The multilingual repertoire of the Haitian community in Chapecó (SC, Brazil): Patterns of linguistic evolution in a South–South migration context

4.1 The shift in diglossia and the new function of French

In Haiti, HC is the language of everyday use, while French takes on the typical functions of a high language (Valdman 2015, 361). Possibly, this HC/French diglossia has been present in Haiti since the nineteenth century (Neumann-Holzschuh 2003, 295). Despite the fact that HC has recently gained official recognition and is also, on paper, a co-official language (and despite the potential for a functional space for hybrid HC/French discourses: Cothière 2016/17), the co-existence of both languages in Haiti can still generally be considered a case of diglossia (Valdman 2015, 14). This was defined by one of my informants in the following terms (cf. also Benítez Oviedo 2023, 73–4):

Other accounts leave no doubt that the sociological-identity value comes from HC, while French is often perceived as a foreign language:

Even though there isn’t necessarily a negative estimation of French, the Haitian diglossia is indeed accused of being a source of social inequality, to the extent that it creates a ‘linguistic barrier’ in Haiti (unlike the urban Brazilian reality surrounding my informants, which is strongly monolingual, in Haiti large sectors of the population do not have the opportunity to learn to actively use the language that is then utilized in institutional contexts and at the university):

The issue of diglossia in Haiti was emphasized by many students – often ‘good’ speakers of French – during the interviews. They suggested ‘parar com esse negócio de francês’ (‘stopping this French business/thing’) and instead advocated for further promoting HC. Indeed, far from being speculative, the idea that such a measure would likely contribute to increased development in Haiti has been demonstrated in socio-linguistic studies (Hebblethwaite 2012). Moreover, there is a strong sense of solidarity with Haitians who do not speak French (as in the case of F among the group of students in example (4)), as well as a clear perception that this inequality vanishes in the Brazilian context, which is thus viewed as fairer. In Brazil, at least linguistically, all Haitian migrants are on equal footing when it comes to the need to learn Portuguese.

Moreover, it is noteworthy that HC not only constitutes the quintessential identity language but also carries a predominantly positive linguistic ideology regarding its origin. Gone are, at least among my informants, the well-known colonial heritage ideology according to which Creoles are corruptions or deformations of the European languages from which they derive. If remnants of this ideology still exist today, it seems that the youth definitively abandon it upon leaving Haiti. And, in cases where the conception of HC as a deformation of French does emerge, it is accompanied by a sense of pride and attributed to the free will of the ancestors (voluntary deformation as a rejection of French colonization). In this context, the arguments put forward by two Haitian students in my presence are interesting – both computer science students and therefore, we could assume, unaware of the ideas about Creolization that many linguists have been disseminating (in fact, the opinion of the second student clearly aligns with well-known ‘anti-exceptionalist’ linguistic postulates like those of Chaudenson, Mufwene, etc.):

The linguistic ideologies discussed so far sufficiently explain why HC remains strong in the diaspora, now turned into a minority language for intragroup use, while French, in principle, is abandoned: if this latter language never became part of Haitian identity, even less so in the diaspora in Brazil, where French plays no role in public life. In this country, the position of high language is occupied by Portuguese, so that, simplifying things, we would say Haitian-Brazilians replace a HC/French diglossia with a HC/Portuguese one.

Despite the above, there are two phenomena inherent to the new socioeconomic and ideological context surrounding the Haitian-Brazilian migrant community that allow for a certain presence of French; in a way, claiming to know French (whether or not this is actually the case) acquires new semiotic value, signaling both globality and education. First, as I will further analyze in Section 4.3, Haitians in the diaspora have an open attitude toward other languages and cultures and positively value the knowledge of any language; thus, knowing French implies greater ‘globality’. Second, the French language can serve the social function of indicating high status, specifically that corresponding to skilled migration (although it is unclear to what extent Brazilians are aware of the existence of a HC/French diglossia in Haiti and, more importantly, its social implications). In this sense, it is no coincidence that, although French has limited representation in the linguistic landscape of Chapecó, the owner of a Haitian snack bar displays a sign at the door of the establishment indicating that it is open in both Portuguese and French (see Figure 3): aberto/ouvrir (interestingly, the use of the infinitive ouvrir – instead of ouvert ‘open’– as a translation of the Portuguese past participle aberto must be related to the influence of HC^[6]).

