

    2. (b) De bons romans n'ont pas été publiés l'été dernier.DE good novels NE.have not been published the.summer last‘Good novels were not published last summer.’‘There are good novels and it is not the case that I read them last summer.’ ‘It is not the case that I read good novels last summer.’

In English, PPIs are also strange under the scope of negation, as shown by (54). They necessarily involve wide scope over negation.

  1. (54)

    1. (a) ??I didn't see someone.‘There is someone and it is not the case that I saw them.’ *‘It is not the case that I saw someone.' (i.e. ‘I didn't see anyone.’)

    2. (b) Someone didn't come.‘There is someone and they didn't come.’ * ‘It is not the case that someone came. (i.e. ‘No one came.’)

Second, when the modifier of a de nominal is a quantifier, e.g. nombreux ‘many’, as in (55), it is only possible to obtain the de variant. The alternative with des is ruled out. If de was simply des in disguise, we would expect both versions to be acceptable in such contexts.

  1. (55) J'ai vu de/??des nombreux films l'été dernier.I.have seen DE/DE+LES many movies the.summer last‘I saw many movies last summer.’

Third, as shown in Section 3 above, when adjectives surface post-nominally, as in the case of colour or nationality terms, des must be used. However, when modifiers appear pre-nominally, de is preferred. Again, this indicates that de is not des in disguise, for if it was, both versions should be equally licit in such environments.

Finally, common phrases or near-compounds tend to reject de (Wilmet Reference Wilmet1986: 123). This is shown in (56). Des is preferred in these cases. With des the meaning seems to be ‘youngsters’ (compound reading) whereas with de the meaning appear to be ‘young people’ (quasi-compound reading).

  1. (56)

    1. (a) Des/?de jeunes gens sont entrés.DE+LES/DE young people are entered‘Youngsters/young people came in.’

    2. (b) Des/??de vieilles filles sont entrées.De+LES/DE old girls are entered‘Spinsters came in.’

In summary of this brief section we can say that the distribution of de and des nominals in adjectival contexts is not identical. Therefore, one conclusion we can safely put forward is that de bons romans is simply not the phonologically reduced equivalent of des bons romans.

5. Modified de nominals and focus

As already mentioned in Section 2, the distribution of bare nouns in Spanish and Italian changes when a modifier to the bare noun is added. A solution in terms of focus has been proposed for this generalization and in this section I examine the possibility that the French facts under review in this paper might find an explanation through this focus solution.

I begin with Landau's (Reference Landau2007) minimalist account of ECP effects. The reason why I introduce his theory is because it is an interesting minimalist take on why bare nouns and similar elements are not licensed in subject positions and because he gives an explanation in particular as to why de nominals in French are restricted to object positions in negative environments. Although he does not deal with the problem under review in this paper, namely that modified de nominals behave differently from non-modified de nominals, I speculate that we could in principle appeal to his proposal about modified bare nouns in Spanish and Italian in order to solve the French problem. I propose to review such a solution. In a nutshell, the idea is that modified bare nouns can appear in subject positions because they are focused. This is in fact exactly what Chierchia (Reference Chierchia1998) proposes for similar facts. I will show that this idea cannot possibly be applied to the French cases under review.

Landau (Reference Landau2007) uses Holmberg's (Reference Holmberg2000) version of the EPP, where the EPP has PF requirements: it is composed not only of a categorial [D] feature, but also of a phonological [P] feature. Landau (Reference Landau2007) argues that TP in languages like English needs more than simple overt material in its specifier, it also requires a phrase whose head is phonologically realized. Whereas this requirement is imposed on DP subjects (they cannot appear with null heads), object bare nouns are exempt from it, since V bears no [P] feature.

To illustrate, compare the structure in (57a) – where ∂ is empty determiner – with the structure (57b), where the noun is fully clothed, i.e accompanied by a determiner. The structure in (57a) leads to ungrammaticality (the visibility requirement is not respected, since the head of the subject DP is phonologically empty). On the other hand, (57b) is perfectly well-formed (the visibility condition is satisfied, since the head of the subject DP is filled by an overt determiner). The crossed-out [P] means that the feature P has been verified/checked.

  1. (57)

    1. (a) No visibility

    2. (b) Visibility

There are two reasons why we should abandon the ECP as a tool to explain the distribution of bare nouns and other constructions involving an empty category. One is theoretical, the other empirical. First, the notion of government has disappeared from mainstream generative linguistics (Chomsky Reference Chomsky1995 and subsequent work), therefore appealing to the ECP, which relies heavily on the concept of government, is ruled out. Second, the traditional ECP cannot account for two specific constructions with empty categories: one is the case of bare nouns in small clause subjects and the second is the case of conjoined bare nouns in Spanish and Italian (I choose not to introduce these examples for lack of space).

Landau (Reference Landau2007) extends his analysis of bare nouns in Romance to the case of French de nominal in negative environments. The facts are as follows. A bare de nominal is not licit in subject position (58a), but it is fine in object position (58b).Footnote 9

  1. (58)

    1. (a) *De romans n'ont pas été publiés l'été dernier.DE novels NE.have not been published the.summer lastIntended: ‘No novels were published last summer.’

