Existential Constructions, Definiteness Effects, and Linguistic Contact: At the Crossroads between Spanish and Catalan

Abstract

Existential sentences in Spanish are sensitive to the definiteness or quantification restriction or effect, which prevents personal pronouns, proper nouns, and definite constituents from occupying the pivot position. Contact varieties between Spanish, a robust language as regards the effect, and Catalan, which has a weaker version, remain largely unexplored. This paper fills this void. A large corpus was gathered to quantitatively study the variation between definite and indefinite pivots. Examples involving definite, specific pivots and even proper names, hitherto unnoticed, are brought to the fore. The pivot of the existential in Spanish is argued to bear Partitive case, as shown by (i) pronominal existential pivots in other Romance languages, (ii) the phi-feature defectiveness of the clitic out of the pivot position, (iii) and partitive pronouns with unaccusatives in Spanish. The hypothesis is put forth that varieties of Spanish in contact with Catalan no longer relate Partitive case to the non-definiteness of the pivot.

1. Introduction

The aim of this paper is to provide some insight into how existential sentences in Spanish, a robust language as regards Milsark’s (1974, 1977) definiteness effect, allow for definite, specific, or even personal names as pivots of the existential in those varieties in contact with Catalan, which has a weaker (or null) version of the definiteness effect. This research paper is based on data that has remained hitherto unnoticed. Quantitative results, when obtained from rural varieties, show linguistic variation between Catalan and non-Catalan regions. The hypothesis will be put forth that Spanish in contact with Catalan amnesties the definiteness effect. Linguistic variation will be argued to boil down to the fact that pivots of existential haber, ‘to be’, in contact varieties are full-fledged Determiner Phrases, and they are hence endowed with person features.

Even though the proposal I develop here includes quantitative modeling of dialectal results, the explanations I pursue are couched within Chomsky’s Minimalist Program (Chomsky 1995, 2000). I assume, particularly, Abney’s (1987) Determiner Phrase (henceforth DP) Hypothesis (see Eguren 2006, 2008; Cañas García 2022 for Spanish), further developed by Giusti (1991) and Cardinaletti (1994), among many others.

1.1. Basic Descriptive Data

Existential sentences in Spanish are known to be sensitive to what Milsark (1974, 1977) referred to as definiteness or quantification restriction or effect, which bars definite or specific constituents from the pivot position1. This fact is illustrated in (1):

(1)	a.	Había	gente	por		todas		partes.
		there.were	people	for		every		parts
		‘There was people everywhere.’ (CORPES XXI, Mexico)
	b.	Había	un	olor		raro
		there.was	a	smell		weird
		‘There was a weird smell.’					(CORPES XXI, Argentina)
	c.	*Había	él	en		la		habitación.
		there.was	he	in		art		room
		‘There was him in the room.’
	d.	*Hay	Inés	en		la		habitación.
		there.is	Inés	in		art		room
		‘There is Inés in the room.’
	e.	Hay	{*el ~ *tu}		libro	en		la	habitación.
		there.is	{the ~ your}		book	in		art	room
		‘There is {the ~ your} book in the room.’
	f.	Los	hay		que	descansan		sobre	sial.
		acc.3pl	there.is		rel	rest		on	sial.
		‘There are those that lie on sial.’				(CORPES XXI, Peru)

The pivot position of the existential sentence can be occupied by bare noun phrases (1a) and indefinite noun phrases (1b). Definiteness, nevertheless, seems to be restricted. Let us tentatively assume Aissen’s (2003) Definiteness Scale for explanatory purposes, as stated in (2):

(2)	Definiteness Scale
	Personal pronoun > Proper name > Definite NP > Indefinite specific NP > Non-specific NP
	(Aissen 2003, p. 437)

The definiteness effect prevents personal pronouns (1c), proper nouns (1d), and definite constituents (1e) from occupying the pivot position. In the examples, the asterisk marks the ungrammaticality of the sentence. Pronominal clitics (1f) are grammatical. I assume here, namely as a matter of execution, that they receive Partitive case, as does the pivot of the existential (Belletti 1987, 1988); this is an idea that I have argued for elsewhere2 (see also De Benito Moreno 2016 and Fernández-Ordóñez 2019 for some observations on this clitic). Quantified constituents, such as those headed by cada, ‘each’, or todo, ‘every’, are also barred, unless they quantify over types (McNally 1997).

This prohibition against definiteness (or specificity) seems to be amnestied under certain conditions (cfr. RAE and ASALE 2009, §15.6; Leonetti 1999), such as negation (3a), locative interpretation (and locative coda preposing; see Leonetti 2008) (3b), list reading interpretation (3c), superlatives (3d), or certain restrictive modifiers (Rando and Napoli 1978; Ziv 1982; Lumsden 1988; Abbott 1993, 1999; McNally 1997, 2016).

(3)	a.	No	había		el	ferricarril.
		neg	there.was		art	train
		‘There was not the train.’
		(CORPES XXI, Spain)
	b.	En	la		calle	solo	había	el
		in	art		street	only	there.is	art
		movimiento	de		las	ramas.
		movement	of		art	branches.
		‘In the street, there was only the hanging of branches.’
		(CORPES XXI, Chile)
	c.	Había	el	declive		y	la	brecha.
		there.was	art	decay		and	art	gap
		‘There was the decay and the gap.’
		(CORPES XXI, México)
	d.	No	había		el	más	mínimo	presagio.
		neg	there.was		art	most	minimum	foreboding
		‘The was not a single foreboding.’
		(CORPES XXI, Colombia)

A complete discussion of the examples in (3) lies far beyond the scope of this paper. Nevertheless, a few observations are in order. It is tempting to analyze (3) as being headed by a weak definite determiner (cfr. Abbott 1993, 2006 for alternative analyses). I will leave this possibility aside for the moment. Irrespective of the specific analysis of the examples in (3), it stands to reason that these definite DPs share their distribution with bare and indefinite noun phrases. Personal pronouns (1e) or proper nouns (1d), nevertheless, are not grammatical as pivots of haber, ‘to be’, even if negative, locative, or list readings, as those in (3), are triggered. If we assume Cardinaletti’s (1994) proposal about the derivation of strong pronouns, these facts can be derived in a straightforward fashion: what is ruled out of existentials in Spanish is movement from N° to D°. I will not follow this possibility any further.

The definiteness effect Is thought to be subject to wide cross-linguistic variation, even within the same linguistic family. It is robust in languages like Spanish (cfr. (1), (3)), Galician, or French (cfr. Bouchard 1997; Paykin and Van de Velde 2021), whereas Catalan and Italian have weaker versions of it (see, among others, Leonetti 2008). Catalan allows proper nouns (4a) and definite constituents (4b), as shown in Rigau (1988) or Brucart and Rigau (2006), in the pivot position of haver-hi, ‘to be’, without additional interpretations, contrary to the Spanish cases of (3):

(4)	a.	Hi	havia	la	Marta.
		loc	there.was	art	Marta.
		‘There was Marta.’			(Ramos 1998, p. 45)
	b.	Hi	havia	el	president.
		loc	there.was	art	president
		‘There was the president.’			(Rigau 1997, p. 396)

Personal pronouns, at the beginning of Aissen’s (2003) Definiteness Scale, have been labeled as ungrammatical by some grammarians (e.g., Rigau 1988; Ramos 1998; or Brucart 2006). In fact, Rigau (1988) derives this ungrammaticality from Belletti’s (1987, 1988) proposal that the pivot of the existential is marked with Partitive case, and not with Accusative case (see also Chomsky 1995, §2. 6. 4); hence, nominative personal pronouns are excluded. Several authors (e.g., Villalba 2016; Gràcia i Solé and Roca Urgell 2017) have casted some doubt on this generalization: there are thought to be some rather strict conditions under which personal pronouns in Catalan are acceptable, such as coordination, or the use of adverbials such as nomès, ‘only’, or també, ‘also’ (cfr. Cruschina 2016, who reports similar facts in Italian). What is relevant here is that Catalan, when compared to Spanish, goes a step further as regards the definiteness effect: definite and specific constituents are grammatical in the pivot position of the existential construction without the negative, locative, or list readings in (3) that are required in Spanish for definite pivots. Additionally, proper nouns are acceptable in Catalan, and, quite plausibly, so are personal pronouns under certain strict conditions. Existential haber, ‘to be’, in Spanish rejects them categorically, no matter the circumstances.

1.2. Aim and Scope of the Paper

The data from varieties in contact between Spanish and Catalan, nonetheless, remain largely unexplored. The question as to whether the definiteness effect holds whenever Spanish and Catalan coexist in the same territory has not been addressed. Accordingly, the aim of this study is to shed some light onto the weakening (or even lack) of definiteness effects in a linguistic contact setting between Spanish, a robust language as regards the effect, and Catalan, which has a weaker version of it. Particularly, the contrast between (1, 3) and (4) will be tackled from a quantitative perspective.

The significance of this study is threefold. First, grammatical restrictions, to which close attention was paid in the Government and Binding framework (Chomsky 1980, 1981), remain an interesting area of study, as they may help sharpen our understanding of more abstract principles of grammar when subject to corpus-based studies. Second, the study of grammatical variation in linguistic contact settings may further our knowledge of how grammatical variables are distributed across a geographic continuum. Third, and most importantly, traditional dialectology has concentrated mainly on phonetic, phonological, or lexical variation. As a consequence, syntactic variation within contact varieties still awaits further research.

1.3. Previous Accounts

Grammatical restrictions, which were a prolific line of inquiry in both Standard Theory and Principles and Parameters theory (e.g., Ross 1967), have raised intriguing issues relating to theoretical and empirical adequacy since then. Recent minimalist proposals have tried to make grammatical restrictions dependent on more general principles of grammar, so that their effects are derived in a natural and elegant fashion. The study of long-distance syntactic dependencies is a representative example in this respect: what was derived by resort to constrains on movement is now accounted for in terms of syntactic island or phases (see Chomsky 1995; Boeckx 2006 and references cited therein). There remain, nevertheless, some restrictions which have mounted strong resistance to this tendency.

Existential sentences have been the focus of attention of a growing amount of literature (e.g., Milsark 1974, 1979; Breivik 1981; Freeze 1992; Dobrovie-Sorin 1997; Francez 2009; McNally 2016; Partee and Borschev 2007). Since Milsark (1974) and Safir (1982), existential sentences are known to be subject to the universality, quantification or definiteness restriction or effect. The relevant data are shown in (5, 6). The definiteness effect bars definite (5), specific (Enç 1991), or quantificational constituents (6) (cfr. Keenan and Stavi 1986; McNally 1998; Zucchi 1995, 2003) from the pivot position of existential there.