In the same vein, I interpret the fact that some informants during the interviews start by asserting that they speak French without actually doing so (or, at least, displaying notable insecurity and seeking to switch to Portuguese when asked in French) as an attempt to present themselves – in a sense, market themselves – as qualified individuals (this might also explain the apparent contradiction observed in the responses to the questionnaire about languages spoken at home in Benítez Oviedo (2023, 67–8)). Haitians are aware of the prestige of French as an international and cultural language, and moreover, they grew up in a context where French is a sign of education. The following example showcases this tendency to claim to speak French but avoid using this language in favor of Portuguese during the conversation:

Naturally, those Haitians whose knowledge of French is partial (i.e., mostly passive) decide to present themselves or not as ‘French speakers’ following, in the first instance, idiosyncratic aspects of their personality. Conversely, other informants acknowledged from the start that they do not know how to speak French actively (note the informant’s laughter at even imagining if he speaks it):

4.2 The role of Spanish in the repertoire

In Haiti, many people have had some experience with Spanish, influenced not only by the international prominence of the language but, more significantly, because it is the language of the neighboring Dominican Republic. Haiti shares the island of Hispaniola with the Dominican Republic, along with a complex history of both connections and conflicts (Barzen 2022, 66–89). The interaction of Haitians with Spanish can primarily be categorized into three types: first, Haitians who have completed several years of schooling, particularly younger Haitians, have often taken Spanish courses at school. Second, Spanish is somewhat present in Haiti, especially in border areas, due to the interactions between the Haitian population and those living across the border (or raya), where Haitians and their descendants also learn to speak Spanish (showing unique features often resulting from transfer from HC: Ortiz López 2010). Moreover, it is plausible that seasonal workers in Dominican bateyes (Jansen 2013, 80–1), as well as those returning after being deported from the Dominican Republic to Haiti, help disseminate insights into Dominican reality, possibly including elements of the country’s language. Third, there is a significant migrant population that settles more or less permanently in the Dominican Republic – not just near the border, but especially in Santo Domingo – and quickly acquires Spanish. The number of Haitian migrants in the Dominican Republic significantly outnumbers those in other countries – except for the USA (Audebert 2012, 98–103) – though, as noted, the Dominican Republic is generally the worst regarded among Haitians as a potential host country (Audebert 2012, 112, Martínez 2011, 51–70). Notably, the view of Dominicans as racists by Haitians – a perspective I won’t judge here – remains prevalent in the migration context outside the island (indeed, such comments were not uncommon among my Haitian-Brazilian informants).

Generally speaking, Haitian-Brazilians who know Spanish learned it during their time in the Dominican Republic:

Among my informants, there are seven people who, for various reasons, have some knowledge of Spanish. Nevertheless, for the considerations of this section and the selection of examples of structural phenomena in Section 5, I will only focus on two of them (not related to each other). Both informants are women who had been working in the Dominican Republic for about a decade (11 years in one case, 8 in the other); in both cases, the current length of residence in Brazil is similar to the time they spent in the Dominican Republic, during which time they had not initially planned to leave the country. The other informants who spoke some Spanish had either studied it in school or learned it after having stayed in the Dominican Republic for short periods, usually a few months, without clear prospects of wanting to settle there definitively.

In both cases, as with many Haitians around the world, the connection with the Dominican Republic remains to some extent present since they continue to have family in the Dominican Republic. The account below offers a striking example of what Audebert (2012, 77–81) terms famille transnationale (‘transnational family’), where siblings from the same family are spread across different countries (in this case, Brazil, Chile, Mexico, and the Dominican Republic, with the latter country hosting the most family members):

The experience with Spanish is strongly influenced by the Dominican context. The prominence of exposure to the Dominican variety of Spanish is such that the Spanish spoken by Venezuelans – who are notably present in Chapecó (Krug and Horst 2022, Benítez Oviedo 2023) – can be considered enredado (‘tangled, confusing’). This attribute was also applied to European Spanish (my native variety, which I used during the interviews in Spanish). This leads to a perception that sets Dominican Spanish (the ‘normal’ variety) apart from all other Spanish varieties (seen as somewhat confusing):

Although these informants may have taken some Spanish courses in the Dominican Republic (one of them explicitly stated this), it is reasonable to assume that the acquisition of Spanish – as would later happen with Portuguese in Brazil – predominantly occurred in an informal manner, through daily contact with the surrounding environment.