    2. (b) Je n'ai pas lu de romans l'été dernier.I NE.have not read DE novels the.summer last‘I didn't read novels last summer.’

For (58a) to be grammatical, an indefinite plural determiner des needs to be used. This is shown in (59).

  1. (59) Des romans n'ont pas été publiés l'été dernier.DE+LES novels NE.have not been published the.summer last‘(Some) novels were not published last summer.’

This contrast is traditionally accounted for by appealing to the notion of government: only the object is governed by the verb. Landau (Reference Landau2007) brings the French facts under his EPP account on the (reasonable) assumption that de nominals contain an empty determiner.

However, if we modify de romans in (58a), then the sentence becomes grammatical. This is shown in (60). In the case where a modified de nominal is in an object position, the positive effect of modification is masked because of course the de nominal would be licit in such an environment even without a modifier (the reason being that in this case the de nominal is licensed by negation).

  1. (60) De bons romans n'ont pas été publiés l'été dernier.DE good novels NE.have not been published the.summer last‘No good novels were published last summer.’

Although Landau's (Reference Landau2007) account has extensive empirical coverage, it does not address the issue pertaining to French de nominals that are modified. It does, however, deal with focalized bare nouns in subject positions and we may find here a tentative solution for our problem. Spanish or Italian bare nouns, as already mentioned, are banned from subject positions (see Contreras Reference Contreras, Bordelois, Contreras and Zagona1986, among many others). When the subject bare noun is focused, however, the sentence becomes grammatical as (61) shows. Here, escritores ‘writers’ is contrastively focused. Following Rizzi (Reference Rizzi and Haegeman1997) and others, we can assume that escritores raises to the specifier of a focus projection after it has moved to Spec-TP, the canonical subject position.

  1. (61) ESCRITORES han aceptado su invitación (pero periodistas no…)writers have accepted his invitation but journalists no‘WRITERS have accepted his invitation (but not journalists).’

The question that arises is why should a null determiner be excluded in Spec-TP but allowed in Spec-FocP, given the (plausible) assumption that both T0 and Foc0 are [P]-bearing heads? Landau tentatively proposes that the answer might lie with the different anchoring features for [P]. On T0, [P] is anchored by Case/φ-features on Foc0, the anchor is simply [+foc]. Suppose that within DP, Case/φ-features are specified on D0 (possibly spread to N0 by concord) but [foc] is a feature of N0. This is motivated by the observation that D0 is the prototypical locus of morphological contrasts in Case/φ-values, while it is N0 that bears the semantic content targeted by the focus interpretation. Now we understand why (61) is grammatical. Since [P] on T0 is anchored by a feature that is checked against D0, it is D0 that must be phonologically visible to satisfy [P]; hence, Bare Nouns (with a null determiner are excluded in Spec-TP. By contrast, since [P] on Foc0 is anchored by a feature checked against N0, it is N0 that must be phonologically visible to satisfy [P]; hence, a null determiner is tolerated in Spec-FocP. Note that Spec-TP can but need not be lexicalized in Romance (e.g., in post-verbal subject constructions).

It seems to me that this account is not unlike the one proposed by Chierchia (Reference Chierchia1998). Chierchia's hypothesis has the advantage of addressing the issue as to why modified bare nouns in Romance languages are also acceptable in subject positions. Chierchia (Reference Chierchia1998) argues that we have in this case weak focus. Weak focus presupposes a set of alternatives. To interpret a modified noun phrase such as famosos escritores ‘famous writers’, we intersect the set that is the extension of the adjective (people that are famous) with the set that is the extension of the noun (people that are writers), giving the set of things that are both famous and writers, that is famous writers. This is similar to, yet different from, the case of contrastive focus above, since in the latter case not only is a set of alternatives presupposed, but it is also evoked and excluded. Following Longobardi (Reference Longobardi1994), Chierchia (Reference Chierchia1998) argues that bare nouns contain a bare determiner δ and that ‘in Italian the licensing conditions for δ appear to be either licensing by a lexical head (perhaps by a process of LF incorporation) or by the functional head of a Focus Phrase, via Spec-Agreement (under the assumption that focalized constituents are moved to the Spec of FP; see Rizzi Reference Rizzi and Haegeman1997 and references quoted there)’(Chierchia Reference Chierchia1998: 386–387).

The problem is that the focus solution will not work for modified de nominals in French. As already mentioned in the introduction, it turns out that, in the case of French de nominals, only modifiers that are in a pre-nominal position are able to save a given sentence from ungrammaticality. De nominals with post-nominal modifiers are not licit in subject positions. This is shown in (62). The first of these examples, (62a), involves a bare modifier, (62b) involves a PP modifier, and (62c) involves a CP modifier. Pre-nominal and post-nominal modifiers are equally focused. Therefore, they should behave the same, but they do not. The same facts hold in the case where no negative element is present in the derivation.

  1. (62)

    1. (a) *De romans français (n')ont (pas) été publiés l'été dernier.DE novels French (NE)have not been published the.summer last‘No French novels were (not) published last summer.’