(5)	a.	*There is {the ~ John’s ~ that} dog in the room.
	b.	*There is {John ~ him} in the room.
(6)	a.	*There are {all ~ both} dogs in the room.
	b.	*There is {every ~ each} dog in the room.
		(Milsark 1974, p. 195)

Milsark (1977, 1979) subsequently refined some of his explanations. Milsark’s observations were soon tested not only in English, but also vis à vis some Romance languages. The advent of Abney’s (1987) Determiner Hypothesis popularized the Definiteness Effect as a diagnosis of definiteness, specificity, generic interpretations, or the strengths or weaknesses of certain determiners (e.g., Keenan and Stavi 1986; Dobrovie-Sorin 1997; Zamparelli 2000; Eguren 2008; Barker 2019; Ladusaw 1994; Leonetti 1999; Fernández Soriano and Baylín 1999; Gutiérrez Rodríguez 2008; Cruschina 2012; Mare 2016; Barker 2019; Keenan 2019).

Theoretical explanations of the definiteness effect may be classified as to whether more or less abstract syntactic principles are invoked (e.g., Safir 1982; Longa et al. 1996, 1998; Belletti 1987, 1988; La Fauci and Loporcaro 1997; Rodríguez-Mondoñedo 2006, 2007a, 2007b, 2019; Bentley 2013), semantic ones (e.g., Milsark 1974, 1977, 1979; Lumsden 1988; Blutner 1993; Enç 1991; Keenan 2003; Fischer 2016), or pragmatic-discursive ones (e.g., Abbott 1993, 2006; Ward and Birner 1995; Zucchi 1995; Pons Rodríguez 2014). It is widely accepted, particularly in relation to Romance languages, that the theoretical complexity of the definiteness effect boils down to a conspiracy between syntactic, semantic, and discursive properties; Zucchi (1995, 2003), McNally (1998, 2016), or Leonetti (2008), for instance, acknowledge that the definiteness effect seems to be at the interplay between syntax, semantics, and some discursive-pragmatic factors. Particularly, McNally (1997) hypothesizes that the definiteness effect has a Janus-like nature; quantificational noun phrases are excluded from existentials due to semantic principles, whereas other definite descriptions, such as proper names or pronouns, are barred as a consequence of independent pragmatic principles. There does not seem to remain an intuitive force to the idea that the definiteness effect can be made dependent on solely syntactic, semantic, or pragmatic principles. Instead, a conspiracy between several related principles seems to be at play. If this line of reasoning is taken to be on the right track, I believe we have arrived at a comprehensive explanation for the definiteness effect.

The definiteness effect has been devoted close theoretical and descriptive attention. Our understanding of the phenomenon, its crosslinguistic distribution, and its theoretical relevance has deepened accordingly. Nevertheless, there seem to remain some puzzling facts that I will only touch upon here. The question as to whether the definiteness effect is a restriction on subjects or a restriction on objects should be raised. Particularly, the syntactic nature of the pivot of existential haber, ‘to be’, in Spanish has been a matter of a long-standing controversy since early grammarians. At the very least, two analyses must be distinguished:

• The pivot occupies the object position. The hypothesis that the pivot of haber, ‘to be’, occupies the direct object position has been argued for by traditional grammarians (e.g., Kany 1945; the Esbozo written by RAE 1973; Gili Gaya 1961, §62; Fernández Ramírez 1987, §146; Alcina and Blecua 1975, p. 891). More recently, Fernández Soriano and Baylín (1999), and Longa et al. (1996, 1998) have also supported this view.
• The pivot occupies the subject position. The alternative hypothesis has also been put forth that the pivot of existential haber, ‘to be’, occupies the subject position of an intransitive construction. It comes as no surprise that this approach has revolved around plural agreement between the pivot and the verb (e.g., Montes Giraldo 1982; Fontanella de Weinberg 1992; Nicita 1997; Martínez 1999; Castillo Lluch and Octavio de Toledo y Huerta 2016; Claes 2014, 2016; Arteaga and de la Mora 2022).

An exhaustive analysis of each of the hypotheses would take us rather far afield. Nevertheless, some brief observations are in order. I will assume here that the pivot of Spanish haber is not a subject, but an object. First, number agreement must not be taken as evidence that the pivot occupies the subject position; person and number agreement relate to independent features (even to independent projections), and the latter is not necessarily triggered by subjects, as Rigau (1991, 1993, 1997, 2008) has convincingly argued. A similar line of reasoning is pursued in the works by Rodríguez-Mondoñedo (2006, 2007a, 2019), where it is argued—I believe, quite convincingly—that the pivot of existential haber, ‘to be’, is not endowed with (person) features. See Section 4 for further details. Second, number agreement in haber constructions, which has been tackled quite frequently (e.g., Claes 2014, 2016; Pato 2016 and references cited therein), is compatible with the pronominalization of the pivot (see Gràcia i Solé and Roca Urgell 2017; De Benito Moreno 2016; Rodríguez-Mondoñedo 2019; Agulló, forthcoming for further details). These facts would come as no surprise if the pivot was thought to receive Partitive case, along Belletti’s (1987, 1988) lines. I take these two pieces of evidence in support of the hypothesis that the pivot is an object, albeit it might be the case that it does not receive Accusative case. Therefore, the definiteness effect is taken here to apply to objects in Spanish.

Interestingly, the two analyses in (a) and (b) have their theoretical counterparts. La Fauci and Loporcaro (1997) proposed that the pivot is a predicate (see, among many others, Francez 2009, 2010; Cruschina 2012; Bentley 2013; Fábregas 2022). The analysis of the pivot as a constituent of a small clause was put forth by Stowell (1978) and followed by Chomsky (1981), Safir (1982, 1983), and Freeze (1992), among many others. Villalba (2013) has convincingly called this approach into question. Milsark (1977, 1979) or Lumsden (1988), on the contrary, conceive the pivot as an object.

The crosslinguistic distribution of the definiteness effect has been tackled and partially unveiled (see, for instance, McNally 2016). In fact, there are thought to be deep crosslinguistic differences as regards the effect. Italian or Catalan have been argued to be immune to the effect (e.g., Longa et al. 1996, 1998; La Fauci and Loporcaro 1997; Bartra 1987; Villalba 2013) (cfr., nonetheless, Leonetti 2008; Bentley 2013), whereas Spanish or Galician adhere more consistently to the effect (see Rodríguez-Mondoñedo 2007a, 2007b for Spanish). The hypothesis that the definiteness effect can be parametrized has been contended by several authors: crosslinguistic differences can be attributable to a difference between locative-flavored constructions and proper existentials (e.g., Cruschina 2016). It has come to be known that languages generally allow for definite pivots under certain circumstances (cfr. Bentley 2013 and the data in (3)), but what are subject to crosslinguistic variation are (i) the conditions under which definite pivots are acceptable and (ii) the positions of Aissen’s (2003) Definiteness Hierarchy (see (2)) that might be grammatical as pivots of existential sentences under those conditions.

1.4. Preliminary Hypotheses

Even though the crosslinguistic distribution of the definiteness effect has been delved into, linguistic contact settings remain largely unexplored. The question has not been raised as to whether linguistic contact strengthens or alleviates the definiteness effect, and if so, what mechanisms underlie syntactic variation. Linguistic interaction between Catalan and Spanish has been paid close attention at the phonetic, morphological, and lexical levels; nevertheless, it comes as no surprise that the syntax of contact varieties has been barely touched upon. Gràcia i Solé and Roca Urgell (2017), for instance, reported that Catalan behaves more loosely as regards the effect, but no real examples are offered, and no quantitative result is shown, either. This paper aims to fill this void.

The differences between Spanish and Catalan, at least prima facie, are striking. Definite pivots may be acceptable in both languages, but there seem to be some differences as regards the conditions that make definite pivots grammatical; while definite pivots in Spanish need to trigger list, negative, weak definite, or superlative readings to be grammatical, Catalan allows for definite pivots without those additional readings. The grammaticality of proper names and personal pronouns makes Catalan existentials differ greatly from their Spanish counterparts. A general research hypothesis, hence, reveals itself, as formulated in (7):

(7)	Varieties of Spanish in contact with Catalan amnesty the definiteness effect.

Some new data will be brought to the fore that support the validity of the claim in (7). It will then naturally follow that definite pivots are more frequent in varieties of Spanish in contact with Catalan. Additionally, list, negative, superlative, or weak definite readings are not needed to make definite pivots grammatical in these varieties. The hypothesis (7) will be sharpened and argued to be derived from more general principles of grammar.

2. Materials and Methods

This study is based upon the reflection and observation of a rather large database (N = 4995) of haber, ‘to be’, existential constructions in European Spanish gathered from the Audible Corpus of Spoken Rural Spanish (henceforth COSER, after its translation in Spanish) (Fernández-Ordóñez 2005, dir.), a lemmatized corpus of rural varieties of European Spanish. The quantitative aim of this study is to test the hypothesis in (7), from which stems a prediction: definite pivots should be more frequent in varieties of Spanish in contact with Catalan. Quantification is thus needed to model grammatical variation. This methodology aims to test the hypothesis that definite pivots are more readily used in varieties of Spanish that are in contact with Catalan. Accordingly, existential constructions with definite and indefinite pivots have been collected. Data used in this paper were retrieved automatically from transcribed interviews, which are available on the web. After a close examination of the large database, several types of constructions were excluded, namely due to two reasons: (a) possessive (and thus transitive) uses of haber, ‘to have’, and (b) hesitations of the speaker or interruptions, which seemingly cast some doubt on the linguistic construction being used. Each existential construction was classified vis à vis the definiteness of the pivot. If the pivot is definite, there arise further possibilities; the mechanism or interpretation that definiteness triggers was tagged as ‘List reading’, ‘Locative’, ‘Relational possessive determiner’, etc. The label ‘DE’, which stands for ‘definiteness effect’, is used when definiteness does not trigger any of the aforementioned readings. Frequencies, which were obtained using R language programming (R version 4. 3. 1 Beagle Scouts) (R Core Team 2021), are depicted with QGIS (QGIS Association 2023), an open-source geographical information system.

The choice of COSER as the fundamental corpus for this work is by no means a hazardous or trivial decision. The COSER project has become a cornerstone in the study of vernacular varieties of Spanish since its inception for several reasons: (a) data from rural varieties are collected with semidirected sociolinguistic interviews, which has resulted in a significant number of transcribed interviews (i.e., it amounts to 3,596,205 words); (b) it is the only corpus of current Spanish that allows for the study of spatially continuous varieties. Thus, the COSER can be safely considered the most substantial corpus of vernacular Spanish to date. Illustrative in this respect is the vast amount of research that hinges on the COSER data and the doctoral dissertations based on them. I refer the reader to the works by Fernández-Ordóñez (2004, 2009) and Fernández-Ordóñez and Pato (2020) for further details.

