In this paper, the properties of Spanish DPs including a possessive pronoun and a relative clause are thoroughly described and analyzed. Adopting a raising analysis for both prenominal possessives and restrictive relatives, it is claimed that the incompatibility of a determiner possessive and a restrictive relative in current standard Spanish is due to the violation of an interpretive constraint sanctioning subextraction from [Spec, CP]. It is further proposed that, in constructions in which a possessive pronoun does combine with a relative clause, the possessive is not subextracted from [Spec, CP]. It is finally shown that this proposal accounts for different well-formed dialectal and Old Spanish patterns with a prenominal possessive and a restrictive relative and also applies to data from other Romance languages.
As described in Brucart (1994, 1999), the combination of a prenominal possessive and a restrictive relative clause in current general Spanish is deviant (1), whereas postnominal possessives are fully compatible with restrictive relatives (2), and a prenominal possessive does co-occur with an appositive relative (3): 
In this paper, I offer a new account of the facts in (1)–(3). Adopting Kayne’s (1994) raising analysis of restrictive relatives clauses, and also assuming the widely shared view that determiner possessives are first base-generated as nominal modifiers and then move to the DP domain, I will propose that the incompatibility of Spanish prenominal possessives and restrictive relatives illustrated in (1) is due to the violation of an interpretive constraint banning subextraction from [Spec, CP]. The idea is thus that, in this construction, the possessive, which is part of the raised relative “head” in [Spec, CP] of the relative clause, moves out of this position to reach the clause-external DP, giving rise to an anomalous output. As for the well-formed combinations in (2) and (3), I will argue that, in these cases, subextraction of the possessive from [Spec, CP] does not take place.
The content of the paper is organized as follows. In Section 2, previous analyses of the phenomenon under study are first critically reviewed. Section 3 sets out my assumptions on the derivation of possessives and relative clauses. In Section 4, my proposal on the incompatibility of possessives and restrictive relatives is presented. Section 5 addresses dialectal and Old Spanish data bearing on this topic. In Section 6, I extend my proposal to other Romance languages. Section 7 finally contains the main conclusions of this research.
2 Previous analyses
The incompatibility of determiner possessives and restrictive relative clauses is discussed, within the generative framework, in González Escribano (1995) and Brucart (1994). In this section, I will examine the proposals in these two papers, showing that they both have conceptual and empirical problems.
González Escribano (1995) focuses on present-day Standard English, where DPs containing a prenominal possessor and a restrictive relative are also prohibited (4a,b), while postnominal genitives are compatible with restrictive relatives (4c,d) (his examples and judgments):
González Escribano builds his account of the ungrammaticality of English examples like the ones in (4a,b) on the following main assumptions: (a) restrictive relative clauses have an identificational meaning and are adjoined to DPs; (b) the basic semantic import of the D head is “identification”, and their typical exponents are th-words (the, this...) and the ’s clitic; (c) the D clitic ’s, being a Case assigner, must be adjoined to a lexical DP in [Spec, DP] at S-Structure; and (d) inherently Case-marked specifiers of N (possessives) must also rise into [Spec, DP] and have their Case feature checked.
Under these assumptions, this author argues that, if the ID(entification) feature under D gets discharged by determiners like the or this, the ID feature will percolate up to the DP node and c-command the adjoined relative clause, allowing the correct coindexation and interpretation of the wh-item. However, if the ID feature “jumps” under the form –s onto a lexical DP landing in [Spec, DP], then it cannot percolate up to the node that functions as the antecedent of the relative (because the feature is not under the head D but under a Spec position), thus not being available as the required ID specification of an appropriate antecedent for the wh-item.
There are serious problems with this account, among them the following. Note, first, that González Escribano’s analysis presupposes that his Identification Procedure operates at S-Structure, which seems to be unmotivated (i.e. why comes that the ID feature on the D clitic –s does not percolate up to the DP node at D-Structure?). Moreover, for this proposal to be operative, it must be taken for granted that –s is merged under D, which is controversial: alternative analyses can be found in the literature, stating, for example, that –s is a possessive marker (i.e. a functional head Poss) heading a Poss(essor) Phrase in the extended projection of the noun, or a nominal copula that surfaces in contexts of DP-internal Predicate Inversion (see, e.g., Corver 2009: 87–88, and the references therein).
The main problem with the proposal under discussion is, in any case, that it calls for a [DP DP CP] analysis of restrictive relative clauses, which, as González Escribano (1995: 732) himself acknowledges, “clashes head-on with traditional doctrine”. It is generally assumed, at this respect, that restrictive relatives are interpreted as intersective modifiers of the nominal “head” (Partee 1975; Heim and Kratzer 1998), whereas in non-restrictive (appositive) relatives, a coreference relation is established between the relative pronoun and the whole DP “head”. This semantic difference has been related to different syntactic configurations: a restrictive relative is c-commanded by the determiner, which is not part of the relative “head”; in appositive relatives, the determiner is part of the relative “head” and does not c-command the relative clause. This has been structurally captured, for instance, by claiming that restrictive relatives are adjoined to a nominal projection (NP or N’), while appositive relatives are adjoined to the whole DP. The structural representation attributed to restrictive relatives by González Escribano (1995) therefore exactly corresponds to a well-known representation that has been traditionally assigned to appositive relatives (i.e. [DP DP CP]), thus blurring the distinction between these two main types of relative clauses.