    2. (b) *De romans de grande qualité (n')ont (pas) été publiés l'été dernier.DE novels of great quality (NE)have not been published the.summer last‘(No) novels of great quality were(not) published last summer.’

    3. (c) *De romans qui se vendaient mal (n')ont (pas) été réédités l'été dernier.DE novels that self sell badly (NE)have not been reedited the.summer last‘No novels that were selling badly were (not) reedited last summer.’

On the other hand, in the case of Spanish (and Italian) bare nouns, there is no difference between pre- and post-nominal modifiers: as exemplified by (63), both are capable of saving a sentence from ungrammaticality. Let us suppose that in this case, a focus functional head is capable of licensing the nominal (via Agree or Spec–Head agreement at LF or other such mechanism). However, this will work only for Spanish (and Italian), not for French.

  1. (63)

    1. (a) Escritores franceses han aceptado su invitación.writers French have accepted his invitation‘French writers have accepted his invitation.’

    2. (b) Famosos escritores han aceptado su invitación.famous writers have accepted his invitation‘Famous writers have accepted his invitation.’

The idea that focus is not responsible for the licensing of modified French de nominals is reinforced by the fact that, contrary to what happens with bare nouns in Spanish and Italian, the addition of contrastive focus on a subject de nominal does not improve grammaticality. This is shown in (64a). This is also true in the case of an object de nominal that surfaces without a proper licensor, as shown by (64b).

  1. (64)

    1. (a) *De ROMANS n'ont pas été publiés l'été dernier.DE novels NE.have not been published the.summer last‘No NOVELS were published last summer.’

    2. (b) *J'ai lu de ROMANS l'été dernier (pas des magazines).I.have read DE books the.summer last not DE+LES magazines‘I read BOOKS last summer (not magazines).

To conclude Section 5: In French only pre-nominal modifiers allow de nominals to appear in otherwise illicit environments. There thus appears to be a syntactic constraint on the distribution of modified de nominals that is independent from the notion of focus.

6. Dayal (Reference Dayal2004)

The positive effect that modification has on noun phrases has been considered elsewhere, by Veneeta Dayal. In Dayal (Reference Dayal2004) (see also Dayal Reference Dayal, Simons and Galloway1995 and Reference Dayal1998), she discusses the case of English any, a lexical item that normally occurs in negative and other downward entailing contexts or in modal and generic environments, but which becomes acceptable in non-modal, non-negative contexts if modified by a relative clause, a prepositional phrase or an adverbial phrase (as first noted by Legrand Reference Legrand1975).

Let me illustrate. In normal cases of formal licensing we expect the licensor to take scope over the licensee. In the case of licensing by negation, for example, any in subject position cannot be licensed by sentential negation because the subject position in English is thought to be higher than negation. This is why (65a) is unacceptable but (65b) is grammatical. Of course, if there is no negation, the sentence is simply ungrammatical, as shown in (65c).

  1. (65)

    1. (a) *Any student didn't sign the petition.

    2. (b) The students didn't sign any petition.

    3. (c) *The students signed any petition.

However, if a modifier is added, then the sentences are well-formed, as shown in (66) and (67). This is true for cases with and without negation. In each case the noun is modified by a post-nominal modifier.

  1. (66)

    1. (a) Any student who went to the meeting signed the petition.

    2. (b) Any student who went to the meeting didn't sign the petition.

    3. (c) The students signed any petition that was circulated by the professors.

  1. (67)

    1. (a) Any student at the meeting signed the petition.

    2. (b) Any student at the meeting didn't sign the petition.

    3. (c) The administrators scolded any students at the meeting.

Dayal (Reference Dayal2004) argues that these cases are grammatical because the modifier is modally independent of the head noun: it introduces an independent spatio-temporal variable. This is possible, however, because the post-nominal modifier is phrasal (see Sadler & Arnold Reference Sadler and Arnold1994). The idea is that only phrasal modifiers provide the necessary situation variable that is crucial to the licensing of any: a pre-nominal modifier does not introduce an independent spatio-temporal variable. This explains why (68) is ungrammatical. In this case, the pre-nominal modifier forms a head structure with its associated noun rather than a phrase (more on this below).

  1. (68) *Mary talked to any angry student.

Dayal's proposal goes like this: Any is a universal quantifier over situations and individuals. The sentence in (68) is unacceptable because in this case there is a clash between the presupposition that the domain of quantification ranges over possible individuals and a predication that can only apply to actual individuals. This is shown in (69).

  1. (69) ∀s,x [student (x,s) & angry (x,s) & C(s)] ∀s′[s<s′ & talk(m,x,s′)]

The version of (68) with a post-nominal modifier (in (70) below) avoids this problem by restricting the domain of quantification to those possible individuals who also fulfill the property in the relative clause, including the property of being in the temporal slice denoted by the temporal specification of the relative clause. This is shown in (71).

  1. (70) Mary talked to any student who was at the meeting.