4.3 Global identity

Generally, Haitian-Brazilians expressed a very positive attitude toward multilingualism as part of a personal growth process in relation to their internationalization experience. As is clear from the testimony of the following informant, embedded in a part of the conversation where we alternated between French and Portuguese, languages are, above all, means of communication with other people.

In the words of another informant, Haitian people like to speak several languages, which is related to the fact that they are a traveling people, spread all over the world:

This ideology, highly positive about the knowledge of several languages, should, in my view, be linked to two factors inherent in the diasporic context of Haitians that promote the idea of inclusion in a global society: on one hand, the significance of the ‘transnational family’ (Section 4.2), which can foster a global awareness as well as a feeling of closeness to countries with a strong Haitian presence – and, indirectly, to the languages spoken in these countries; on the other hand, the entrepreneurial mindset of those who leave their country under the assumption that they likely will not return does not tie them to a host country; that is, they do not, on principle, close themselves off to possible new opportunities elsewhere. In the words of a student, one becomes a ‘global citizen’ the moment they obtain their first visa to leave their country:

5.1 What can be expected from a multilingual repertoire?

When speakers possess a multilingual repertoire, the emergence of code-switching within their discursive practices becomes an option. We have observed examples of extra-sentential code-switching interspersed throughout this work (6, 7), as well as some instances of inter-sentential code-switching (8, 11). Additionally, there is intra-sentential ^[8] code-switching, such as in example (14), where the switch also signals the transition from French to Portuguese for the remainder of the answer.

In the example above, the object pronoun is taken from Portuguese, whereas the first part of the sentence (containing the subject and the verb) is in French. Interestingly, the morphosyntactic properties of the object pronoun result from a combination of grammatical rules from both (Brazilian) Portuguese and French. First, this object pronoun does not take the form of an object clitic, as would have been the case in French (if the example had been entirely formulated in French, we would have had le ‘him’: elle l[e]’a appelé). Instead, it takes the form of the ‘independent’ subject pronoun (Haspelmath and the APICS Consortium 2013): ele. This is a regular feature of spoken Brazilian Portuguese, where, for example, a sentence like ‘I saw him’ would be formulated as eu vi ele rather than (eu) vi-o, the typical European Portuguese variant, or eu o vi (in either of the latter two cases with the clitic o ‘him’). Second, this object pronoun, being formulated in Portuguese, is used as an accusative object, as would be expected in French, where appeler is constructed with accusatives (this is why the accusative clitic le would have been used, but never the dative lui: *elle lui a appelé). This is a substantial departure from Portuguese, where the equivalent verb ligar ‘to call’ is constructed with an (analytical) dative/oblique object, preceded by the preposition p(a)ra: ela ligou p(a)ra ele/*ela ligou ele.

However, in this and the following sections I will address cases where the use of multiple languages does not occur through code-switching (the most iconic way to utilize the different possibilities of the repertoire), but through other means, such as the combination of elements formally belonging to the inventory of one language with the semantic and syntactic properties of (partially equivalent) elements in another language.

Multilingual speakers can use all the resources of their repertoire, either consciously or, more often, unconsciously. In other words, speakers do not always ‘comply with social norms and context-appropriate selection of features’ but often make ‘efficient use of a variety of elements in a multilingual repertoire’ (Matras 2009, 10). Thus, features from one language can filter into discourses or sentences primarily conducted in another language, even more frequently in the case of minority languages (Barzen 2022, 221–2 and references therein). In managing a complex repertoire globally, it is also necessary to consider the idiosyncratic aspects, namely, the impact of informants’ experiences with different languages throughout their unique life stories (cp. Dombrowsky-Hahn and Fanego Palat 2024), as well as to take into account the momentary and situational nature of efficiently producing a specific linguistic formulation at a given point in discourse.