In contrast to the proposal in González Escribano (1995), Brucart (1994) does adhere to the standard view that restrictive relatives are intersective modifiers, and adopts, in particular, a well-known version of the N-CP analysis, whereby these subordinate clauses are analyzed as N’ adjuncts (5a). As for determiner possessives, this author considers that they are XPs, which are first base-generated as complements of the noun, next move to [Spec, NP] (the position he assigns to prenominal possessives in Catalan-like languages), and end up landing in [Spec, DP], where they check their [+definite] feature (5b).
Against this framework, Brucart claims that the incompatibility of a determiner possessive and a restrictive relative in Spanish results from the violation of the Principle of incidence of nominal modifiers in (6), where the notion of “incidence” can be formalized in terms of Higginbotham’s (1985) thematic-identification in modification relations:
For Brucart, a nominal expression like the one in (1a), repeated in (7a), is thus ungrammatical because the relative clause, which follows all restrictive modifiers in a nominal expression, does not c-command the raised possessive, violating the principle in (6). Brucart’s analysis is represented in (7b):
The proposal in (7b) also faces some problems. First, the principle in (6) must apply at S-Structure: otherwise, the relative clause, which does c-command the possessive in its base position, would “affect” it in this position, and the example in (7a) should be well-formed. Moreover, it must be assumed that traces of moved items cannot be “affected” by a restrictive complement of the nominal head, but this seems to be just a stipulation, since no independent evidence is given for why this should be the case.
Second, Brucart’s account predicts that Spanish prenominal possessives should not combine with participial reduced relatives, given that this type of restrictive relative clauses, like full restrictive relatives, also occupy the right-most position in a nominal expression. This prediction is not borne out, however, as shown by the grammaticality of examples like the following: 
Third, as Brucart (1994: 52) points out, Spanish prenominal possessives are compatible with AP and PP restrictive modifiers, which can optionally follow a postnominal possessive pronoun or a prepositional genitive phrase (9b). Under Brucart’s analysis, the combination in (9a) should therefore be ruled out, since the determiner possessive is outside the c-command domain of the restrictive modifier also in this case (9c).
Finally, in some varieties of American Spanish (and in Old Spanish as well), a prenominal possessive preceded by the indefinite article co-occurs with a restrictive relative: 
If Brucart’s analysis in (7b) were applied to this case, as in (11), the prenominal possessive would now move to [Spec, NP] (as in Romance Catalan-like languages; cf. supra), a position that is not c-commanded by the restrictive relative clause either, again leading to a violation of the principle in (6). The American/Old Spanish construction illustrated in (10) should therefore be ungrammatical, contrary to facts.
In the light of the observations above, I must conclude that an alternative account of the (in)compatibility of prenominal possessives and restrictive relatives is required. Before developing my proposal in Section 4, I will make explicit my assumptions on the derivation of both possessives and relative clauses.
3 Assumptions on the derivation of possessives and relative clauses
I will first assume the commonly shared view in generative research that the prenominal and the thematic position of possessives are transformationally related (see, e.g., Cardinaletti 1998; Alexiadou et al. 2007; and the references therein). Focusing on alienable possession, I will suppose, in particular, that Romance possessive pronouns are generated as follows:  (a) postnominal possessives remain in its base position, and the postnominal order results from leftward N(P) movement across the possessive (see Cardinaletti 1998 for Italian; Bernstein 2001 for Spanish; and Brito 2007 for Portuguese) (12a); (b) prenominal possessives preceded by an overt determiner in Italian-like languages move to the specifier position of a functional nominal projection lower than D (cf. Picallo 1994; Cardinaletti 1998; Brito 2007; and Alexiadou et al. 2007, among others), arguably being attracted by the person feature in IP/AgrP (Brito 2007) (12b); and (c) the determiner possessive of Spanish (and French), which I take to be an XP, further rises to [Spec, DP] to have its [+definite] feature checked (cf. Brucart 1994; Ihsane 2003) (12c). 
I will further adopt Kayne’s (1994) version of the raising/promotion analysis of restrictive relatives, whereby the relative clause is a complement of an external D (i.e. [DP D0 CP]), and the relative “head” is first generated in the relativization site within the relative clause and then moves to [Spec, CP], as implemented, in particular, in Bianchi (1999, 2000) and de Vries (2002).  As represented in (13b), both Bianchi and de Vries hold that the relative “head”, being an argument within the relative clause (Borsley 1997), always originates as an indefinite relative DP with a relative morpheme in the D0 position (Drel), which is overt in wh-relatives and empty in non-wh-relatives. This DP then raises to [Spec, CP] for wh-checking, and the NP complement of Drel subsequently moves to the Spec of the relative DP in order to establish a local agreement relation with the external D, and check the φ-features and the [+N] selectional feature of the determiner. 