Some words of caution are, nevertheless, in order. The point should be made clear, as suggested by a reviewer, that the COSER data are obtained from older speakers of rural varieties of Spanish and are, hence, dependent upon the criteria used for the selection of the informant. As a consequence, statistical results are equally dependent on these criteria and could possibly be different if another age or social group was selected.

3. Results

Definiteness in the pivot position of the existential clause is by no means restricted to a specific variety of Spanish. Evidence in support of this claim will be argued to come from a more or less general approach, developed in (Section 3.1): if we take definiteness to be a purely morphosyntactic notion, we can take samples as those in (3) to be properly definite. I will first provide some insight into the dialectal distribution of definite vis à vis indefinite pivots from a merely morphosyntactic viewpoint. Then, the hypothesis will be put forth in (Section 3.2) that the study of the syntactic variation at hand can only be dealt with if morphologically definite, but semantically indefinite constructions, as those in (2), are excluded from the corpus. The quantitative results stemming from excluding constructions, such as those in (3) from the corpus, are shown in (Section 3.3).

3.1. Definiteness and the Existential Sentence: Preliminary Insights

Definite pivots, as argued in relation to (3), are grammatical if certain conditions are met. Let us assume, only tentatively, that definiteness can be regarded as a purely morphological notion; if we abstract away from notions such as uniqueness, presuppositionality, or weak definiteness, a conventional distinction can be drawn between DPs headed by an indefinite determiner, as in (8a), and DPs headed by definite articles, as in (8b).

(8)	a.	Había	un	pastor	que	cuid-aba	a todos.
		there.was	a	pastor	rel	keep-pst.3sg	dom all
		‘There was a pastor that took care of them all.’
		(Elburgo/Burgelu (Alava), COSER-0103_01)
	b.	Pues	aquí	había	el	ayuntamiento	antiguo.
		Then	here	there.was	art	town hall	old
		‘Then there was the old town hall.’
		(Mura (Barcelona), COSER-0804_01)

The difference between definite pivots and indefinite ones, stated only formally, seem clear-cut (see Prince 1992 for an early objection). Namely as a matter of execution, I will assume that examples of the sort of (8a) can be regarded as indefinite, whereas (8b) is properly definite. If this distinction between definite and non-definite pivots is taken to be taxonomical and more or less stable, it is tempting to shed some light into its dialectal distribution. The large database of existential constructions (N = 4995), which can be safely considered as representative of the different varieties of European Spanish, allows for robust generalizations. In Figure 1, the frequency of indefinite (and non-definite) pivots versus definite ones is depicted.

A preliminary insight into the frequency data yields two conclusions: (a) indefinite (and non-definite) pivots show a pervasive frequency; and (b) definiteness in the pivot position of the existential clause is by no means restricted to a specific variety of Spanish. The former fact clearly stems from a more general principle: existential sentences have been argued to introduce new referents into the discourse or to re-state them (cfr., among many others, Prince 1997; Abbott 1993, 1999; Ward and Birner 1995; Francez 2009; Cruschina 2012). Indefinites, as seem natural, are the device par excellence for doing so. Negative indefinites, by definition, pose a problem—pinpointed by McNally (1997)—that I will not delve into.

Definite pivots, on the contrary, are attested through different dialects, even though there are clear-cut quantitative differences; indefinite pivots are pervasive, contrary to definite ones, which show a more restricted distribution. Indefinite pivots, as an anonymous reviewer points out, are attested irrespective of the dialectal area. Nevertheless, definite pivots are not attested in all varieties of European Spanish; southern regions of the central area of the Peninsula, as well as some varieties of north-eastern Castile, seemingly avoid the use of the definite pivot. More generally, the fact that definite pivots are possible in Spanish should come as no surprise. In fact, definite pivots have been also attested in the history of Spanish (e.g., Pons Rodríguez 2014), and have been argued to be grammatical under certain conditions (see, among others, Leonetti 2008 for Spanish). Even in a preliminary observation of the data in Figure 1, there seem to be some dialectal differences with regard to the frequency of definite pivots; some villages from the territories where Catalan is spoken reach up to 100% of definite pivots, as seen in Sant Climent (Mallorca), or to 62.5%, as seen in Torregrossa (Lleida).

Care should be exercised, nevertheless, as regards the distinction in (8). The distinction between definite pivots and indefinite ones only in morphological terms poses some theoretical problems. There is not thought to be a straightforward relation between definiteness and whatever is taken to be its semantic import, whether it be presuppositionality, familiarity, or the like (see Abbott 1992, 1993, 1999, 2006). The opposite, naturally, also holds true; indefinites are not uniformly associated with hearer-new or brand-new entities (i.e., Ward and Birner 1995 false indefinites), as is known since Prince’s (1992) work on indefinites. Strong evidence in this respect has been brought to the fore (e.g., Fodor and Sag 1982; Heim 1982; Ward and Prince 1991).

That being the case, it stands to reason that the distinction between definite and indefinite pivots in (8) must be refined. Purely morphological distinctions between definite and indefinite pivots seem inadequate. The question arises as to whether these formally definite pivots should be considered semantically definite or indefinite. The issue, I believe, is nontrivial. If some formally definite pivots are argued to convey indefinite meanings, they can naturally circumvent the definiteness effect. Hence, they must not be labeled as definites and have to be excluded from the statistics in Figure 1. I will briefly sketch in the next section different kinds of formally definite pivots that, nevertheless, seem to adhere to indefinites (or even bare nouns). As a consequence, they must be regarded as formally definite indefinites. I will touch upon some theoretical problems that arise, but an explicit theoretical proposal will not be offered. The reader is referred to Rando and Napoli (1978), Heim (1982, 2019), Ward and Birner (1995), Abbott (1993), McNally (1997), and the references cited therein.

3.2. What Amnesties the Definiteness Effect in Spanish?

Formally definite Ds may be grammatical in the pivot position of the existential sentence when certain circumstances are met. Several types are usually distinguished, as follows: list reading interpretations and locative coda preposing.

List reading interpretations have been claimed to circumvent the definiteness effect since at least Rando and Napoli (1978) and Ziv (1982). These constructions have been referred to as contextualized existentials (e.g., Abbott 1992, 1993) or enumerative existentials (e.g., Abbott 1999). List reading interpretations are held to be triggered when the noun in question is part of a wider set, which can be either covertly or overtly expressed. As a consequence, the relation between the noun and the wider set it is part of can be traced covertly, i.e., retrieved from previous discourse or simply understood. Some examples of pivots triggering list interpretations are given below:

(9)	a.	Y	después	hay	la	mosca	también.
		And	then	there.is	art	fly	too
		‘And then there’s the fly too.’
		(Santa Eulalia de Oscos (Santa Eulalia de Oscos) (Asturias), COSER-0543_01)
	b.	Después	había	la	guadaña,	después	había
		Then	there.was	art	scythe,	then	there was
		la	máquina	de	segar.
		art	machine	of	mowing.
		‘Then there was the scythe, then there was the mower.’
		(La Serra (La Torre de Claramunt) (Barcelona), COSER-0803_01)
	c.	Ten-dría	treinta	años	cuando	empezar-on	a
		have-cond.1sg	thirty	years	when	start-pst.3pl	to
		haber	los	televisores	y	las	radios.
		There-be.inf	art	televisions	and	art	radios
		‘I would be thirty years when there started to be the televisions and the radios.’
		(Alanis (Sevilla), COSER-3809_01)
	d.	Había	el	casino	y	el	círculo.
		There was	art	casino	and	art	circle
		‘There was the casino and the circle.’
		(Antequera (Málaga), COSER-3001_01)

It should be underscored here that the pivot need not appear in an explicit coordinate structure, as in (9b–d). Adverbs such as también (‘also, too’), as in (8a), can also trigger list interpretation. An observation regarding the grammatical variation at stake is in order. List readings, as seems obvious, are not necessarily linked to definiteness. Samples such as those in (9) allow for an alternation between the definite pivot and the bare noun phrase. Example (9c), repeated here as (10) without irrelevant details, is enlightening in this regard:

(10)	Tendría treinta años cuando empezaron a haber {Ø ~ los ~ ??unos} televisores y {Ø ~ las ~ ??unas} radios.

The oddity of the indefinite, marked with (^??), determiner holds only under the relevant interpretation. Note that, in (9c), an existential predication is built around types or sorts (i.e., what McNally 1997 labels nominalized functions). Bare nouns, which have been taken to denote either classes or types (e.g., Carlson 1977, p. 59; Chierchia 1997, 1998; Doron 2003) or properties, under McNally’s (2004) proposal, are fully grammatical in the pivot position of the existential. The indefinite determiner, in these cases, seems to trigger a referential interpretation, as defined in Fodor and Sag (1982); a member of a set is picked out or selected among others. The alternation between definite and indefinite determiners in (9), hence, cannot be regarded as free.

It is widely agreed upon that pivots of the sort illustrated in (9), which are formally definite, are nonetheless interpreted as members of an indefinite set (see Abbott 1992 for an alternative proposal). Ward and Birner (1995) and Hartmann (2006) have further developed some of Rando and Napoli’s (1978) ideas. If list reading interpretations are taken to be dependent on more general principles of grammar, I believe it might be promising to derive these interpretations from the concept of D(iscourse) linking or partitivity, as borrowed from Enç (1991), Ionin (2003), and Rizzi (2005). D-linked constituents refer to members of a previously introduced set. Partitivity has been devoted close attention in relation to doubling object constructions; indefinites in clitic left dislocations and in clitic doubling constructions, for instance, are acceptable if a partitive or D-linked interpretation is triggered (e.g., Suñer 1988; Fernández Soriano 1993; Alexiadou and Anagnostopoulou 2000; Villalba 2000; Belloro 2007, 2012; Agulló 2023a, 2023b). Essential in this respect is that this member-of-set relation need not be explicit; partitivity may be overt or covert, as argued in Enç (1991).