While, theoretically, it is not necessary to further classify linguistic contacts (other than as an organizational tool for upcoming sections), such classification can be explored based on the direction of linguistic influence, which presupposes a sequence in which languages were acquired. Thus, adopting the now classical perspective of Thomason and Kaufman (1988), we could examine substrate interference (or transfer, according to the terminology used by Selinker 1969, 1972, originating first from psychology) for the influence of the L1 on the L2 (or other non-native languages: L3, L4, etc.) and borrowing for the influence of the L2 (or other non-native languages) on the L1 (see also Clements 1996). This approach is illustrated (and applied to the migrant community under study) in Figure 4.

Figure 4

Linguistic features moving across languages in the repertoire of Haitian-Brazilians.

The recipient languages are not only those represented at the ends of the scheme: that is, HC also interferes with French or Spanish, in the same way that these two languages also borrow elements from Portuguese. Additionally, languages in intermediate positions can influence the others; most importantly, languages learned first can influence those learned later, as Riehl (2014, 92) concludes after reviewing the existing literature on L3 acquisition that “die Sprachen im Gehirn nicht nur alle untereinander vernetzt sind, sondern auch dass beim Erlernen von weiteren Sprachen bereits bekannte Sprachen effektiv genutzt werden” (‘the languages in the brain are not only interconnected but also that when learning additional languages, already known languages are effectively utilized’).

Lastly, it is crucial to highlight the emergence of two additional processes related to the development of a multilingual repertoire. First, languages acquired in adulthood, as well as those learned during childhood or adolescence exclusively in school settings – especially when the exposure time is quite limited – can be considered ‘non-native’ (L2, L3, L4, etc.) and thus are susceptible to phenomena associated with the fossilization of interlanguages (Roche 2013, 82–6, Selinker 1972). These phenomena, which in our case affect all languages except for HC, extend beyond substrate interference to encompass processes of simplification, regularization of patterns, and overgeneralization of a specific form across an entire paradigm. Second, language obsolescence (Aikhenvald 2012) triggered by disuse – which may incidentally result in structural consequences partly analogous to those observed in interlanguage grammar – affects the two types of languages that, throughout this work, have been identified as ‘forgotten’ within the biography of migrants, namely French and Spanish. The erosion of such languages, as Riehl (2014, 93) describes, constitutes an “umweltbedingte Attrition einer Zweit- oder Drittsprache, die nur teilweise beherrscht wurde” (‘environmentally induced attrition of a second or third language that was only partially mastered’)

5.2 Interlanguage grammar

When speaking non-native varieties, among other processes, “strategies of second-language learning” (Selinker 1972, 269–70) occur, involving simplifications and overgeneralization in the use of certain linguistic forms. I highlight only two of the phenomena that, as a result of these strategies, are more recurrent in my corpus. First, it is common for the masculine form to prevail over the feminine in words that in French and Spanish have gender inflexion:

As expected, the same phenomenon also occurs in Portuguese (to facilitate comparison, I provide examples with the corresponding cognates of the non-canonical forms presented in the previous examples):

Second, one of the most characteristic features of interlanguages – the overgeneralization of 3SG verbal forms across the entire grammatical person paradigm, leading to an ‘agreement mistake’ effect from the perspective of canonical monolingual varieties – is quite prevalent in my data. This phenomenon has been observed before, for instance, in interlanguages of Spanish spoken by Chinese migrants in Clements (2005, in press), and Schuchardt as early as 1888 noted that jargons – which correspond to both fossilized interlanguages (Selinker 1972) and to what today would be called pidgins – often overgeneralize the 3SG, unlike Creoles, which tend to overgeneralize INF forms (Schuchardt 1888, 251). These latter forms can also appear in migrants’ interlanguages, particularly in their earliest stages (cf. Ortiz López 2010, 81–2), but in my data, cases where the resulting form from the simplification of the verbal paradigm is the 3SG predominate.