With respect to past participle relatives, I will follow Bhatt (1999) and Giurgea and Soare (2010) in assuming, contra Kayne (1994),  that reduced relatives are not CPs. As these linguists point out, the idea that reduced relatives lack a CP layer is supported by the fact that, in contrast to full restrictive relative clauses, they never include an overt complementizer or relative pronoun. Moreover, in reduced relatives, the relativized element always occupies the subject position, whereas in full relatives relativization is not restricted to any specific argument. Against this background, I will share, in particular, Giurgea and Soare’s (2010) view that participial constructions project to PredP: 
Finally, I will adopt a non-raising analysis of appositive relatives, like the one developed by de Vries’ (2002, 2006).  As depicted in (15b), de Vries claims that an appositive relative is a semi-free relative with a pronominal empty head (DP2), which is connected to the clause-external relative “head” (DP1) by specifying coordination (i.e. Annie, who is our manager = Annie, (namely/or, she/the one) who is our manager). In (15b), &: represents the head of a specifying coordination phrase, j and i have the same referent, and, at a discourse level, k = i.
With these assumptions in mind, I will next expose my proposal on the (in)compatibility of possessive pronouns and relative clauses in Spanish.
4 The proposal
The idea I would like to put forward is that the incompatibility of prenominal possessives and restrictive relatives in Spanish results from the violation of a well-known Condition on Extraction Domains establishing that [Spec, CP] renders syntactic objects in this position internally frozen (i.e. they can be extracted as a whole, but their subparts cannot be extracted) (see Gallego 2009, 2010, and the references therein). The effects of this restriction are illustrated by the anomalous English sentences in (16), in which wh-movement targets a subpart of an already wh-moved phrase: 
My account of the deviance of DPs like the one in (17a) would then go as follows. As depicted in the simplified structure in (17b), in this construction, the possessive is base-generated as a nominal modifier. It is thus part of the relative “head”, which originates in the relativization site within the relative clause and raises to [Spec, CP]. In order to have its [+definite] feature checked, the Spanish prenominal possessive further moves to the specifier position of the external D out of the moved relative “head”. The possessive is therefore subextracted from [Spec, CP], the condition above banning subextraction from this position is violated, and the construction is ill-formed.
The analysis in (17b) is inspired by the observations in Brucart (1999: §220.127.116.11), who attributes the incompatibility of prenominal possessives and restrictive relatives in Spanish to the double role the prenominal possessive plays in a nominal expression: on the one hand, it is interpreted as a modifier of the noun; on the other, it functions as a definite determiner (see (12c) in Section 3). This being the case, and assuming that the relative “head” in restrictive relatives includes both the noun and its complements, but excludes the determiner (see Section 2), there then arises a conflict, Brucart argues, between the fact that the possessive, as a nominal modifier, belongs to the relative “head” and its being a determiner, which implies that it cannot count as part of the antecedent in a restrictive relative (see also RAE: §44.8g).
Following Gallego (2010: 52), I take the failure to subextract from [Spec, CP] to be related to an interface constraint sanctioning ambiguous outputs. This interface constraint, which, more generally, applies to island effects on XPs moved to edge positions in the sentential left periphery, can be seen as a consequence of Chomsky’s (1986) Principle of Full Interpretation, requiring derivations, in the case at hand, to provide unambiguous instructions to the external systems (cf. Gallego 2009, 2010). I thus share Gallego’s position that the Condition on Extraction Domains banning subextraction from [Spec, CP] is not narrow syntactic, but has to be reinterpreted as an effect of the Principle of Full Interpretation. 
In my view, the interpretive (non-computational) nature of this constraint lies behind the following paradigm of general Spanish, which, to my knowledge, has never been observed (nor discussed) before: 
Under my judgments, and those of the native speakers I have asked about these kind of data, when uttered out of context, the combination of a prenominal possessive and a restrictive relative clause in a Spanish DP is just degraded, but not ungrammatical (18a) (see fn. 2). This combination improves if the relative “head” gets more descriptive content by adding restrictive modifiers to the noun (18b). And it is felicitous in a contrastive setting (18c). It comes as no surprise then that a number of examples of this construction, like the ones in (19), can in fact be found in the Spanish CREA database with an implicit partitive/contrastive interpretation: in the sentence in (19a), for instance, it is understood that the relatives of yours that had met him are a subset of your relatives, so that there are relatives of yours that had not met him. 
As said above, taking the ban on subextraction from [Spec, CP] to be an interpretive interface constraint can help us explain the paradigm in (18): the construction is anomalous, but not ungrammatical, and it improves under appropriate interpretive conditions. In a Spanish DP including a prenominal possessive and a restrictive relative, the determiner possessive plays a major role in establishing the reference of the whole DP, but it is also part of the content of the moved relative “head”, which is interpreted deep inside the relative clause. This causes a problem with the identification of the individual the DP refers to, which can be reduced (or even solved) through the addition of restrictive modifiers to the noun and/or the insertion of the expression in a contrastive context.
A similar scenario obtains in other cases of subextraction from XPs situated in the specifier position of a sentential left-peripheral projection. In the example in (20a), taken from Gallego (2009: 42), for instance, topicalization affects a subpart of an already topicalized DP, and the output is again anomalous. The acceptability of this sentence clearly improves, however, if the topic in the matrix clause has a contrastive interpretation and the DP in topic position of the subordinate clause is endowed with more descriptive content (20b).