Locative coda preposing also allows for definite pivots, as illustrated in (11):

(11)	a.	Allí	hay	el	Parque	Móvil		de
		there	there-is	art	park	mobile		of
		la	Guardia	Civil.
		art	guard	civil
		‘There was the vehicle fleet of the Civil Guard.’
		(Povedilla (Albacete), COSER-0222_01)
	c.	En	toas	parte-s	había		las	fiestas
		in	all	part-s	there.were		art	parties
		de-l	año.
		of-art	year
		‘Everywhere, there were the village festivals.’
		(Gordexola (Vizcaya), COSER-4506_01)
	d.	Allí	abajo	hay	el	molino.
		there	below	there.is	art	mill
		‘Below, there is the mill.’
		(Ledantes (Vega de Liébana) (Cantabria) COSER-1212_01)

Locative coda preposing is known to amnesty the definiteness effect, as was first noticed for Spanish by Leonetti (2008). Rigau (1997) previously noted that a similar fact holds for Catalan (see also Villalba 2013 for eventive existentials and Remberger 2013 for deontic existentials). Cruschina (2015) (also Cruschina 2012) regards the preposed coda, as the ones exemplified in (11), as a sort of aboutness topic, as I will replicate here. To put it simply, aboutness topics are constituents external to the core predication that assert what the sentence is about. Syntactically, locative codas occupy peripheral or detached positions. If Leonetti’s (2008) Coda Constraint is taken to be on the right track, locative codas must be dislocated, whether it be left- or right- dislocated, to allow for definite pivots. Thereupon, in (11a), the pivot el Parque Móvil, ‘the vehicle fleet’, which, under Cruschina’s (2015), acts as a predicate, is modified by the preposed locative coda Allí, or ‘there’; additionally, following Leonetti (2008), this locative coda preposing also legitimates the definite pivot, which would be otherwise ungrammatical. Even though existentials with definite pivots have been argued to have a ‘strong locative flavor’ (e.g., Zucchi 1995; Zamparelli 2000; Cruschina 2012), I will consider them a subtype of existential sentences. The examples in (11), consequently, can be safely treated as a subtype of existential sentences, regardless of their locative flavor.

If the previous lines of reasoning are pursued, it seems obvious that formally definite pivots do not correspond in a straightforward manner to definite interpretations. Particularly, to the extent that list readings and locative coda preposing are thought to amnesty the definiteness effect, they might be excluded temporarily from the corpus to test the hypothesis that varieties of Spanish in contact with Catalan are not sensible to the definiteness effect. As seems obvious from Figure 1, list readings and locative coda preposing are not restricted to certain dialectal varieties.

3.3. The Immunity to the Definiteness Effect: Some Findings

The role of list readings, as seen in (9), and locative coda preposing, as seen in (11), in amnestying the definiteness seems obvious. The study of frequency data without further modifications led us to the map in Figure 1, which shows certain more or less subtle dialectal differences. Indeed, those data provided some insight into the restrictive distribution of definite pivots when compared to indefinite ones, as well as some patterns of geographical variation. I will now take things a step further. If the question is raised as to whether varieties of Spanish in contact with Catalan, list readings, and locative flavored constructions need to be tentatively isolated from the corpus, mechanisms that circumvent the definiteness effect will be removed, and we will be left with those samples that truly show immunity to the definiteness effect.

The result of removing locative coda preposing constructions and list reading constructions from the corpus is shown in Figure 2.

The frequency data in Figure 2 seem to sharpen some patterns of linguistic variation. When locative flavored and list reading constructions are sorted out, some dialectal differences become clearer. First, it should be emphasized that indefinite pivots are equally pervasive, no matter the dialectal variety of European Spanish. The Eastern area of the Iberian Peninsula still resorts to definite pivots in existential constructions more readily than the central, western, and southern areas. Moreover, previously noted dialectal differences become even stronger. A striking difference between the data in Figure 1 and the data in Figure 2 is that varieties of Spanish in contact with Catalan stand in stark contrast with the rest of the varieties in their uses of definite pivots. As can be clearly seen, sorting out locative and list existential constructions reduces the usage of definite pivots in the central, western, and southern areas of peninsular Spanish to zero (or near zero). Additionally, these data show the quantitative strength of locative and list interpretations in linguistic variation.

If we still want to pursue our analysis of linguistic variation, there appears to be more to the story. As the quantitative data were explored in both Figure 1 and Figure 2, I will now abstract away from frequency to offer some insight into how syntactic variants are distributed.

3.4. Definite, Specific, and Personal Pivots in Contact with Catalan

Upon closer inspection, the existential constructions gathered from the COSER data will be argued to provide useful theoretical insight. In this section, some previously unnoticed data will be brought to the fore. Varieties in contact with Catalan not only use definite pivots more frequently than other varieties, as Figure 1 and Figure 2 have shown, but also allow for specific and even personal referents in the pivot position. Figure 3 depicts the samples of constructions that show immunity to the definiteness effect without list or locative interpretations.

It seems clear, even under a preliminary approach, that immunity to the definiteness effect is found mainly in varieties of Spanish of the coastal or mediterranean area of the Iberian Peninsula. Samples of immunity to the definiteness effect are far more frequent in varieties of Spanish that are in contact with Catalan, as was argued in relation to Figure 1. Some other cases are found, but they are isolated and do not conform to a dialectal area. On the contrary, those cases of immunity to the definiteness effect found in the eastern part of the Iberian Peninsula do conform to a more or less unitary area; from the north-eastern region to the south of the Community of Valencia, there seems to be a strong dialectal pattern. Immunity to the definiteness effect can thus be geographically traced.

Samples of immunity to the definiteness effect must be grouped, I believe, with regard to the animacy of the pivot. I will thus tease apart nonhuman or inanimate pivots and personal ones. The reasons underlying this decision will be developed immediately below. The examples in (12) illustrate definite inanimate pivots:

(12)	a.	Cuando	vine	de	la	mili
		when	came.1sg	of	art	military service
		había	la	carretera.
		there-was	art	road
		‘When I returned from the military service, there was the road.’
		(La Serra (Torre de Claramunt) (Barcelona), COSER-0803_01)
	b.	Ahora	hay	la	calefacción.
		now	there.is	art	heating
		‘Now, there is the heating.’
		(Mura (Barcelona), COSER-0804_01)
	c.	Había	los	tapones	de	gasoil	en
		there.were	art.pl	caps	of	diesel	on
		el	suelo.
		art	floor
		‘There were the caps of diesel on the floor.’
	d.	Hay	los	balcones	abiertos,	yo.
		there.are	art	balconies	opened,	me.
		‘There are the balconies opened, me’.
		(La Serra (Torre de Claramunt) (Barcelona), COSER-0803_01)
	e.	Había	la	iglesia,	la	iglesia	abierta.
		there.was	art	church,	art	church	opened.
		‘There was the church opened.’
		(Torrebesses (Lleida), COSER-2711_01)
	f.	Después	había	la	República.
		after	there.was	art	republic.
		‘Then, there was the Republic.’
		(Son Macià (Manacor) (Mallorca), COSER-4915_01)
	g.	Había	el	colegio	por	la	mañana
		there.was	art	school	in	art	morning
		y	por	la	tarde.
		and	in	art	afternoon.
		‘There was the school in the morning and in the afternoon.’
		(San Climent (Mao) (Menorca), COSER-5003_01)
	h.	Había	la	cocina	encendida.
		there.was	art	kitchen	on.
		‘There was the kitchen turned on.’
		(San Climent (Mao) (Menorca), COSER-5003_01)

Definite inanimate pivots are, in fact, somehow exceptional. No additional mechanism (i.e., list or locative reading) is found in these examples; nevertheless, these definite pivots are attested. A distinction must be made, I believe, between what seem to be (a) weak uses of the definite (in the sense of Poesio 1994) and (b) proper definites. Pivots of the examples in (11a, b, g) may be thought of as weak definites to the extent that they show the following properties: (i) they are not related to uniqueness (Löbner 1985; Ward and Birner 1995; Carlson et al. 2006; Schwarz 2014); (ii) they disallow, quite generally, anaphoric reference (e.g., Scholten and Guevara 2010), as do negative indefinites (Karttunen 1968, 1969); and (iii) their distribution overlaps with that of bare noun phrases (e.g., Carlson and Sussman 2005). In fact, the pivot la carretera (‘the road’) in (12a) or la calefacción (‘the heating’) in (12b) do not refer uniquely; in that sense, they are not referential, nor are they definite or specific. No specific referent is being invoked or restated. Instead, what (12a, b, h) seem to denote is a type or kind, that is, what McNally (1997) refers to as nominalized functions. Evidence in support of this hypothesis is that the definite article in (12a, b, h) alternates with the corresponding bare noun phrase, but the indefinite is ungrammatical only under the relevant interpretation: Ahora hay {Ø ~ ^??la ~ *una} calefacción (‘Now, there is (the) heating’). It is interesting to note that, under Schwarz’s (2014) or Aguilar-Guevara and Zwarts’ (2010) analyses, weak definites and bare noun phrases are hypothesized to denote types.

Leaving weak definites aside, it stands to reason that the pivots in (12c, d, e, f, h) are used as proper definites. Differently put, they are strong determiners in Milsark’s (1974) terms. For instance, la República (‘the Republic’) in (12f) or la cocina (‘the kitchen’) in (12h) are referential stricto sensu and even specific, inasmuch as they refer univocally. These data stand in stark contrast with Spanish, which, quite generally, only allows for nonreferential pivots (i.e., McNally’s 1997 nominalized functions) and bars referential nouns from the pivot position of existential haber, ‘to be’.

Pivots with personal or animate features are also found among the samples that show immunity to the definiteness effect. Some examples are found in (13) below:

(13)	a.	Habían	los	Condes	de	Barcelona.
		there.were-3pl	art.pl	counts	of	Barcelona
		‘There were the Counts of Barcelona.’
		(Sant Antoni de Vilamajor (Barcelona), COSER-0805_01)
	b.	Menos	mal	que	había	el	vecino.
		less	bad	comp	there.is	art	neighbor.
		‘Luckily, there was the neighbor.’
		(Torrebesses (Lleida), COSER-2711_01)
	c.	No	había	los	señores.
		neg	there.is	art	men.
		‘There weren’t the old men.’
		(Torregrossa (Lleida), COSER-2712_01)
	d.	Había	el	médico	de	cabecera.
		there.was	art	doctor	of	primary attention
		‘There was the general practitioner.’
		(San Climent (Mao) (Menorca), COSER-5003_01)
	e.	En	esta	ocasión,	había	el	cura.
		in	this	occasion	there.was	art	priest
		‘Then, there was the priest.’
		(Guimerà (Lleida), COSER-2715_01)
	f.	Te	obligaban	a	hablar	en	castellano,
		you.acc	forced	to	speak	in	Castilian
		cuando	había	Franco.
		when	there.was	Franco
		‘You were forced to speak in Castilian when there was Franco.’
		(La Serra (La Torre de Claramunt) (Barcelona), COSER-0803_01)

The examples in (13) are a novelty in themselves, as far as they show definite, specific nouns in the pivot position of existential haber, ‘to be’. Some observations are in order with regard to the distinction, just sketched, between nonhuman or inanimate pivots (12) and personal pivots (13). Quite generally, pivots in Spanish have been argued to ban (person)) features (Rodríguez-Mondoñedo 2006, 2007a, 2007b, 2019). I believe the samples in (13), particularly, can further our knowledge on existentials in contact situations between Spanish and Catalan and also sharpen some ideas about φ (‘phi’) features. A brief explanation will be developed in (§4). Let us note, for the moment, that all of the pivots in (13) seemingly have (person) features. The pivots los Condes de Barcelona (‘the Counts of Barcelona’) in (13a), el cura (‘the priest’) in (13b), los señores (‘the old men’), etc., denote human referents. Furthermore, all of these pivots are both definite and specific.