  2. (71) ∀s,x [student(x,s) & C(s) & ∀s″[s<s″& P(s″) & angry(x,s″)]] ∀s′[s<s′& talk(m,x,s′)]

Dayal (Reference Dayal2004) concludes that the apparent licensing effect of modification is derivative on independent parts of the meaning and is derived from the semantics of the parts, obviating the need for a radical shift in the role of directionality in licensing. Dayal's (Reference Dayal2004) paper is interesting because it tries to link the intriguing facts about any with the facts about bare nouns in Romance reviewed above. She gives a pair of examples from Italian, reproduced in (72) below, to show that post-nominal modifiers can suddenly save a sentence from ungrammaticality. The sentence in (72a) is not possible in Italian (a determiner needs to be inserted in order to obtain the generic interpretation) but, surprisingly, (72b) is possible. In this case, the bare noun is modified.

  1. (72)

    1. (a) *Leo odia gatti.Leo hates cats‘Leo hates cats.’

    2. (b) Leo odia gatti di grandi dimensioni.Leo hates cats of great dimension‘Leo hates cats of large size.’

Dayal (Reference Dayal2004) also notes that whereas English plural definite phrases generally do not yield a generic interpretation, they suddenly can if the plural definite phrase is modified, as in (73a) below. The fact that English plural definites do not allow a generic interpretation can be attributed to the familiarity requirement associated with the English definite. In (73b) the underlined condition is, according to Dayal (Reference Dayal2004), meant to capture the familiarity requirement, along the lines of Kamp & Reyle (Reference Kamp and Ryle1993) . The variable introduced by the definite must be identified with a variable in the superordinate Discourse Representational Structure (DRS) resulting in reference, in this case, to some entity in the world of evaluation.

  1. (73)

    1. (a) The students work hard.

    2. (b) Gen s,x [students-in-s(x) & x=?] [are-successful-in-s(x)]

When the English plural definite phrase is modified, the presupposition that the students be anchored to some salient entity in the context is clearly inert. The underlined condition in (74b) is not part of the representation and the generic reading results.

  1. (74)

    1. (a) The students who work hard are successful.

    2. (b) Gen s,x [students-in-s(x) & work-hard(x) & x=?] [are-successful-in-s(x)]

Despite the merits of Dayal's (Reference Dayal2004) analysis of modified any in English, it is clear that her account cannot be generalized to French modified de nominals. The reason is because in French, post-nominal modifiers (which are supposed to be phrasal on Dayal's view) do not save a sentence from ungrammaticality. On the contrary, it is a modifier that is pre-nominal that is capable of saving a sentence from ungrammaticality. There is nothing to suggest that in French pre-nominal modifiers are phrasal but post-nominal modifiers are heads. In fact, quite the opposite has been claimed (see Valois Reference Valois1991a, Lamarche Reference Lamarche1991 and Bouchard Reference Bouchard1998).

It should also be mentioned that in Spanish and Italian, pre-nominal modifiers are quite capable of saving a sentence from ungrammaticality. As already mentioned earlier in the paper (see Section 2 above), it makes no difference whether modifiers appear pre- or post-nominally with Spanish and Italian bare nouns: in both cases the sentence is grammatical. On Dayal's (Reference Dayal2004) account this is unexpected.

  1. (75)

    1. (a) Bravi studenti hanno telefonato. Italiangood students have phoned‘Good students have phoned. ’ ( Chierchia Reference Chierchia1998: 384)

    2. (b) Famosos escritores han aceptado su invitación. Spanishfamous writers have accepted his invitation‘Famous writers have accepted his invitation.’

Dayal's (Reference Dayal2004) proposal (as she acknowledges herself) cannot account either for the fact that other factors, such as coordination and focus (see Section 5 above), improve the status of bare plurals in Romance. The licensing of any is clearly different from the licensing of Romance bare nouns. For example, coordination does not improve the status of any without a licensor, as shown in (76a), and nor does focus, as illustrated in (76b).

  1. (76)

    1. (a) *Any student and any professor signed the petition

    2. (b) *ANY STUDENT signed the petition.

Finally, it seems to me that the set of data in (72) should receive an analysis slightly different from other cases where modification makes a difference for the distribution of noun phrases. For example, in French, it is marginally possible to use des noun phrases in generic contexts if the noun phrase is modified (see Vogeeler & Tasmowski Reference Vogeeler, Tasmowski, Vogeleer and Tasmowski2005, Reference Vogeeler, Liliane, Vogeleer and Tasmowski2006 for a brief discussion); the definite determiner les is otherwise commonly used. However, the modification can come not only in the form of a pre-nominal adjective, as in (77b), but also in the form of a post-nominal adjective, as in (77c), or even a PP modifier, as in (77d). These cases are clearly different from the cases of de nominal.

  1. (77)

    1. (a) Je déteste les/*des chats.I hate the/DE+LES cats‘I hate cats.’

    2. (b) Je déteste les/?des petits chats.I hate the/DE+LES small cats‘I hate small cats.’

    3. (c) Je déteste les/?des chats noirs.I hate the/DE+LES cats black‘I hate black cats.’