Given my assumptions on the derivation of the different types of possessive pronouns and relative clauses in the previous section, the proposal on the deviance of the Spanish construction with a determiner possessive and a restrictive relative I have just presented also provides a straightforward explanation for all the well-formed combinations involving possessive pronouns and relative clauses described so far: what these constructions have in common is that the possessive is not subextracted from [Spec, CP] in the relative clause, so that the restriction sanctioning subextraction from an XP occupying this position is not violated.
Take first postnominal possessives, which, as we already know (see Section 1), combine with a restrictive relative (21a). As shown in (21b), a postnominal possessive remains in situ within the raised relative “head” (see Section 3). Therefore, as compared to prenominal determiner possessives, subextraction out of [Spec, CP] does not take place in this case.
Under de Vries’ (2002, 2006) non-raising coordination analysis of appositive relatives, there is no subextraction from [Spec, CP] in the well-formed construction with a Spanish prenominal possessive and an appositive relative illustrated in (22a) either. As represented in (22b), the determiner possessive now moves to the specifier of the clause-external DP relative “head”, and not out of [Spec, CP] of the relative clause, as in restrictive relatives.
Furthermore, in the acceptable sequence formed by a prenominal possessive and a participial reduced relative (23a), the determiner possessive is again not subextracted from [Spec, CP], since the participial relative, lacking a relative operator, does not contain a CP projection at all (cf. Giurgea and Soare 2010) (23b): 
To end up this section, I will take up again the observation in Section 2 that the Spanish determiner possessive is compatible with restrictive modifiers of the noun others than relative clauses (24a), and see how this falls in with my proposal. Focusing on restrictive adjectives, and assuming Cinque’s (2010) idea that adjectives are merged in the specifier of a series of functional projections dominating the NP, in nominal expressions including an adjectival modifier, the determiner possessive just moves out of its base position in the nominal domain and into [Spec, DP], as it does in nominal phrases without adjectival modifiers (24b).  The combination of a prenominal possessive and a restrictive adjective is therefore well-formed, as expected.
5 Dialectal variation and Old Spanish facts
Up to this point, I have dealt with core phenomena regarding the (in)compatibility of possessives and relative clauses in current general Spanish. I will now try to show that my proposal also applies to relevant dialectal and Old Spanish data bearing on this topic, which, as far as I know, have not been addressed in previous research within the framework of generative grammar. For ease of exposition, I will divide this section into two parts, corresponding to two different patterns: the [Poss + N + restrictive relative] pattern found in certain American Spanish dialects (and in Old Spanish), in which the prenominal possessive is a determiner introducing the nominal expression, and the [Det + Poss + N + restrictive relative] construction, where the prenominal possessive is preceded by a determiner, which co-existed with the [Poss + N + restrictive relative] construction in Old Spanish, and is also used, even though in a limited way, in some varieties of American Spanish nowadays.
5.1 [Poss + N + restrictive relative]
In the dialects of Spanish spoken in wide areas of Mexico and Central America, as well as in Andean countries like Peru or Bolivia, a determiner possessive does combine with a restrictive relative when the subordinate clause includes a verb denoting possession and/or a verbal affix or personal pronoun that identifies the same individual the possessive refers to.  This construction, in which there is a double reference to the possessor within the same DP, is illustrated in (25), (26) and (27) for Mexican, Bolivian and Peruvian Spanish, respectively (the examples are taken from Huerta Flores 2009 and CREA).  In the rest of Spanish-speaking countries, including Spain, the construction in (25)–(27) is not used, and the definite article consistently replaces the possessive.
As has been noticed (RAE 2009: §18.3k, §44.8g), it is just in the aforementioned dialects of Central and South American current Spanish that the so-called doubled possessive construction can also be found. As shown by the American Spanish examples in (28) (from Company Company 2001; Huerta Flores 2009), in the doubled possessive construction, the possessor is again expressed twice in a nominal expression, by means now of a third person prenominal possessive pronoun, on the one hand, and through a postnominal prepositional genitive phrase containing a DP with the same reference, on the other: 
Meaningfully, the [Poss + N + restrictive relative] pattern in (25) – (27) and the doubled possessive construction in (28) are also widely attested in Old Spanish (examples taken from the CORDE data-base, Company Company 1993 and Huerta Flores 2009): 
There thus seems to be a strong correlation, both in American Spanish and in Old Spanish, between the occurrence of the doubled possessive construction and the existence of the pattern with a prenominal possessive and a restrictive relative containing morphological or lexical material that also makes reference to the possessor. A natural way to capture this correlation is by assigning the same analysis to the possessive determiner in the two constructions.
The idea I would like to entertain on this point, building on the observations in Company Company (2001) and RAE (2009: §18.3k), is that, in the Spanish doubled possessive construction, the possessive is a definiteness marker that is in place of the definite article, lacking a true possession content, which is conveyed by the postnominal possessive phrase: i.e. an American Spanish nominal expression like su mamá de él ‘his mum of he’ in (28a), for example, is basically interpreted as la mamá de él ‘the mum of he’.