3.5. Two Types of Possessives in the Pivot Position: Relational/Emphatic and Strong

In this section, some brief observations will be developed with regard to possessive determiners. Two types of possessive determiners will be argued to be attested in the pivot position of haber (‘to be’) in Spanish: strong possessives and emphatic possessives. It seems, at least prima facie, that existential haber (‘to be’) in Spanish prevents possessive determiners from occupying the pivot position (see Eguren 2018, 2023). The ungrammaticality of the examples in (14) supports this observation:

(14)	a.	*Había	{mi ~ tu ~ su}	libro	en	la	habitación.
		there.was	{my ~ your ~ his}	book	in	art	room
		‘There was {my ~ your ~ his} book in the room.’
	b.	*Había	{mi ~ tu ~ su}	tío	en	la	habitación.
		there.was	{my ~ your ~ his}	uncle	in	art	room
		‘There was {my ~ your ~ his} uncle in the room.’

Neither non-animate (14a) nor human (14b) referents seem to be grammatical in the pivot position of existential haber (‘to be’).

The hypothesis will be put forth here that existential sentences in Spanish allow for two different types of possessive determiners in the pivot position. Sharpening the issues here, there appear to be some cases which show immunity to the definiteness effect, while others do not. I will support the claim that pivots in Spanish existentials must be classified as to which type of possessive occupies the D° head position: (a) semantically strong possessives (in the sense of Milsark 1974, 1977) and (b) what is labeled in Eguren (2018, 2023) and Del Barrio de la Rosa (2022), emphatic possessives.

(15)		Strong possessive determiners
	a.	Hasta	cuando	había	mi	hermano	vivo.
		even	when	there.was	my	brother	alive
		‘Even when there was my brother alive.’
		(Torregrossa (Lleida), COSER-2712_01)
	b.	Si	no	hubiera	habido	mi	madre.
		if	neg	aux	were	my	mother.
		‘If there hadn´t been my mother there.’
		(Torregrossa (Lleida), COSER-2712_01)
	c.	Aquí	había	la	abuela	mía.
		here	there.was	art	grandmother	poss.1
		‘Here, there was my grandmother.’
		(Gordexola (Bizkaia), COSER-4506_01)
	d.	Había	tu,	tu	hermano,	tu	madre
		there.was	your	your	brother	your	mother
		o	tu	mujer.
		or	your	wife
		‘There was your brother, your mother or your wife.’
		(Fuente del Pino (Jumilla) (Murcia), COSER-3107_01)
		Emphatic possessive determiners
	e.	En	todos	los	pueblos	hay	sus
		in	every	art	villages	there.is	poss.3-pl
		médicos.
		doctors
		‘In every village there are the doctors.’
		(Aguilar de la Frontera (Córdoba), COSER-1503_01)
	f.	Hay	sus	neveras.
		there.is	poss.3-pl	refrigerators
		‘There are the refrigerators.’
		(Arjona (Jaén), COSER-2301_01)
	g.	En	cada	sitio	hay	sus	costumbres.
		in	each	place	there.is	poss.3-pl	habits
		‘In every place there are the customs.’
		(Lorenzana (Cuadros) (León), COSER-2614_01)
	h.	Había	su	carretero	en	el	pueblo.
		there.is	poss.3-sg	impolite	in	the	village
		‘There was the impolite in the village.’
		(Muñoveros (Segovia), COSER-3707_01)
	i.	Había	su	sociedá	y	ahí	había
		there.was	poss.3-sg	society	and	there	there.was
		sus	socios.
		poss.3-pl	members
		‘There was the society, and there were its members.’
		(Constantina (Sevilla), COSER-3814_01)

There seems to be a neat difference between the possessive determiner in (15a–d) and the one in (15e–i). The pivots mi hermano (‘my brother’) in (15a), mi madre (‘my mother’) in (15b), la abuela mía (‘my grandmother’) in (15c), or the coordinate conjunct tu hermano, tu madre o tu mujer (‘your brother, your mother, or your wife’) in (15d) differ from those in (15e–i) in both their definiteness and their specificity. The pivots in (15e–i) are properly definite and specific, to the extent that they denote univocally. Furthermore, all of these examples pertain to the same dialectal area: that of Spanish in contact with Catalan. Example (15d), albeit from Murcia—to the South of the Catalan-influenced region—is no exception to the generalization; the second person possessive determiner seems to trigger a generic interpretation, and hence, no specific interpretation is triggered. The samples in (15a–d), apart from (15d), can thus be safely regarded as cases of immunity to the definiteness effect.

Several differences arise between possessive determiners in (15a–d) and those in (15e–i). The former are referential stricto sensu, but the latter seem to stablish a different type of relation. Note that the pivot sus medicos (‘their doctors’) in (15e) or sus neveras (‘their refrigerators’) in (15f) do not denote a possessive relation, nor do sus costumbres (‘their customs’) in (15g) or su carretero (‘their impolite (man)’) in (15h). In addition, these emphatic possessives occupy a preposed position, and there is no alternation with postnominal possessives; this is contrary to what happens in (15a–d), where strong possessives alternate with postposed, as in example (15d). As Eguren (2023) puts it: “they do not properly express possession” (112). Instead, what seems to be at work is an emphatic, non-contrastive interpretation of a previously stablished possession relationship. Eguren (2018, 2023), quite convincingly, I believe, hypothesizes that the difference between canonical possessives, such as those in (15a–d), and emphatic possessives, such as those in (15e–i), lies in that emphatic ones are generated in situ, i.e., they are not transformationally related to an argument position, as happens with prenominal possessives (see also Eguren 2016). The pivots in (15e–i) are thus non-definite and compatible with non-specific interpretations. Emphatic possessive determiners, such as those in (15a–d), are hence not immune to the definiteness effect and constitute no exception to the generalization that the definiteness effect holds quite robustly in Spanish.

4. Discussion

In what follows, I will argue that the differences between Spanish and Catalan stem from a more general principle: dialects of Spanish in contact with Catalan allow for definite, specific DPs in the pivot position. As such, I will take them to be full-fledged DPs whose head D hosts person, number, and gender features. More specifically, in (Section 4.1, Section 4.2, Section 4.3, Section 4.4 and Section 4.5), I shift gears to put forth the hypothesis that existential pivots in so-called general Spanish are assigned Partitive case, as I state in (Section 4.1). Evidence in support of this hypothesis comes from several related facts: Romance languages with dedicated partitive pronouns pronominalize the pivot position of the existential with partitives (Section 4.2); clitics out of existentials in Spanish are person-defective and thus phi incomplete, even when haber (‘to be’) bears person agreement morphemes (Section 4.3); if person features are thought to be located in the head D, these clitics cannot be related to a full-fledged DP projection (Section 4.4). The fact that the hypothesis of Partitive case assignment to the pivot is not ad hoc is proven by the fact that Spanish, as I argue in (Section 4.5), has partitive pronouns with unaccusative verbs. In (Section 4.6), syntactic variation in varieties of Spanish in contact with Catalan is argued to boil down to a structural condition on Partitive case assignment. Since Belletti’s (1987, 1988) account, Partitive case assignment is dependent on the non-definiteness of the pivot. I specifically argue that, as a consequence of contact between Spanish and Catalan, Partitive case is no longer tied to non-definiteness.

4.1. The Situation in General Spanish: Partitivity and Non-Referentiality

There seems to be wide agreement over the fact that, in Spanish, the pivot of the existential construction is not related to referential readings. The non-referentiality of the existential pivot has been accounted for in rather different ways, be it with regard to the semantics of bare noun phrases, to the grammatical category of the pivot, or in relation to case assignment. I believe a unified proposal, as the one I will develop here, to be on the right track: there is a conspiracy between semantics and case assignment that accounts for the relevant facts.

The semantics of bare noun phrases has been hypothesized to be related to either (a) classes (or types), which are sets of properties attributed to individuals, as in Carlson’s (1977, p. 59 and seq.), Chierchia’s (1998), Doron’s (2003), or Fábregas’ (2022) proposals, or (b) simply properties, as put forth by McNally (1998, 2004). I believe the distinction between the two to be non-trivial, but irrespective of its precise semantics, what seems clear is that the pivot does not convey a determined reference, as seen in Farkas’ works (2000, 2002a, 2002b); it denotes neither tokens nor individuals, i.e., it does not pick up specific individuals or exemplars. It is in this sense of non-determined reference that the term partitivity, alongside parti-genericity, will be used here (see, among others, Belletti 1987; Laca 1990, 1996; Demonte and Masullo 1999; Schurr 2020; Agulló 2023a).

If this hypothesis is taken to be essentially correct, we can easily break down the ungrammaticality of the samples in (16):

(16)	a.	*Había		Juan	en	la habitación
		there.was		Juan	in	the room
		‘There was Juan in the room.’
		(Rodríguez-Mondoñedo 2007b, p. 13)
	a.	*Has	tú.
		There.is-2.sg	you
		‘There was/were you.’
		(Rodríguez-Mondoñedo 2007b, p. 16)

Neither Juan in (16a) nor tú (‘you’) in (16b) can be said to convey properties or sets of properties, and thus, they are ungrammatical in the pivot position. Nor can they be used in a predicative sense. Note that the ungrammaticality of (16) is robust in general Spanish, but not so in varieties of Spanish in contact with Catalan. Proper names and personal pronouns, in the upmost positions of Aissen’s (2003) Definiteness Hierarchy, are straightforwardly excluded from the pivot position in semantic terms. Given that the pivot position can only be partitive or parti-generic, both proper names and personal pronouns are barred from the position: the pivot can only pick up classes or types, and not tokens or individuals, such as Juan (16a) or tú (16b).

4.2. Partitive Case Assignment and Partitive Pronoun Series

The complexity of the definiteness effect, however, is not restricted to (16). On the contrary, there appears to be more to the story. I will argue, particularly, that sequences such as (16) can be ruled out if Partitive case assignment is invoked. Upon closer inspection, the hypothesis will prove both tenable and promising.