Example (19) is noteworthy for displaying the alternation – typical in my data – between the overgeneralization of 3SG (produit) and forms with the morphology proper to the corresponding grammatical person (3PL produisent). Example (20) seems to alternate a 3SG form trabaja [tɾa'baha] (<Sp. trabaja) with an INF trabajá [tɾaba'ha] (< Sp. trabajar, with loss of -r, likely influenced by Portuguese). Further research is required to evaluate this issue, particularly focusing on how the prosodic and tonal properties of HC are transferred to Spanish and the other languages in the repertoire. These properties often make it challenging to distinguish between 3SG forms and infinitives, except in cases with irregular verb morphology, such as the 3SG form in example (21). On the other hand, a form like veni[r] in example (25) clearly displays INF morphology.

Once again, the same phenomenon is widely documented in the varieties of Portuguese spoken by Haitian migrants (the predominance of 3SG forms over INF is more evident here, but it should be noted that the overgeneralization of 3SG morphology is a marginal yet possible variant in some diastratically low, native varieties of Brazilian Portuguese^[9]: Castilho and Elias 2011, 458, Koch et al. 2011, 450–6, Steffen 2021, 144):

Lastly, in several examples interspersed throughout the work, we observe the most evident manifestation of simplification processes, namely, the absence of some elements – in this case, prepositions (or any other equivalent strategy for introducing adjunct NPs). This absence cannot be traced back to the syntax of either of the contributing languages:

5.3 Transfer from Haitian Creole

HC continues to be widely used in the diaspora and generally dominates the space of intragroup language, which means its influence on the other languages in the repertoire can spontaneously appear at any moment in discourse. Some of these transfer phenomena will be analyzed in this section; however, it is important to note two methodological issues regarding my data.

First, it is often challenging to determine whether a specific transfer phenomenon was already present in the learner’s variety before arriving in Brazil, that is, if it was already established in the L2 French varieties circulating in Haiti or in the L2 Spanish varieties of Haitian migrants in the Dominican Republic (cf. Ortiz López 2010). For example, one might wonder whether the use of the infinitive ouvrir with the meaning of an attributive participle, as discussed above (Section 4.1), was a relatively stable feature of ‘Haitian (L2-)French’ or emerged spontaneously in a Brazilian context.

Second, in terms of French, I mainly focus on those speakers who have had regular experience actively using this language. I reserve for future studies the utterances produced by those who are not accustomed to actively speaking French but who were forced to briefly engage in it with us. Here is just an example of such utterances, where the influence of HC is so pronounced that one could discuss a hybrid French/HC variety or an acrolectal or debasilectalized variety (Mufwene 2004) of HC, which, given the diasporic context I am addressing here, may feature code-mixing with Portuguese. In the example, a speaker attempts to switch from Portuguese, the language predominantly used in the prior conversation, to respond to a question posed in French (regarding how Haitians arrive in Chapecó); uncertain about the appropriateness of continuing in Portuguese – which one of the interviewers asserts not to understand – he temporarily shifted to French. Although he managed to articulate the idea, recognizing his inability to express himself effectively in French, he seizes the opportunity to request switching to HC.

Setting aside the lexical insertions from Portuguese (the pronouns eu and você – see further below – the non-canonical form falá – stemming either from a 3SG form with accentual shift or from an infinitive form^[12] – and the adverb aqui – perhaps used creatively as a preposition meaning ‘to/in’), each element of the French/HC parts merits separate analysis:

1. ou [u] originates from HC (deriving from and cognate with the French vous [vu]).

2. veni[r] [ve'ni] is from French, specifically from the infinitive venir [vəniʀ], but adapted to the phonetics of both HC (where there is no /ə/ and the final -r of French infinitives was never transmitted) and Brazilian Portuguese (which also lacks [ə] and where, dialectally, the deletion (apagamento) of final -r is common). There is no doubt that veni[r] is a French form, as in Portuguese the equivalent is vir [vi(ʀ)], and in HC, it is vin∼vini [vin(i)] (Valdman 1988, 264, Valdman et al. 2017, s.v. to come); the Spanish form venir [be'niɾ] could not have played any role, as this particular speaker did not speak Spanish.