In addition to being a definiteness marker, the possessive in the doubled possessive construction, as compared with the alternating nominal expression with the definite article, evaluates the relation between two entities, indicating that, in the speaker’s view, there exists an intrinsic or close relation between the possessor and the possessum (cf. Company Company 1993, 1994, 1995a, 1995b, 2001 for Old Spanish and American Spanish; Fernández 2013 for Bolivian Spanish; and Risco 2013 for Peruvian Spanish). The speaker will thus choose between the definite article and the possessive depending on the meaning he wants to convey. This characterization of the prenominal possessive in the doubled possessive construction of both American Spanish and Old Spanish is supported by the fact that this construction preferably includes nouns that express inherent or close relations, like kinship terms, body parts, and nouns denoting private or daily use objects, social or personal links, intrinsic abstract properties of the possessor, etc. (see Company Company 1993, 1994; Huerta Flores 2009; Fernández 2013).
The insight that the prenominal possessive in the Spanish doubled possessive construction is a definiteness marker without possession content can be formally expressed by directly merging it in the head position of the DP (31b): 
In my view, the construction including a determiner possessive and a restrictive relative in American and Old Spanish illustrated in (25)–(27) and (29) is to be analyzed in a similar way. Take, for instance, the American Spanish example in (25a): Su cámara digital que se compró ‘His digital camera that he bought’. It is crystal clear that the possessive does not have a possession content here: this expression cannot mean that the individual at hand bought a camera that he already owned, which is nonsense. Therefore, in this construction, as in the doubled possessive construction, the possessive does not have a true possession meaning and serves as a definiteness marker: putting it in RAE’s (2009: 1352) words, “it can be thought that there is no incompatibility between the prenominal possessive and the restrictive relative clause because, in the former, the reference to the possessor is redundant, so that only the definiteness features are interpreted”. Furthermore, as in the doubled possessive construction, also in this case, by using the possessive instead of the article the speaker seeks to emphasize that, in his view, the possessor has an inherent or close relation with the possessum. 
The analysis I have in mind for American Spanish or Old Spanish DPs like the one in (32a) is represented in (32b): the prenominal possessive merges as a definiteness marker under the clause-external D head; it is therefore not subextracted from the raised relative “head” in [Spec, CP], and the construction is felicitous.
In support of this proposal, it can be argued that an analysis along these lines is independently needed in order to account for nominal expressions introduced by the so-called emphatic possessive, which is attested in all dialects of current Spanish. The emphatic possessive, illustrated in the examples in (33), is a reflexive determiner possessive (i.e. it must have an antecedent in the same sentence) that “is used in expressions denoting overrating, closeness, and various affective nuances”, or indicates “the attribution of typical or characterizing properties to an entity, as well as its participation in events or situations that affect it, especially if the speaker associates them with some stereotype” (RAE 2009: 1367):
The emphatic possessive is also devoid of a possession content, as corroborated by the fact that it never has the contrastive interpretation a true possessive is amenable to have, and can often be either omitted, or replaced by a determiner: instead of uttering, for instance, the sentences in (33), we can also say in Spanish Yo también tengo problemas ‘I also have problems’ or Todos los días se compra el periódico y se toma un cafetito ‘Everyday, she buys the newspaper and takes a cup of coffee’, without a significant loss in meaning.
The properties of the Spanish emphatic possessive can again be structurally captured by positing that this lexical item is base-generated within the DP domain, and not as a complement of the noun:
Under this analysis, it is predicted that an emphatic possessive should combine with a restrictive relative clause. This prediction is borne out, as shown by the following examples:
My proposal for the well-formed construction in (35) with an emphatic possessive and a restrictive relative is represented in (36b): as in the [Poss + N + restrictive relative] pattern of American Spanish and Old Spanish in (32), the emphatic possessive is inserted under D0, and does not move out of the raised relative “head” in [Spec, CP].
This analysis could also be applicable to examples like the ones in (37), which are documented in the CREA database, in which a determiner possessive and a restrictive relative co-occur:
In my view, in the examples in (37), the possessive does not properly express a relation of possession or belonging either, but rather seems to behave as the Spanish “affective possessive” preceding a proper name illustrated in the sentences in (38), taken from RAE (2009), where “a particular individual is not chosen from a set of people with the same name, but the existence of an affective tie between the person designated by the possessive and the one the proper name refers to is marked instead” (RAE 2009: 1355).
5.2 [Det + Poss + N + restrictive relative]
In Old Spanish, a prenominal possessive could function as a determiner, as in (29) and (30) above, but it could also follow the definite article and, in this case, the possessive once more combined with a restrictive relative clause including a possession denoting verb and/or a verbal affix or personal pronoun that refers to the individual the possessive refers to. This is illustrated in the CORDE examples in (39):
Not surprisingly, the doubled possessive construction also obtained in Old Spanish when the possessive was preceded by the definite article (examples found in the CORDE database):
This correlation calls for a uniform analysis of the prenominal possessive in (39) and (40). As represented in (41) and (42), I propose at this respect that, in both constructions, the possessive, which again lacks a true possession meaning, does not originate as a complement of the noun, but is merged as an agreement marker located in the head position of an IP/Agr Phrase (see Section 3) in the extended projection of the noun:
As illustrated in the examples in (43) from Company Company (2009) and CREA, a different instantiation of the [Det + Poss + N + restrictive relative] pattern, which was already presented in Section 2, is used in some areas of Central and South America, where a prenominal possessive can follow the indefinite article, and the possessive is fully compatible with a restrictive relative:
As shown in (44), this construction is also widely attested in Old Spanish (cf. Company Company 2009):
As a matter of fact, in Old Spanish, a prenominal possessive could also follow a demonstrative, a cardinal numeral or a quantifier (cf. RAE 2009: §18.2l; Company Company 2009) and, in all these cases, it combined with a restrictive relative as well (examples from CORDE):
The well-formedness of the pattern in (43)–(45) easily falls under my proposal. As depicted in (46b) and (47b), in this case, the possessive moves to [Spec, IP] (the landing site of the prenominal possessive in Italian-like languages; see (12b) in Section 3) within the raised relative “head” in [Spec, CP] of the relative clause, but it does not reach the clause-external DP. There is thus no subextraction from [Spec, CP], and the construction is O.K.