Let us assume the semantics of indefinite and bare noun phrases to be closely related to case assignment. The semantic import of case assignment need not be emphasized, as there is a long-standing tradition that links case to the semantics of the noun. The reader is referred to Kagan (2020) and the references cited therein. The hypothesis I will test out here crucially hinges on Belletti’s (1987, 1988) and Lasnik’s (1992) accounts; according to them, the pivot of the existential clause behaves in a similar fashion as the subject of unaccusatives in that it bears Partitive case, i.e., it does not receive a Nominative nor Accusative case. Fábregas (2022) and Gràcia i Solé and Roca Urgell (2017) also assume that Accusative case is not at work in existential sentences, as I assume here. My proposal departs from them in that I explicitly assume Partitive case assignment to be at work in existentials. A distinction was drawn in Chomsky (1981) between an inherent case, associated with theta-marking, and structural case, which independent of theta-marking and is construction-dependent. The question shall be raised as to whether Partitive case is inherent or structural, but I believe that this issue would take us rather far afield. Let us only briefly touch upon it. Belletti (1987, 1988) takes Partitive case to be inherent (cfr. Eguzkitza and Kaiser 1999), but Lasnik (1992), Vainikka and Maling (1996), and Kiparsky (1998) have casted some doubts on the assumption. Partitive case in Finnish, for instance, has been related to both configurational and semantic properties (e.g., Lasnik 1992; Csirmaz 2012), but I believe there remains some doubts as to whether it is inherent or structural.

Evidence in support of the fact that the pivot of the existential bears Partitive case comes in various guises. I will only be concerned, mainly for explanatory purposes, with the three following claims:

(17)	a.	Romance languages that have a specifically partitive inventory of pronouns use it to pronominalize the pivot of the existential.
	b.	Western and Central Ibero-Romance recycle Accusative clitic pronouns (Rigau 1988; Longa et al. 1996, 1998) are partitives. As such, these pronouns are endowed with number features, but not person features.
	c.	These accusative-as-partitives pronouns are not ad hoc stipulations for existentials, as they can be found in some unaccusative constructions.

If the claims in (17) are on the right track, as I will argue, the hypothesis that the pivot position of the existential is assigned Partitive case would fare better than other alternative hypothesis.

The claims in (17a) and (17b) shall be reviewed jointly. Romance Languages can be said to belong to one of the two following classes: (a) languages with dedicated partitive pronoun series (e.g., Catalan, Aragonese, Italian, French, and Occitan) and (b) languages without dedicated partitive pronouns (e.g., Galician, Portuguese, Spanish, and Rumanian) (cfr., specially, Kabatek and Pusch 2011). Let us concentrate on Catalan’s partitive pronouns, as those italicized in (18), and Spanish’s accusative-as-partitive pronouns, as in (19):

(18)	a.	N’hi ha	que	diuen	que	els	redacta	ell.
		part-loc is	that	say	that	cl.acc-3pl	writes	he
		‘There are those that say that he writes them himself.’ (CTILC 1967)
	b.	N’hi ha	que	pretenen	que	és	viu.
		part-loc is	that	pretend	that	is	alive
		‘There are those that pretend he is alive.’ (CTILC 1969)
(19)	a.	Los	hay	que	nacen	con	estrella.
		cl.acc-3pl	is	that	born	with	star.
		‘There are those that are born with a star.’ (CORPES XXI, Spain)
	b.	Los	hay	que	son	guardias	de	tráfico.
		cl.acc-3pl	is	that	are	guard	of	traffic.
		‘There are those that are traffic guards.’

The data in (18) and (19) are descriptively simple but theoretically relevant. Catalan resorts to the dedicated partitive clitic en / n’ to pronominalize the pivot of the existential sentences. There is thought to be a tendency to substitute the partitive in these cases for the accusative form, but not in all contexts (e.g., Bartra 1987; Jané 2001). Spanish, on the contrary, uses morphologically accusative forms (and even the covert pronoun; see De Benito Moreno 2016; Agulló 2022) to pronominalize the pivot. As the inquiry proceeds, I will demonstrate that the term accusative does not improve matters here; I will argue that proforms, such as as those in (19), are best regarded as partitives, on account of its φ-feature defectiveness.

4.3. Phi-Feature (in)Completeness and Types of Clitics

The samples in (19), in fact, are potential counterarguments to Belletti’s (1988) and Lasnik’s (1992) hypothesis that the pivot receives Partitive case; if these forms are morphologically accusative, why they should be instances of Partitive case? As an anonymous reviewer gently pointed out, if the assumption is not made to follow from more general principles of grammar, it seems somewhat unnatural or arbitrary. The answer to this lies, I believe, in the φ-feature defectiveness of the Spanish forms in (19). Note that, if we take these forms to be purely accusative, we are left with no explanation of the person asymmetry in (20):

(20)	a.	A	los	españoles	se	{los / os / nos}	ve	felices.
		dom	the	spaniards	imp	{3pl / 2pl / 1pl}acc	see	happy
		‘One can see {you / us / they} Spaniards are happy.’
	b.	{Los / *Os / *Nos}		hay	felices.
		{3pl / 2pl / 1pl}acc		is	happy
		‘There are some {of them / of you / of us} happy.’

It is generally agreed upon that accusative clitics can agree in first, second, and third person, as in (20a). I will now remain neutral as to whether cases like (20a) are instances of anti-agreement in the sense of Boeckx (2008) or of unagreement in the sense of Rivero (2004), as the decision does not hinge on the validity of the argument developed here. Los españoles (‘the Spaniards’) in (20a) can trigger first-, second-, and third-person agreement in the clitic, contrary to what happens in (20b) (see Rodríguez-Mondoñedo 2006, 2007a, 2007b, 2019; Gràcia i Solé and Roca Urgell 2017 for similar data); the clitic related to the pivot can only bear third-person agreement. The clitic pronoun in (20b) can hence be regarded as φ-feature-defective or incomplete; it lacks both first- and second-person agreements. Digressing briefly, φ-feature defectiveness or incompleteness makes clitics out of existentials bear close resemblance to so-called predicative clitics in Spanish of the sort, as illustrated in (21), which are also φ-feature-defective:

(21)	a.	Los	otros	no	siempre	lo	son.
		the-m-pl	others	neg	always	cl.acc	are
		‘Not always the others are so.’ (CORPES XXI, Mexico)
	b.	Son	pecados	–claro	que	los	son–.
		are	sins	of course	that	cl.acc-pl	are
		‘Those are sins; of course they are so.’ (CORPES XXI, Dominican Republic)
	c.	La	cordillera	no	es	nadie,	pero
		the-f	range	neg	is	nobody	but
		todos	la	somos.
		all	cl.acc-f	are
		‘The range is nobody, but we all are the range.’ (CORPES XXI, Chile)

The clitics in (21) pronominalize a predicate, but, quite generally, the person, number, and gender features are stripped away, as in (21a); the neuter clitic lo is used no matter the person, number, or gender specification of the predicate subject. However, there can be seen to be two variation phenomena at stake. (a) In certain varieties of Spanish, mainly Caribbean and Central, the predicative clitic can retain number features, as shown in (21b): the clitic los agrees in number with pecados (‘sins’), contrary to expectation. (b) Mainly in Chilean Spanish, the predicative clitic can also bear gender agreement, as seen in (21c); the clitic la agrees in gender with cordillera (‘mountain range’). Predicative clitics are generally held to be neuter- and thus φ-feature-defective, as seen in (21a); the proform is void of person, number, or gender features. φ-feature enrichment can take place, and predicative clitics may acquire number feature agreement, as seen in (21b), or gender feature agreement, as seen in (21c).

Defectivity or incompleteness as regards φ features, hence, holds across different types of clitics, but to varying degrees: (a) clitics out of existentials, which are person-defective, and (b) predicative clitics, which are person-, number-, and gender-defective. φ-feature enrichment can take place in the latter class, as predicative clitics may acquire number and gender features in certain varieties. As should be noted, φ-feature defectivity in clitics out of existentials holds regardless of the dialectal variety: no φ-feature enrichment can take place. Let us build upon this claim. Haber (‘to be’) existential sentences are generally held to display third-person agreement by default: the pivot position, be it singular or plural, does not trigger number agreement. Plural agreement is, nevertheless, widely attested and is even more frequent than par default singular agreement in some varieties (see, among others, Claes 2014, 2016; Pato 2016). In some varieties, even person agreement with existential haber (‘to be’) is found. The reader is referred to Castillo Lluch and Octavio de Toledo y Huerta (2016) for essential data and to Gràcia i Solé and Roca Urgell (2017) and Fábregas (2022) for some theoretical analyses. The basic descriptive data are shown in (22):

(22)	a.	Aquí	habemos	seis	vecinos.
		here	are	six	neighbors
		‘There are six of us neighbors.’
		(COSER, Casas de Soto (Valencia) COSER-4308_01)
	b.	Habemos	poca	gente.
		are	few	people
		‘There are few people among us.’
		(COSER, Mahíde (Zamora) COSER-4617_01)
(23)	a.	Imbéciles	los	habemos	en	todos	lados.
		idiots	cl.acc-m-pl	are	in	every	places
		‘Idiots, there are some of us everywhere.’ (Corpus del Español, Uruguay)
	b.	Los	habemos	mucho	más	bestias.
		cl.acc-m-pl	are	much	more	savage
		‘There are some of us that are much more savage.’

As it seems, seis vecinos (‘six neighbors’) in (22a) or poca gente (‘few people’) in (22b) trigger both number and person agreement. That being the case, the definiteness effect seems to hold at any rate (Rodríguez-Mondoñedo 2006; Fábregas 2022); only indefinite pivots are allowed. Person agreement, hence, cannot be said to lift the definiteness effect. Essential in this respect is that, whenever the pivot of the existential is pronominalized, the clitic does not show person agreement, as shown in (23); only the three-person-by-default los may be used. Thus, the ungrammaticality of first and second personal pronouns, as illustrated in (20b), remains, no matter the person agreement with the verb. I believe the φ-feature incompleteness of the clitic pronoun in (19, 20b, 23) to be strong evidence in support of the hypothesis that these clitics belong to a different class of clitics. Ultimately, differences across classes of clitics will be shown to be derived from the Partitive case assignment hypothesis.