    4. (d) Je déteste les/des chats de petite taille.I hate the/DE+LES cats of small size‘I hate small size cats.’

In conclusion, the fact that only pre-nominal modifiers can help French de nominals to appear in subject positions and more generally without a licensor suggests that a formal, i.e. syntactic, operation is responsible for the grammaticality of such cases. It appears that the influence that modification has on English any is of a different nature from the influence that modification has on French de nominals. There may be overlapping properties between the two sets of constructions, but French de nominals are clearly sensitive to the syntactic distinction between pre- and post-nominal modification. Therefore, Dayal's account, although very interesting and promising, cannot be applied to the case of French de nominals.

7. Conclusion

In this paper, I have shown that it is possible for modifiers to extend the distribution of de nominals in French. I have shown that this is not an isolated phenomenon. It has been noted in the linguistic literature that modification can be used to license elements (e.g. Romance bare nouns, English any) in a sentence that would otherwise be illicit. The accounts offered for such phenomena have been shown here not to be applicable to the French facts (three different types of accounts have been reviewed: the allomorphy analysis, the focus account and the extra situation variable account). In view of the special distribution that modified French de nominals exhibit (only pre-nominal adjectives can save a sentence from ungrammaticality if a determiner like les is not present), I proposed a new account. I argued that modifiers can come to play the role of determiners in French as long as they are accompanied by a head de, which is the spell-out of a Cardinal head (see Lyons Reference Lyons1999). In French, the role of determiners is thus shared between the specifier and the head of that category. In the absence of the determiner les, an adjective can raise to the specifier of CardinalP. This is achieved via phrasal rather than head movement.

Footnotes

[1]

I wish to thank two anonymous JL referees for very helpful comments. Funding from SSHRC (Social Sciences and Humanities Research Council of Canada) is gratefully acknowledged (#410-2005-1781 awarded to Éric Mathieu).

[2] This determiner is usually claimed not to be the definite determiner les as in les romans ‘the books’ (since the interpretation is not definite or in many cases partitive). According to Milner (Reference Milner1978) and Jones (Reference Jones1996), les is in this case an expletive with no presuppositional and/or definiteness force for the DP, since des romans does not refer to a known set of novels. Instead, it is an indefinite phrase. Although, I will also decompose des into de les, this does not mean necessarily that I think les is in this case (indefinite use of des) an expletive determiner. I remain agnostic. It could well be that les is a full determiner. For example, Dayal (Reference Dayal2004) argues against les being considered an expletive on the grounds that it is restricted from appearing in the nuclear scope where expressions without existential presuppositions are interpreted – this distinguishes English bare plurals from Italian definite plurals.

[3] As a JL referee points out, the wide versus narrow scope reading can be shown through the form of a following pronoun: je cherche de grandes enveloppes et je ne les trouve pas ‘I am looking for big envelopes and I can't find them [wide scope] versus je cherche de grandes enveloppes et je n'en trouve pas ‘I am looking for big envelopes and I can't find any’[narrow scope].

[4] In the absence of scopal elements in the sentence, it is clear that in the right contexts, des nominals are ambiguous between an existential non-partitive reading and a partitive interpretation (see Galmiche Reference Galmiche, David and Kleiber1986a).

[5] To the exception of non-predicative adjectives: as is well-known, un ancien president ‘a former president’ is possible, but not *un président ancien ‘a former president’. Also, nationality terms resist appearing in pre-nominal positions. We return to this point below, in Section 3.

[6] What I mean by predicative adjectives here is: those adjectives that are normally in predicative position but can also appear pre-nominally; I will call those adjectives that can appear only predicatively, pure predicative adjectives – see next page of in the main text.

[7] It is worth noting that modifier raising as XP movement within the DP domain is not an isolated phenomenon. It applies not only in the case of de nominals, but also in the case of demonstratives in languages such as Greek. In this and other languages, it is possible for both definite articles and demonstratives to surface in a DP. For many researchers, the demonstrative raises from a lower position (Giusti Reference Giusti and Haegeman1997, Panagiotidis Reference Panagiotidis2000, Brugè Reference Brugè and Cinque2002, Shlonsky Reference Shlonsky2004) to surface in Spec-DP. This is shown in (ia); (ib) represents the base position for the demonstrative.

  1. (i)

    1. (a) [Spec-DP afto [D to vivlio <afto>]] this the book

    2. (b) [D0 to vivlio afto] the book this‘this book’

[8] ‘We therefore witness today, in contemporary French a divorce between the role of de in des and in du/de la’ – my translation.

[9] De nominals are neither bare nouns (they cannot surface on their own) nor Negative Polarity Items (see Kayne Reference Kayne1984, Heyd Reference Heyd2003, Heyd & Mathieu Reference Heyd and Mathieu2005): unlike English any phrases they cannot appear in questions or conditionals, as shown by the examples in (i), unless they are modified, as illustrated by (ii).

  1. (i)

    1. (a) Avez-vous vu *d'/des étudiants?have-you seen DE/DE+LES students ‘Did you see any student(s)?’

    2. (b) Si *d'/des étudiants arrivent, dis-le moi.if DE/DE+LES students arrive tell.it me.dat‘If any students arrive, let me know.’