6 Extending the analysis to other Romance languages 
In this final section, I will apply my proposal to data from other Romance languages. First, my analysis of the incompatibility of prenominal possessives and restrictive relatives in current general Spanish (see Section 4) naturally extends over to French, where a prenominal possessive is also a determiner. As illustrated in (48), in French, as in Spanish, DPs including both a determiner possessive and a restrictive relative clause are anomalous, which can be equally attributed to a violation of the condition sanctioning subextraction from [Spec, CP]. 
Second, as shown in (49), in languages in which the prenominal possessive is preceded by a determiner, like Catalan, Italian or Portuguese, a possessive following the indefinite article co-occurs with a restrictive relative clause.  The analysis in (46b) for this pattern in Old Spanish (and in some dialects of American Spanish) applies to the well-formed Catalan, Italian and Portuguese data in (49) as well.
Finally, as illustrated in (50a) for Catalan, in these Romance languages, the combination of a restrictive relative clause and a prenominal possessive preceded by the definite article is deviant, as pointed out in Brucart (1994, 2002). Nevertheless, the grammaticality of this combination improves when a restrictive modifier is added to the noun in the relative “head” (50b), and the construction is acceptable in a contrastive setting (50c): 
Note that the paradigm in (50) exactly parallels that of the Spanish construction with a determiner possessive and a restrictive relative clause in (18) (see Section 4), repeated in (51):
This fact can be accounted for, in my view, by analyzing anomalous Romance data like the one in (50a) along the lines of my account of the corresponding Spanish construction in (51a). I would thus like to argue that in the [definite article + possessive] sequence of Catalan, Italian and Portuguese (e.g., Cat. el seu llibre ‘the his book’), the article is an expletive filling the D0 position, and the possessive rises to [Spec, DP] at LF. Therefore, in the deviant Catalan example in (50a), the possessive, which is pronounced in [Spec, IP] of the relative “head” (see Section 3), covertly moves out of [Spec, CP], and the construction is ruled out by the restriction sanctioning subextraction from [Spec, CP]. As in Spanish, given the interpretive nature of this restriction (see Section 4), the construction in (50a) gets better when restrictive modifiers are added to the noun (50b), contributing to the identification of the individual the DP refers to, and is fully acceptable in a contrastive context (50c).
The insight behind this proposal is that the sequence formed by the definite article and a prenominal possessive in Catalan-like languages constitutes a CHAIN (in Chomsky’s 1986 sense) that semantically corresponds to the determiner possessive in Spanish-like languages, so that the interpretation of this sequence is determined by the possessive, and not by the article, which lacks substantive semantic content.  Independent evidence for this idea comes from paradigms like the ones illustrated in (52) and (53): 
In (52a), the Catalan article has a quantitative or qualitative interpretation, as expressed in the English translation. In (52b), the presence of the prenominal possessive makes it impossible for the article to convey this meaning, and the sentence is ungrammatical. This shows that it is the possessive, and not the article, that determines the interpretive properties of a DP introduced by the definite article and a possessive in Catalan-like languages.
As for the paradigm in (53), concerning emphatic possessives (see Section 4), the sentence in (53a) is ill-formed due to the definiteness effect affecting the complement position of the verb tenir ‘to have’. The very same sentence is, however, well-formed if the nominal expression in object position includes an emphatic possessive, as in (53b). This clearly indicates that the article in a sequence like el seus problemes ‘the his problems’ in (53b) cannot have a definite interpretation, and must be an expletive. Otherwise, this sentence should also be ruled out, just as (53a) is.
In this paper, I have analyzed a number of constructions including a possessive pronoun and a relative clause in Spanish (and in Romance more generally). The main conclusions to be drawn from this research are the following. First, adopting a raising analysis for both determiner possessives and restrictive relatives, I have argued that the incompatibility of determiner possessives and restrictive relative clauses in Spanish (and French) is due to the violation of an interpretive condition sanctioning subextraction from [Spec, CP]. Second, I have proposed that, in all well-formed Romance constructions in which a possessive co-occurs with a relative clause, subextraction from [Spec, CP] does not take place. And third, I have finally claimed that my proposal can also explain the anomalous status of the Catalan, Italian or Portuguese [Art[+def]+ Poss + N + restrictive relative] pattern, under the assumption that, in the sequence formed by the definite article and a prenominal possessive, the article is an expletive and the possessive rises to the DP domain at LF.
I would like to thank Ángel Gallego, Francesc Roca, and two anonymous reviewers for their comments on the content of this paper. Research for this study has been supported by a grant to the project FFI2014-56968-C4-3-P.