4.4. Layers in the DP: D as the Locus of Person

Noteworthy is the fact that, both in clitics out of existentials (20b) and in predicative clitics (21), person agreement seems to be highly restrictive. In fact, person agreement will be taken here to be at the core of the contrast between standard accusative clitics, as in (20a), and clitics out of existentials, as in (19, 20b, 23). I believe the asymmetry in person features in (20) to be essential for the architecture of the grammar; by no means is it merely accidental or hazardous. With this in mind, it is safe to conclude that Spanish clitics out of existentials are not endowed (person) features, nor are predicative clitics of the sort illustrated in (21). I will take this gap as evidence in support of the following diagnosis: clitics out of existentials belong to a different class and are not just person-stripped accusatives. I take this class to be the Partitive series, which, similarly to (18a), does not show person agreement. I am using the term partitive—and will use this term henceforth—much along the line of Cardinaletti and Giusti’s (1992, 2006) use of quantitative (see also Zamparelli 2000, §4.2.3); the clitic is a subnominal clitic, following Pollock (1998), that pronominalizes not a DP, but lower layers of it. In that sense, it can be regarded as quantificational. The reasons behind my decision to assign different structural layers to each type of pronoun are multiple. Fábregas (2022), for instance, takes the pivot of the existential to be a Noun Phrase, which is a proposal I believe to be essentially correct. To sharpen the issues here, let us assume that different positions within the nominal phrase are associated with different interpretations, i.e., Longobardi’s (1994) topology within the nominal domain. It has come to be known that referentiality and predominantly, Person features are closely correlated to the DP projection (see, among others, Longobardi 1994; Martin et al. 2021). Longobardi (2008) and Bernstein (2008) have specifically argued for the functional head D(eterminer) to be the locus of person features. If person features are, in fact, located in D, it is clear that clitics out of existentials, along with predicative clitics, cannot be related to the same layer as fully fledged personal pronouns (i.e., those in (20a)). Recall that, in relation to (19, 20b, 23), it was derived that clitics out of existentials are not provided with person features, nor are predicative clitics in (21). That is, they lack person features and are thus φ-feature incomplete. If this lack of person features is taken to be the basis of a configurational diagnosis, as in the work by Martin et al. (2021), clitics out of existentials and predicative clitics cannot be related to a DP position. To put it differently, they are not related to the highest layers of the DP, as that is where person features are located. Zamparelli (2000, §1.1.4.2) noted similar contrasts for Italian and also attributed different layers to (a) object clitic pronouns, (b) predicative lo, and (c) partitive ne.

Let us recapitulate what has been tackled thus far. The claims in (17a) and (17b) have been used in support of the hypothesis that clitics out of existential clauses in Spanish behave like partitive pronouns in several respects: (i) they are used to pronominalize the pivot of the existential; (ii) they are void of person features, much like predicative clitics; and (iii) they are presumably related to a non-DP position, that is, to a Noun Phrase position, also lacking person features.

4.5. Partitive Pronouns in Spanish with Unaccusative Predicates

The validity of the claims in (17a) and (17b) has been argued to stem from different areas of grammar. There remains to account for the claim in (17c). Note that, if the claim in (17c) proves true, the partitive nature of clitics in existentials would by no means be an ad hoc theoretical move. A complete analysis of the emergence of a class of unaccusative partitive clitics in Spanish, as I will hereinafter refer to them, lies far beyond the scope of this paper. I refer the reader to Agulló (forthcoming), where some of these ideas are fully developed. Let us review, at least, the basic empirical support for the claim in (17c). The data suggest that a close resemblance can be traced between partitive pronouns of the sort in (18) (and, by extension, (19)) and unaccusative partitive pronouns, such as those italicized in (24) (taken from Agulló, forthcoming):

(24)	a.	Los	existimos	que	buscamos	respetar	las	normas
		cl.acc-m-pl	exist	that	search	respect	the	regulations
		de	tráfico.
		of	traffic
		‘There are some of us that respect traffic regulations.’ (X (formerly Twitter), @LaPatataComuni1, 9/04/2023)
	b.	Los	estamos	que	nos	gusta	echar	unas
		cl.acc-m-pl	are	that	cl.dat.1-pl	like	throw	some
		risas.
		laughs
		‘There are some of us that like to laugh.’ (X (formerly Twitter), @Naxo82026228, 20/11/2022)
	c.	De	esos	los	abundan	por	ahí
		of	them	cl.acc-m-pl	abound	for	there
		‘Of that kind, they abound everywhere.’ (X (formerly Twitter), @PabloGa98541569, 3/12/2022)

The fact that verbs such as existir (‘to exist’), presentative estar (‘to be’), or abundar (‘to abound’) belong to the unaccusative class, as formulated in Perlmutter (1978) and Burzio (1986), is clear from a bundle of aspectual, syntactic, and semantic properties: (i) unaccusative verbs allow for bare noun phrases as postposed subjects (Torrego 1989; Contreras 1996); (ii) some unaccusatives admit that adjectives ending in -ble, such as variable (‘variable’), are contrary to unergatives, which generally reject them (cfr. the ungrammaticality of nadable ‘swimmable’) (De Miguel 1986a, 1986b); and (iii) unaccusatives are low-agentivity predicates (Perlmutter 1978; Pustejovsky 1995). Existir (‘to exist’), for instance, has been generally held to be an unaccusative predicate (e.g., Batiukova 2004; López Ferrero 2008; Jaque Hidalgo 2010). The verbs in (24) are clearly imperfective, even though unaccusatives have been widely linked to perfectivity (e.g., Bosque 1990, p. 170 and seq.; Hernanz 1991; De Miguel 1992).

The data in (24) seem puzzling, at least prima facie, due to several reasons. The syntactic subject of unaccusatives behaves similarly to a semantic object (see, among many others, Mendikoetxea 1999; Campos 1999; Batiukova 2004; Alexiadou et al. 2004; Baños 2015; Bosque and Gutiérrez-Rexach 2009), but it triggers both person and number agreement with its verb. It is thus safe to conclude that the argument of the unaccusative predicates in (24) occupies the subject position. That being the case, how is the pronominalization of the subject to be accounted for? To put it differently, if the subject of the examples in (24) triggers number and person agreement, how is it that it can be pronominalized? If the unstressed pronouns in (24) were to be ascribed to the accusative pronoun series in Spanish, we would expect neither person nor number agreement with the verb. Invoking object agreement in cases like (24) would not improve matters here; a lack thereof in Spanish seems clear, and it would be an ad hoc or arbitrary theoretical move.

A possible solution to this paradox lies in a well-known fact about unaccusative verbs in Romance languages: in languages that have explicitly partitive pronoun series (recall the claim in (17a)), the partitive pronoun can be used to pronominalize the subject of the unaccusative verb (Belletti 1979; Belletti and Rizzi 1981; Bentley 2004, 2006, §6. 3). The relevant data are shown in (25) with examples from Italian (25a), French (25b), and Catalan (25c):

(25)	a.	Ne	sono	arrivate	due.
		cl.part	are	arrived	two
		‘There arrived two.’
		(Example taken from Bentley 2006, p. 253)
	b.	Il	en	est	allé	beaucoup	à	Venise.
		expl	cl.part	are	gone	lots	to	Venise
		‘There have gone many to Venise.’
		(Example taken from Manente 2008, his (110c))
	c.	Avui	ja	no	en	surten.
		Today	already	not	cl.part
		‘Today, they no longer come out.’
		(Example taken from Todolì 2002, p. 1376)

Predicates like arrivare (‘to arrive’) in (25a), aller (‘to go’) in (25b), and surtir (‘to come out’) in (25c) are proper unaccusatives. As such, their internal arguments can be pronominalized with the corresponding partitive pronouns: It., ne; Fr., en; and Cat., en/n’. Recall that the puzzling nature of the Spanish examples in (24) stemmed from the fact that, regardless of the pronoun, the argument anyhow triggers number and person agreement with the verb. The same situation holds in the examples in (25); the internal argument is ne-pronominalized and, at all events, it triggers number and person agreement. I will take, from now on, the pronouns in the examples in (24) and (25) to be of the same sort and to belong to the φ-feature-defective class of partitive pronouns. The hypothesis, which still needs to be fully unfolded, awaits further research, but I will regard it here as a diagnosis that a more general mechanism seems to be at work: the pivot, as suggested by Belletti (1988) and Lasnik (1992), bears Partitive case and is thus φ-feature-defective.

4.6. Microcontact and Phi-Feature Enrichment

Throughout this paper, I have tacitly assumed that the φ-feature defectiveness of the pivot is a consequence of it being assigned Partitive case. The assumption is on no account trivial or cursorily, as it ultimately hinges on the nature of the definiteness effect and the principles governing case assignment and interpretation. The account of the effect I have worked out here bears close resemblance to other syntactic accounts of the definiteness effect, along the lines of Safir (1982), Longa et al. (1996, 1998), and Belletti (1987, 1988). The basic intuition underlying these accounts is that configurational principles (i.e., namely, but not exclusively, case assignment), such as syntactic chains under Safir’s (1982) purview or Belletti’s (1987, 1988) Partitive case assignment, are at the core of the definiteness effect.

The φ-feature incompleteness of the pivot position of the existential, which is, under my proposal, a side consequence of Partitive case assignment, is not new in the literature on the definiteness effect. I will flesh out Rodríguez-Mondoñedo’s (2006, 2007a, 2007b, 2019) proposal, whose predictions I believe to be not only accurate but also quite promising for Spanish. Rodríguez-Mondoñedo’s works (2006, 2007a, 2007b, 2019) (drawing heavily on Rigau 1988, 1991, 1993, 1997) have convincingly argued that, in Spanish, pivots of existential haber (‘to be’) are not endowed with person features. The head of the small vP, which case-checks the nominal under the current assumptions, only has number features. These ideas can be easily argued to account for the restriction on person features that seems to hold in the pivot position of existential haber (‘to be’) in Spanish. In fact, proper nouns, such as those in (26a), or personal pronouns, such as those in (26b), are banned in the pivot position (see also Agulló, forthcoming):

(26)	a.	*Hay	Juan.
		is-loc	Juan
		‘There is Juan.’		(Rodríguez-Mondoñedo 2007a, p. 37)
	a.	*Me/Te/Nos		había.
		1.sg.acc/2.sg.acc/1.pl.acc		there.was
		‘There was/were me/you.’		(Rodríguez-Mondoñedo 2007a, p. 38)

A restriction on person features, as illustrated in (26), would bar personal pivots, mainly items in the leftmost positions of Aissen’s (2003) Definiteness Scale. It should be born in mind that, in standard definiteness hierarchies (e.g., Aissen 2003; see, particularly, von Heusinger and Kaiser 2003, 2007, for Spanish), proper names, such as those in (26a), and personal pronouns, such as those in (26b), are thought to occupy the highest positions of the hierarchy. The facts in (26) clearly show that a ban against (person) features seems to be at work. The wisdom of such an approach, however, shall be questioned. Rodríguez-Mondoñedo’s (2006, 2007a, 2007b, 2019) account places the theoretical burden of the definiteness effect on person features. As a consequence, other definite pivots without person features are left unaccounted for. Illustrative in this regard are the examples in (27):

(27)	a.	*Hay	Sevilla.
		is-loc	Seville
		‘There is Seville.’
	b.	*Había	Twitter.
		there was	Twitter
		‘There was Twitter.’

On no account are pivots such as Sevilla ‘Seville’ in (27a) or Twitter in (27b) specified for person features, and interestingly, they are equally prevented from occupying the pivot position of the existential. Note that the ungrammaticality is as robust in the samples in (27) as it is in the samples in (26), no matter the φ features involved. The examples in (27), indeed, clearly suggest a refinement of Rodríguez-Mondoñedo’s (2006, 2007a, 2007b, 2019) hypothesis. Nevertheless, I believe this potential refinement does not compromise the validity of the hypothesis, which remains, at least for Spanish, tenable. I will thus regard pivots of existential sentences in Spanish as φ-feature-defective, even though some additional restrictions are needed to fully account for the definiteness effect in Spanish.