3. moi [mwa] is a French form, clearly different from the HC cognate mwen [mwen] (retaining the old realization – general in French until the eighteenth century – of today’s diphthong [wa]). However, moi adopts the syntactic properties of HC mwen, thus being used as a dependent subject pronoun (in HC, French subject clitics like je do not exist, and the originally independent/topicalizer pronouns like moi were adapted as dependent subjects; moreover, a new cliticization process occurred, giving rise to m’, which can be used as a variant of mwen) (cf. Haspelmath and the APICS Consortium 2013).

4. palé can correspond to both HC palé (the basic verb form, derived from the French infinitive parler) and the French infinitive parler itself (with a slight phonetic adaptation in the interlanguage – the loss of /ʀ/ in implosive internal position of a word).

5. créole [kʀeɔl] is a French form, distinct from the corresponding cognates in Portuguese (crioulo) and in Creole (kreyòl [kɤe'jol]).

6. passapol [pasa'pɔl] diverges from the HC form (paspö) and seems to be based on the French passeport [pɑspɔʀ]; however, the inclusion of the second [a] likely comes from Portuguese passaporte [pasa'pɔʀʈʃi], while the articulation of the final rhotic [ɾ] as [l] is a phonetic feature typical (in Chapecó, even stereotypical) of interlanguages spoken by Haitians, resulting indirectly from a transfer from HC (which lacks the rhotic consonants of French/Spanish/Portuguese: Fattier 2013, §4) (see also Swiderski (2019) on similar changes affecting Spanish rhotics in the learner varieties of Haitian migrants in Mexico).

All things considered, when we examine instances like example (25) and specifically note the transfer from HC to French fragments – keeping in mind the informant’s initial intent to communicate in French – we indeed identify certain phonetic and syntactic-phonetic transfer phenomena, such as changes in rhotics and the usage of moi as a dependent subject pronoun. Nonetheless, it is clear that instances like these effectively showcase how multilingual speakers utilize various elements from their linguistic repertoire (HC, French, and Portuguese), often merging phonetic forms from one language with syntactic properties of another, or even combining phonetic material from multiple languages, while blending these with interlanguage phenomena (such as the simplification of verbal paradigms and the omission of prepositions). Although such hybrid utterances warrant further analysis, I will instead focus here on examples where transfer phenomena from HC can be identified more straightforwardly. The remainder of this section will focus on two such phenomena: one documented in the context of French, and the other observed across all non-native languages in the repertoire.

A notable phenomenon in French as spoken by Hatian-Brazilians is the use of vous ‘you(V)’ and the practical absence of tu ‘you(T)’ in non-formal situations.^[13] This points to the overgeneralization of vous as a neutral 2SG pronoun in terms of T and V (cf. Brown and Gilman 1960). Even more significant is that the common strategy in the oral varieties of many languages worldwide to use 2SG with a generic meaning (similar to the English one) is carried out among my informants through the use of vous, unlike canonical varieties of French, which would use tu with this same meaning^[14]:

This use would be hard to explain without acknowledging that the meaning of vous, as the sole 2SG pronoun, neutral in terms of T/V distinctions and thus the only possible 2SG form used for generic reference, replicates the use of HC’s cognate ou:

Regarding the potential influence of HC on this grammatical aspect in Spanish and Portuguese as spoken by Haitians, further research and data collection are necessary before drawing definitive conclusions.^[15]

Another phenomenon that may be related to HC is the use of present indicative verb forms with a perfective past meaning, which I only find in my data with aspectually dynamic verbs, but never with stative verbs. This phenomenon occurs across all languages in the repertoire (for ease of comparison, I provide examples with the same verb – fr. apprendre/sp. aprender/port. aprender – prototypically dynamic):

In HC, there is a basic dichotomy between marking or not marking the past tense – considering that TMA marking in this language (as in many other Creoles) occurs through the addition of preverbal particles. Unlike stative verbs, where a bare verb (without preverbal TMA markers) is interpreted as present tense, it is more common for the bare verb to be construed as perfective past in the case of dynamic verbs (Valdman 2015, 218):

The hypothesis suggests that multilingual speakers predominantly choose the present indicative form (and, to a lesser extent, the infinitive^[16]) in French, Spanish, and Portuguese as a replication of the bare verbs from HC. While the equivalence between bare verbs and the present indicative morphology with stative verbs matches the canonical varieties of French, Spanish, and Portuguese, the spontaneous extension of the present indicative to cover the other meaning of HC’s bare verbs, that is, the perfective past tense with aspectually dynamic verbs, represents a linguistic innovation by multilingual speakers. Schematically, this linguistic innovation could be represented in four steps (a type of schema that would naturally apply to other transfer phenomena where the L1 expresses through the same word or construction meanings that in the target language are expressed separately).