Alexiadou, Artemis, Liliane Haegeman & Melita Stavrou. 2007. Noun phrase in the generative perspective. Berlin: Mouton de Gruyter. Search in Google Scholar
Bernstein, Judy B. 2001. Focusing the “right” way in Romance determiner phrases. Probus 13(1). 1–29. Search in Google Scholar
Bhatt, Rajesh. 1999. Covert modality in non-finite contexts. University of Pennsylvania dissertation. Search in Google Scholar
Bhatt, Rajesh. 2002. The raising analysis of relative clauses: Evidence from adjectival modification. Natural Language Semantics 10. 43–90. Search in Google Scholar
Bianchi, Valentina. 1999. Consequences of antisymmetry: Headed relative clauses. Berlin: Mouton de Gruyter. Search in Google Scholar
Bianchi, Valentina. 2000. The raising analysis of relative clauses: A reply to Borsley. Linguistic Inquiry 31(1). 123–140. Search in Google Scholar
Bianchi, Valentina. 2002. Headed relative clauses in generative syntax. Part II. Glot International 6(8). 1–13. Search in Google Scholar
Borsley, Robert D. 1997. Relative clauses and the theory of phrase structure. Linguistic Inquiry 28(4). 629–47. Search in Google Scholar
Brito, Ana Maria. 2007. European Portuguese possessives and the structure of DP. Cuadernos de Lingüística del I.U.I. Ortega y Gasset 14. 27–50. Search in Google Scholar
Brucart, José M. 1994. Sobre una incompatibilidad entre posesivos y relativas especificativas. In Violeta Demonte (ed.), Gramática del español, 51–86. México: El Colegio de México. Search in Google Scholar
Brucart, José M. 1999. La estructura del sintagma nominal: las oraciones de relativo. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva de la lengua española, vol. 1, 395–522. Madrid: Espasa. Search in Google Scholar
Brucart, José M. 2002. Els determinants. In Joan Solà, Maria-Rosa Lloret, Joan Mascaró & Manuel Pérez Saldanya (eds.), Gramàtica del català contemporari, vol. 2, 1435–1516. Barcelona: Editorial Empúries. Search in Google Scholar
Cardinaletti, Anna. 1998. On the deficient/strong opposition in possessive systems. In Artemis Alexiadou & Chris Wilder (eds.), Possessors, predicates and movement in the determiner phrase, 17–53. Amsterdam: John Benjamins. Search in Google Scholar
Chomsky, Noam. 1986. Knowledge of language. Its nature, origin, and use. New York: Praeger. Search in Google Scholar
Chomsky, Noam. 2001. Derivation by phase. In Michael Kenstowicz (ed.), Ken Hale. A life in language, 1–52. Cambridge, MA: MIT Press. Search in Google Scholar
Chomsky, Noam. 2008. On phases. In Robert Freidin, Carlos P. Otero & María Luisa Zubizarreta (eds.), Foundational issues in linguistic theory, 133–166. Cambridge, MA: MIT Press. Search in Google Scholar
Cinque, Guglielmo. 1994. On the evidence for partial N-movement in the Romance DP. In Guglielmo Cinque, Jan Koster, Jean-Yves Pollock, Luigi Rizzi & Raffaella Zanuttini (eds.), Paths towards universal grammar: Studies in honor of Richard S. Kayne, 85–110. Washington, DC: Georgetown University Press. Search in Google Scholar
Cinque, Guglielmo. 2010. The syntax of adjectives: A comparative study. Cambridge, MA: MIT Press. Search in Google Scholar
Company Company, Concepción. 1993. Su casa de Juan. Estructura y evolución de la duplicación posesiva en el español. In R. Penny (ed.), Actas del I Congreso Internacional Anglo-Hispano, vol. I, 73–86. Madrid: Castalia. Search in Google Scholar
Company Company, Concepción. 1994. Semántica y sintaxis de los posesivos duplicados en el español de los siglos XV y XVI. Romance Philology 48(3). 111–135. Search in Google Scholar
Company Company, Concepción. 1995a. Cantidad vs. cualidad en el contacto de lenguas. Una incursión metodológica en los posesivos ‘redundantes’ del español americano. Nueva Revista de Filología Hispánica 43(2). 305–339. Search in Google Scholar
Company Company, Concepción. 1995b. Old forms for new concepts. The recategorization of possessive duplication in Mexican Spanish. In H. Andersen (ed.), Historical linguistics 1993, 77–93. Amsterdam: John Benjamins. Search in Google Scholar
Company Company, Concepción. 2001. Gramaticalización, debilitamiento semántico y reanálisis. El posesivo como artículo en la evolución sintáctica del español. Revista de Filología Española 81(1–2). 49–87. Search in Google Scholar
Company Company, Concepción. 2009. Artículo + posesivo + sustantivo y estructuras afines. In Concepción Company Company (dir.), Sintaxis histórica de la lengua española. Segunda parte: La frase nominal, vol. I, 761–880. México: Fondo de Cultura Económica. Search in Google Scholar
CORDE (Corpus de Referencia del Español). http://www.rae.es (accessed May 2014). Search in Google Scholar
Corver, Norbert. 2009. Getting the (syntactic) measure of measure phrases. The Linguistic Review 26. 67–134. Search in Google Scholar