In what follows, the hypothesis will be put forth that Spanish in contact with Catalan amnesties the definiteness effect. A theoretical explanation will be developed that bears upon two mechanisms: (a) Partitive case assignment and (b) φ-feature enrichment of the pivot in contact situations. The core of the hypothesis, which will be argued to follow from the data in a principled manner, can be stated as follows:

(28)	Spanish in contact with Catalan amnesties the definiteness effect as a consequence of Partitive case no longer being linked to indefiniteness.

Note that, under the hypothesis in (28), the definiteness effect is nothing but a descriptive artifact; its consequences are to be accounted for with resort to more general principles of grammar, as Case assignment. The hypothesis in (28) presupposes the claims listed in (29):

(29)	a.	Partitive case is directly related to non-definiteness, as put forth by Belletti (1987, 1988).
	b.	The non-definiteness of the pivot is tied to the defectiveness of the DP, which also accounts for the φ-feature defectiveness of the pivot.

The claim in (29a) directly links Partitive case assignment to non-definiteness. Belletti (1987, 1988) thinks of it as a sort of structural condition on Case assignment; Partitive case is genuinely assigned to non-definite noun phrases. Definite pivots are thus barred as a consequence of this structural condition on Partitive case assignment. Recall that I have taken rather independent pieces of data to support the hypothesis that the pivot position of haber (‘to be’) existential sentences bears Partitive case: (1) the φ-feature defectiveness of clitics out of the pivot position, as seen in (19, 20b, 23); (2) the fact that the φ-feature defectiveness of the clitic out of the existential holds even in cases where haber (‘to be’) shows person agreement, as shown in (23); and (3) the ability of these φ-feature-defective clitics to pronominalize the syntactic subject of unaccusative verbs in Spanish, as in (24). It then naturally follows that the pivot position is, by default, non-definite, as required by Partitive case assignment. If, additionally, it is assumed that D is the locus of person features, along the lines of Longobardi (1994, 2008), Bernstein (2008), and Martin et al. (2021), the claim formulated in (29b) also follows without ad hoc theoretical moves; the pivot position of the existential is D-defective and, as a consequence, φ-defective. The φ-defectiveness of the pivot straightforwardly accounts for the absence of person features in clitics out of existentials, as argued in relation to (19, 20b, 23).

The default situation in existential sentences (i.e., Partitive case assignment and φ-feature incompleteness) has been proven to hold in general Spanish. The hypothesis, as formulated in (28) and decomposed in (29), does not seem to be fully at work in varieties of Spanish in contact with Catalan. In these varieties, as I will argue, Partitive case assignment is not tied to definiteness. The relevant descriptive data have been shown throughout the paper, particularly in (12), (13), and (15), but I will recast it here as evidence that these varieties allow for definite, specific pivots that are properly referential. The data shown in (30) are strong evidence in this respect:

(30)	a.	Había	el	número	uno	caído.
		was	the	number	one	fallen
		‘There was the number one fallen.’
		(Benimodo (Valencia), COSER-4306_01)
	b.	Aquí	había	el	tráfico	muy	grande
		here	was	the	traffic	very	big
		‘Here, there was (the) heavy traffic.’
		(Queixas (Cabanelles) (Gerona), COSER-1707_01)
	c.	Había	el	camión	que	lo	esperaba.
		was	the	truck	that	it	waited
		‘There was (the) truck that waited for him.’
		(San Climent (Mao) (Menorca), COSER-5003_01)
	d.	Ha	habido	los	toros	al		NO-DO
		have	been	the	bulls	in-the		NO-DO
		‘There were the bulls on NO-DO.’
		(Vinebre (Tarragona), COSER-4011_01)

The data in (30) illustrate what has been referred to either as eventive existentials (McNally 1997; Villalba 2013) or as presentative constructions (De Cesare 2007; Cruschina 2016, 2018). In these constructions, the pivot, which is referential and specific, is the subject of a stage-level predication, in the sense of Carlson (1977). Quite interestingly, these examples are thoroughly ungrammatical in general Spanish. I will refrain from attributing a small clause analysis à la Stowell (1978) to the sequences in (30), as I believe there to be several strong counterarguments to the proposal (see Villalba 2013). What is relevant to our purpose here is that the pivot is fully identifiable, and thus specific and referential. I take this semantic interpretation to yield a structural diagnosis; the pivots in (30) are full-fledged DPs. If it is assumed that D is the locus of person features, it then naturally follows that the pivots in (30) are no longer φ-defective.

The question should be addressed as to which mechanisms underlie this case of syntactic variation. The hypothesis that existential pivots in varieties of Spanish in contact with Catalan are full-fledged DPs and hence equipped with φ features is specifically called into question by an anonymous reviewer. It is my belief, accordingly, that if the syntactic variation at work is not made dependent upon more general principles of grammar, the explanation seems ad hoc, arbitrary, or unmotivated. If the hypothesis in (28) and the presuppositions underlying it, as seen in (29), are thought to be essentially correct, some additional principles of syntactic variation and change have to be invoked. To put it differently, how is it that pivots in (30) are full-fledged, whereas pivots outside varieties of Spanish in contact with Catalan are not?

A potential solution to this problem stems from the literature on language contact and, particularly, from the notion of interdialect variant or form. The hypothesis will be put forth, particularly, that examples with definite, specific pivots in the existential clause of the sort in (30) are the result of contact between dialects of Spanish and dialects of Catalan. Theories on linguistic variation and change, particularly since pioneering research by Weinrich et al. (1968) and Labov (1964, 1972, 1978), have furthered our knowledge on the structural and social consequences of linguistic contact. It is generally agreed upon that, whenever grammar is subject to change, the process “entails not merely formal differences but functional differences as well” (Harris 1984, p. 314). Contact between different languages and their dialectal varieties is generally acknowledged to lay the foundations for the emergence of (a) linguistic variants hitherto non-existent or unnoticed, (b) new dialectal varieties, and (c) the obsolescence of others (see, among many others, Siegel 1985; Kerswill 2013; Cerruti and Tsiplakou 2020).

Alternatively, already in-use variants may be reallocated, i.e., recycled or used differently. New linguistic variants—so-called interdialect variants in Trudgill (1989a, 1989b, 1992), Britain (2005), and Almeida (2020), among many others—can thus arise when two languages come into contact with each other. Inter-dialect variants can be thought of as “novel features not found in any of the established contributing dialects” (Tuten 2001, p. 325). Interdialectalisms, which are, essentially, a special kind of dialect mixing, result when “speaker-learners reanalyse or rearrange forms and features of the contributing dialects” (Tuten 2006, p. 187). Interdialect forms stem, thus, from a mixture, rearrangement, or readaptation of different features or constructions belonging to two or more dialects.

Note that, in Catalan, eventive existentials with definite and specific pivots are grammatical, as shown in the examples in (31):

(31)	a.	Al	wharf	hi ha	la Denise	que	m’	espera.
		in-the	wharf	loc is	the Denise	that	me	awaits
		‘In the Wharf, there is Denise that waits for me.’
		(CTILC, Sagarra, Josep M. de: La ruta blava, 1964)
	b.	Hi ha	la Verge	dels	Dolors	asseguda.
		loc is	the Virgin	of-the	Sorrows	seated
		‘There was the Virgin of Sorrows seated.’
		(CTILC, Raventós i Domènech, Jaume: Memòries d’un cabaler, 1932)
	c.	Hi ha	la vella	Kaa	fent rotllanes	per terra
		loc is	the old	Kaa	doing rolls	on ground
		‘There was the old Kaa spinning around on the ground.’
		(CTILC, Manent, Marià (T): El llibre de la jungla, 1920)

The structural parallelism between the examples in (30) and those in (31) is not casual. On the contrary, eventive existentials with definite pivots in varieties of Spanish in contact with Catalan can be accounted for in a principled way. If they are taken to be interdialect variants, the fact that these varieties allow for full-fledged DPs follows quite naturally from the hypothesis in (28). Thus stated, the hypothesis in (28) yields the following prediction: the relation of Partitive case assignment to non-definiteness, as stated in (29a), is overridden. An explanation, in fact, suggests itself: if existentials in Catalan, as those in (31), are taken to supersede the relation of Partitive assignment with non-definiteness, a lack thereof in varieties of Spanish in contact with Catalan would be a direct structural consequence of language contact.

5. Conclusions and Further Prospects

Throughout this paper, the dialectal and grammatical variation of the definiteness effect has been thoroughly inquired into with data from, mainly, spoken corpora, but also written corpora, such as CORPES XXI, Corpus del Español, or X (formerly Twitter). The main assumption underlying this research is that grammatical restrictions, which are thought of as hypotheses about the limits of human possible grammars, are a rich path of scientific inquiry and thus have to be dealt with. Several rather independent claims, as stated in (1–3) below, have borne close inspection:

• The definiteness effect is amnestied in varieties of Spanish in contact with Catalan. As a consequence, existential sentences allow for definite, specific pivots, which can even be the subjects of stage-level predicates in eventive existentials.
• The pivot position of the existential, as hypothesized, is assigned Partitive case. Evidence in this respect has been brought to the fore: (a) Clitics out of the pivot position are φ-feature-defective or incomplete. (b) This φ-feature defectiveness holds even in cases in which haber (‘to be’) bears person agreement. (c) These φ-feature defective clitics can pronominalize the syntactic subject of unaccusative verbs in Spanish. (d) If the head D is taken to host person features, it stands to reason that these clitics cannot be related to a full-fledged DP projection.
• Under Belletti’s (1987, 1988) purview, Partitive case was closely tied to indefiniteness, but the relation is somewhat of a structural condition on Case assignment. Spanish in contact with Catalan overrides the non-definiteness requirement of Partitive case assignment. As a result, definite, specific constituents are grammatical in the pivot position of the existential.

The claims in (1–3) do not exhaust all theoretical possibilities, nor do they account for all of the relevant dimensions of the syntactic variation at stake. Particularly, the question shall be brought up as to whether this instance of variation is neatly syntactic in nature or owes to some alternation within the lexicon. Spanish presentative-existential constructions are widely held to alternate between haber (‘to be’) and estar (‘to be’), but the alternation is far from free or casual; in general Spanish, definite and specific pivots are generally introduced with estar (‘to be’). If a decrease in the frequency of presentational estar (‘to be’) was found in the dialects of Spanish in contact with Catalan, the claims in (1–3) could be easily argued to boil down to a lexical alternation. This and other questions await further research.