    1. (i)

      1. (a) Avez-vous vu de bons étudiants?have.you seen DE good students‘Did you see any good student(s)?’

      2. (b) Si de bons étudiants arrivent, dis-le moi.if DE good students arrive tell.it me.dat‘If any good students arrive, let me know.’

References

REFERENCES

Abeillé, Anne & Godard, Danièle. 1999. La position de l'adjectif épithète en français: le poids des mots. Recherches Linguistiques de Vincennes 28, 9–32.CrossRefGoogle Scholar
Alexiadou, Artemis, Haegeman, Liliane & Stavrou, Melita. 2007. Noun phrase in the generative perspective. New York: Mouton de Gruyter.CrossRefGoogle Scholar
Anderson, Stephen R. 1992. A-Morphous morphology. New York: Cambridge University Press.CrossRefGoogle Scholar
Androutsopoulou, Antonia. 2000. On the (in)ability of pre-nominal attributive adjectives to take complements. WCCFL (West Coast Conference on Formal Linguistics) 19, 29–42. Stanford, CA: CSLI Publications.Google Scholar
Androutsopoulou, Antonia & Echevarría, Manuel Español. 2003. D-raising and Asymmetric Agreement in French. In Griffin, William E. (ed.), The role of agreement in natural language: TLS 5 Proceedings (Texas Linguistics Forum 53), 35–40.Google Scholar
Anscombre, Jean-Claude. 1996. Partitif et localisation temporelle. Langue française 109, 80–103.CrossRefGoogle Scholar
Baker, Mark C. 1988. Incorporation: A theory of grammatical function changing. Chicago, IL: The University of Chicago Press.Google Scholar
Bernstein, Judy B. 1993. Topics in the syntax of nominal structure across Romance. Ph.D. dissertation, City University of New York.Google Scholar
Blinkenberg, Andreas. 1933. L'ordre des mots en français moderne: deuxième partie. Copenhagen: Levin & Munksgaard.Google Scholar
Borer, Hagit. 2005. In name only. Oxford: Oxford University Press.Google Scholar
Bouchard, Denis. 1998. The distribution and interpretation of adjectives in French: A consequence of Bare Phrase Structure. Probus 10, 139–183.CrossRefGoogle Scholar
Bouchard, Denis. 2002. Adjectives, number and interfaces. Oxford: Elsevier.CrossRefGoogle Scholar
Brugè, Laura. 2002. The positions of demonstratives in the extended nominal projection. In Cinque, (ed.), 15–53.Google Scholar
Chierchia, Gennaro. 1998. Reference to kinds across languages. Natural Language Semantics 6, 339–405.CrossRefGoogle Scholar
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press.Google Scholar
Chomsky, Noam. 2000. Minimalist inquiries: The framework. In Martin, Roger, Michaels, David & Uriagereka, Juan (eds.), Step by step: Essays on minimalist syntax in honor of Howard Lasnik, 89–155. Cambridge, MA: MIT Press.Google Scholar
Chomsky, Noam. 2001. Derivation by phase. In Kenstowicz, Michael (ed.), Ken Hale: A life in language, 1–52. Cambridge, MA: MIT Press.Google Scholar
Cinque, Guglielmo. 1993. On the evidence for partial N-movement in the Romance DP. University of Venice Working Papers in Linguistics 3, 21–40.Google Scholar
Cinque, Guglielmo. 1994. On the evidence for partial N-movement in the Romance DP. In Cinque, Guglielmo, Koster, Jan, Pollock, Jean-Yves, Rizzi, Luigi & Zanuttini, Raffaella (eds.), Paths towards Universal Grammar: Studies in honor of Richard S.Kayne (Georgetown Studies in Romance Linguistics), 85–110. Washington, DC: Georgetown University Press.Google Scholar
Cinque, Guglielmo. 1999. Adverbs and functional heads: A crosslinguistic perspective. Oxford: Oxford University Press.CrossRefGoogle Scholar
Cinque, Guglielmo (ed.). 2002. Functional structure in DP and IP: The cartography of syntactic structures, vol. 1, 54–90. Oxford: Oxford UniversityPress.CrossRefGoogle Scholar
Clédat, Léon. 1901. La préposition et l'article partitif. Revue de philologie française et de literature 15, 81–131.Google Scholar
Contreras, Heles. 1986. Spanish Bare NPs and the ECP. In Bordelois, Ivonne, Contreras, Heles & Zagona, Karen (eds.), Generative studies in Spanish syntax, 25–49. Dordrecht: Foris.Google Scholar
Curat, Hervé. 1999. Les déterminants dans la référence nominale et les conditions de leur absence. Genève: Librairie Droz.Google Scholar
Damourette, Jacques & Pichon, Édouard. 1911–1940. Des mots à la pensée: essai de Grammaire de la langue française, 7 vols. Paris: D'Artrey.Google Scholar