CREA (Corpus de Referencia del Español Actual). http://www.rae.es (accessed May 2014). Search in Google Scholar
Donati, Caterina & Carlo Cecchetto. 2011. Relabeling heads: A unified account for relativization structures. Linguistic Inquiry 42(4). 519–560. Search in Google Scholar
Fernández, Víctor. 2013. El español de los inmigrantes de los Andes bolivianos en el Norte Grande de Chile: convergencias y divergencias dialectales en el marco de una situación de contacto. Université de Montréal dissertation. Search in Google Scholar
Gallego, Ángel J. 2007. Phase theory, case and relative clauses. Anuario del Seminario de Filología Vasca Julio de Urquijo (ASJU) 41(2). 71–94. Search in Google Scholar
Gallego, Ángel J. 2009. On freezing effects. Iberia 1(1). 33–51. Search in Google Scholar
Gallego, Ángel J. 2010. Subextraction from phase edges. In E. Phoevos Panagiotidis (ed.), The complementizer phrase: Subjects and operators, 51–75. Oxford: Oxford University Press. Search in Google Scholar
Gallego, Ángel J. 2011. Successive cyclicity, phases, and CED effects. Studia Linguistica 65(1). 32–69. Search in Google Scholar
Gallego, Ángel & Juan Uriagereka. 2007. Conditions on sub-extraction. In Luis Eguren & Olga Fernández Soriano (eds.), Coreference, modality, and focus: Studies on the syntax-semantics interface, 45–70. Amsterdam: John Benjamins. Search in Google Scholar
Giurgea, Ion & Elena Soare. 2010. Predication and the nature of non-finite relatives in Romance. In Anna Maria di Sciullo & Virginia Hill (eds.), Edges, heads and projections, 191–214. Oxford: John Benjamins. Search in Google Scholar
González Escribano, José Luis. 1995. On the incompatibility of genitives and restrictive relative clauses: An explanation within the theory of principles and parameters. Linguistics 33. 711–740. Search in Google Scholar
de Granda, German. 1997. Replanteamiento de un tema controvertido. Génesis y retención del doble posesivo en el español andino. Revista de Filología Española 77(1/2). 139–147. Search in Google Scholar
Heim, Irene & Angelika Kratzer. 1998. Semantics in generative grammar. Oxford: Blackwell. Search in Google Scholar
Hernanz, M. Lluïsa. 1994. Concordancia, rección y aspecto: las construcciones absolutas en español. In Alegría Alonso, Beatriz Garza & José A. Pascual (eds.), II encuentro de lingüistas y filólogos de España y México, 367–402. Salamanca: Universidad de Salamanca. Search in Google Scholar
Higginbotham, James. 1985. On semantics. Linguistic Inquiry 16(4). 547–593. Search in Google Scholar
Huerta Flores, Norohella. 2009. Los posesivos. In Concepción Company Company (dir.), Sintaxis histórica de la lengua española. Segunda parte: La frase nominal, vol. I, 607–757. México: Fondo de Cultura Económica. Search in Google Scholar
Ihsane, Tabea. 2003. A typology of possessive modifiers. In Martine Coene & Yves D’hulst, From NP to DP, vol. 2, 23–42. Amsterdam: John Benjamins. Search in Google Scholar
Kayne, Richard S. 1994. The antisymmetry of syntax. Cambridge, MA: MIT Press. Search in Google Scholar
Lasnik, Howard & Mamoru Saito. 1992. Move α: Conditions on its applications and outputs. Cambridge, MA: MIT Press. Search in Google Scholar
Longobardi, Giuseppe. 1994. Reference and proper names. Linguistic Inquiry 25(4). 609–665. Search in Google Scholar
Partee, Barbara H. 1975. Montague grammar and transformational grammar. Linguistic Inquiry 6. 203–300. Search in Google Scholar
Picallo, Carme. 1994. Catalan possessive pronouns: The Avoid Pronoun Principle revisited. Natural Language and Linguistic Theory 12(2). 259–299. Search in Google Scholar
Real Academia Española (RAE). 2009. Nueva gramática de la lengua española. Madrid: Espasa Libros. Search in Google Scholar
Reyes, Doriam. 2000. La alternancia artículo/posesivo en frases nominales especificadas. Universidad Nacional Autónoma de México master thesis. Search in Google Scholar
Risco, Roxana. 2013. Nuevas perspectivas teóricas y empíricas: el doble posesivo de tercera persona. Escritura y Pensamiento 32. 41–71. Search in Google Scholar
Roca, Francesc. 1996. La determinación y la modificación en español. Universitat Autònoma de Barcelona dissertation. Search in Google Scholar
Sánchez, Liliana. 1996. Syntactic structure in nominals: A comparative study of Spanish and Southern Quechua. University of Southern California dissertation. Search in Google Scholar
Torrego, Esther. 1985. On empty categories in nominals. Ms., U. Mass Boston. Search in Google Scholar
de Vries, Mark. 2002. The syntax of relativization. Utrecht: LOT. Search in Google Scholar
de Vries, Mark. 2006. The syntax of appositive relativization: On specifying coordination, false free relatives, and promotion. Linguistic Inquiry 37(2). 229–270. Search in Google Scholar
© 2017 Walter de Gruyter GmbH, Berlin/Boston