Dayal, Veneeta. 1995. Licensing any in non-negative/non-modal contexts. In Simons, Mandy & Galloway, Teresa (eds.), SALT (Semantics and Linguistic Theory) V, 54–71.Google Scholar
Dayal, Veneeta. 1998. Any as inherently modal. Linguistics & Philosophy 21, 433–476.CrossRefGoogle Scholar
Dayal, Veneeta. 2004. Licensing by modification. Ilha do Desterro Florianópolis 47, 217–238.Google Scholar
Déchaine, Rose-Marie. 1993. Predicates across categories: Towards a category-neutral syntax. Ph.D. dissertation, University of Massachusetts, Amherst.Google Scholar
Delfitto, Denis & Schroten, Jan. 1991. Bare plurals and the number affix in DP. Probus 3, 155–185.CrossRefGoogle Scholar
Demonte, Violeta. 1999. A minimal account of Spanish adjective position and interpretation. In Franco, Jon A., Landa, Alazne & Martin, Juan (eds.), Analyses in Basque and Romance linguistics, 45–76. Amsterdam: John Benjamins.Google Scholar
Dimitrova-Vulchanova, Mila & Giusti, Giuliana. 1998. Fragments of Balkan nominal structure. In Alexiadou, Artemis & Wilder, Chris (eds.), Possessors, predicates and movement in the determiner phrase, 333–361. Amsterdam: John Benjamins.CrossRefGoogle Scholar
Dobrovie-Sorin, Carmen & Laca, Brenda. 2003. Les noms sans déterminant dans les langues romanes. In Godard, Danièle (ed.), Les langues romanes – Problèmes de la phrase simple, 235–281. Paris: Éditions du CNRS.Google Scholar
Frei, Henri. 1960. Tranches homophones: à propos de l'article partitif. Word 3, 317–322.CrossRefGoogle Scholar
Galmiche, Michel. 1986a. Référence indefinite, événements, proprétés et pertinence. In David, Jean & Kleiber, Georges (eds.), Déterminants: syntaxe et sémantique, 41–71. Paris: Klincksieck.Google Scholar
Galmiche, Michel. 1986b. Notes sur les noms de masse et le partitif. Langue française 72, 40–53.CrossRefGoogle Scholar
Giusti, Giuliana. 1997. The categorial status of determiners. In Haegeman, Liliane (ed.), The new comparative syntax, 95–124. London: Longman.Google Scholar
Giusti, Giuliana. 2002. The functional structure of noun phrases: A bare phrase structure approach. In Cinque, (ed.), 54–90.CrossRefGoogle Scholar
Greenberg, Joseph H. 1978. How does a language acquire gender markers. In Greenberg, Joseph H., Ferguson, Charles A. & Moravcsik, Edith A. (eds.), Universals of human language, vol. 3, 47–82. Stanford, CA: CSLI Publications.Google Scholar
Gross, Maurice. 1967. Sur une règle de cacophonie. Langages 7, 105–119.CrossRefGoogle Scholar
Haspelmath, Martin. 1995. Diachronic sources of ‘all’ and ‘every’. In Bach, Emmon, Jelinek, Eloise, Kratzer, Angelika & Partee, Barbara H. (eds.), Quantification in natural languages, 363–382. Dordrecht: Kluwer.Google Scholar
Heyd, Sophie. 2003. L'interprétation des syntagmes nominaux en ‘des’ et ‘de’ en position sujet et objet – généricité, habitualité et incorporation sémantique. Ph.D. dissertation, Université de Strasbourg 2.Google Scholar
Heyd, Sophie & Mathieu, Éric. 2005. The semantics of [de N] structures, incorporation and referential intentions. Presented at Indefinites and Weak Quantifiers, Marie Haps Institute and the Royal Flemish Academy for Sciences and Arts, Brussels, 6–8 January.Google Scholar
Himmelmann, Nikolaus. 2001. Articles. In Haspelmath, Martin, König, Ekkehard, Oesterreicher, Wulf & Raible, Wolfgang (eds.), Language typology and language universals: An international handbook, vol. 1, 831–841. Berlin: de Gruyter.Google Scholar
Holmberg, Anders. 2000. Scandinavian stylistic fronting: How any category can become expletive. Linguistic Inquiry 31, 445–483.CrossRefGoogle Scholar
Holmes, Philip & Hinchliffe, Ian. 1994. Swedish: A comprehensive grammar. London: Routledge.Google Scholar
Jones, Michael Allan. 1996. Foundations of French syntax. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Kamp, Hans & Ryle, Uwe. 1993. From discourse to logic. Dordrecht: Kluwer.Google Scholar
Kayne, Richard S. 1975. French syntax: The transformational cycle. Cambridge, MA: MIT Press.Google Scholar
Kayne, Richard S. 1984. Connectedness and binary branching. Dordrecht: Foris.CrossRefGoogle Scholar
Kayne, Richard S. 1994. The antisymmetry of syntax. Cambridge, MA: MIT Press.Google Scholar
Klavans, Judith. 1985. The independence of syntax and phonology in cliticization. Language 61, 95–120.CrossRefGoogle Scholar
Knittel, Marie Laurence. 2005. Some remarks on adjective placement in the French NP. Probus 17, 185–227.CrossRefGoogle Scholar
Kupferman, Lucien. 2004. Le mot de: domains prépositionnels et domaines quantificationnels. Bruxuelles: de Boeck.CrossRefGoogle Scholar
Laenzlinger, Christopher. 2005. French adjective ordering: Perspectives on DP-internal movement types. Lingua 115, 645–689.CrossRefGoogle Scholar
Lamarche, Jacques. 1991. Problems for N-movement to NumP. Probus 3, 215–236.CrossRefGoogle Scholar
Landau, Idan. 2007. EPP Extensions. Linguistic Inquiry 5, 485–523.CrossRefGoogle Scholar
Legendre, Géraldine. 2000. Positioning Romanian verbal clitics at PF: An Optimality-theoretic analysis. In Gerlach, Birgit & Grijzenhout, Janet (eds.), Clitics in phonology, morphology, and syntax, 219–254. Amsterdam: John Benjamins.Google Scholar
Legrand, Jean. 1975. Or and any:The syntax and semantics of two logical operators. Ph.D. dissertation, The University of Chicago.Google Scholar
Lieber, Rochelle. 1992. Deconstructing morphology: Word formation in syntactic theory. Chicago, IL: The University of Chicago Press.Google Scholar
Longobardi, Giuseppe. 1994. Reference and proper names: A theory of N-movement in syntax and logical form. Linguistic Inquiry 25, 609–665.Google Scholar
Lyons, Christopher. 1999. Definiteness. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Mathieu, Éric. 2004. The mapping of form and interpretation: The case of optional WH movement in French. Lingua 114, 1090–1132.CrossRefGoogle Scholar
Milner, Jean-Claude. 1978. De la syntaxe à l'interprétation: Quantités, insultes, exclamations. Paris: Seuil.Google Scholar
Panagiotidis, Phoevos. 2000. Demonstrative determiners and operators: The case of Greek. Lingua 110, 717–742.CrossRefGoogle Scholar
Reinhart, Tanya. 1997. Quantifier scope: How labor is divided between QR and Choice Functions. Linguistics and Philosophy 20, 335–397.CrossRefGoogle Scholar
Rizzi, Luigi. 1997. The fine structure of the left periphery. In Haegeman, Liliane (ed.), Elements of grammar: Handbook in generative syntax, 281–337. Dordrecht: Kluwer.CrossRefGoogle Scholar
Sadler, Louisa & Arnold, Douglas. 1994. Pre-nominal adjectives and the phrasal/lexical distinction. Journal of Linguistics 30, 187–226.CrossRefGoogle Scholar
Shlonsky, Ur. 2004. The form of the Semitic noun phrase. Lingua 114, 1465–1526.CrossRefGoogle Scholar
Spector, Benjamin. 2003. Plural indefinite DPs as plural-polarity items. In Quer, Josep, Schroten, Jan, Scorretti, Mauro, Sleeman, Petra & Verheugd, Els (eds.), Romance languages & linguistic theory 2001, 295–313. Amsterdam: John Benjamins.CrossRefGoogle Scholar
Spencer, Andrew. 2000. Verbal clitics in Bulgarian: A Paradigm–Function approach. In Gerlach, Birgit & Grijzenhout, Janet (eds.), Clitics in phonology, morphology and syntax, 355–386. Amsterdam: John Benjamins.Google Scholar
Starke, Michal. 2001. Move dissolves into Merge: A theory of locality. Ph.D. dissertation, University of Geneva.Google Scholar
Stowell, Tim. 1981. Origins of phrase structure. Ph.D. dissertation, MIT.Google Scholar
Valois, Daniel. 1991a. The internal syntax of DP and adjective placement in French and English. NELS (North East Linguistic Society) 21, 367–382.Google Scholar
Valois, Daniel. 1991b. The syntax of DP. Ph.D. dissertation. University of California, Los Angeles.Google Scholar
Vogeeler, Svetlana & Tasmowski, Liliane. 2005. Les N, un N et des N en lecture générique. In Vogeleer, Svetlana & Tasmowski, Liliane (eds.), Déterminants indéfinis en français. Travaux de linguistique 50, 53–78.Google Scholar
Vogeeler, Svetlana & Liliane, Tasmowski. 2006. Introduction. In Vogeleer, Svetlana & Tasmowski, Liliane (eds.), Non-definiteness and plurality, 1–32. Amsterdam: John Benjamins.CrossRefGoogle Scholar
Wagner, Robert Léon & Jacqueline, Pinchon. 1962. Grammaire du français classique et moderne. Paris: Hachette.Google Scholar
Wilmet, Marc. 1986. La détermination nominale. Paris: Presses Universitaires de France.Google Scholar
Wolter, Lynsey. 2007. Situation variables and licensing by modification in opaque demonstratives. In Puig-Waldmüller, Estela (ed.), Sinn und Bedeutung 11, 612–626. Barcelona: Universitat Pompeu Fabra.Google Scholar
Zwicky, Arnold M. 1977. On clitics. Bloomington, IN: Indiana University Linguistics Club.Google Scholar