Figure 5

Language replication (transfer) in four steps.

While future quantitative analyses will be required to confirm this hypothesis, the distinct behavior observed between dynamic/processive and stative verbs in the non-native varieties studied here is unlikely to have arisen by chance (Figure 5). Furthermore, the influence of HC is suggested, too, by the rarity (or even impossibility) of other tense alternations within the non-native varieties of my informants: for instance, focusing again on past tenses, it is noteworthy that the present indicative never substitutes for the imperfect tense, which is consistently used in its canonical form to express non-perfective or continuous past.

Finally, the implications of the hypothesis are interesting for another reason: if we accept that HC has a predominantly West African linguistic substrate (Lefebvre 1998, which highlights the role of FonGbe), and we remember that the joint expression of the perfective past for dynamic verbs and the present for stative verbs is a very common feature in West African languages and even in a significant part of Niger-Congo languages (Maurer and The APICS Consortium 2013, §4; Lang 2023, 151; Dom et al. 2018),^[17] we might be observing an African matrix linguistic feature that has indirectly, through HC, been transferred to French, Spanish, and Portuguese varieties.

5.4 Borrowing from Portuguese

Throughout this work, examples have been interspersed in which some Portuguese elements are inserted into mostly French or Spanish fragments. In these cases, they are generally elements classifiable as ‘function words’ (Matras 2007, 35), such as affirmation and negation adverbs (pro-sentential, as in answers to questions) (sim ‘yes’, as in example (6), não ‘no’, as in examples (7) and (14)), connectors (such as the conjunction meaning ‘but’ – Port. mas – the one most easily borrowed in universal terms – Matras 1998, 2007, 54, Stolz et al. 2021–; see example (2)), etc. On the other hand, we have seen that in phonetically similar cognates, there seems to be a greater permeability to the adoption, total or partial, of the Portuguese element: this was the case of the [a] in the predominantly French form passapol (ex. (25)), to which could be added – according to reported evidence from Haitian students – the substitution of French différent [difeʀɑ̃] by Portuguese diferente [ʤife’ɾẽʧi] even when speaking French. Since French and Spanish are the least active languages in the repertoire, it would be particularly difficult to empirically determine the actual degree of integration of elements inserted from another language (this issue has traditionally been central to the discussion on whether such elements are borrowings or switches: see Haspelmath 2009 for conceptual issues concerning lexical borrowings, Matras 2009, 111 on the codeswitching-borrowing continuum, and Deuchar 2020 on the particular role of ‘peripheral’ types of integration in shedding new light on the distinction between these categories). As discussed in Section 5.1 (see especially Figure 4), the concept of borrowing is deliberately broad and encapsulates any kind of influence (matter/pattern replication) of later-acquired languages on earlier-acquired languages. This includes, of course, examples of syntactic borrowing, which this section aims to further elucidate to complement the preceding observations.

A feature present in L2 varieties of French and Spanish, attributable to contact with Brazilian Portuguese, is the use of null generic/arbitrary subjects:

As is well known, French is conventionally classified as a non-pro-drop language. While no Spanish variety can be clearly classified as such, Caribbean Spanish, especially Dominican Spanish (cf. Morales 1999, Gutiérrez Maté 2013 and references therein), exhibits some characteristics of non-pro-drop languages, including a notably high pronoun expression rate. Therefore, theoretically, null subjects might assume interpretations other than the conventional ones in canonical pro-drop languages (generally, referential elements recoverable by context). However, this is crucial: neither in monolingual varieties of French nor in Dominican Spanish have null subjects been observed to take on a generic/arbitrary meaning. Conversely, Brazilian Portuguese is often categorized as a ‘non-consistent null subject language’ (Holmberg et al. 2009), due to its tendency to use overt subject pronouns while still permitting their omission in certain contexts, including those with an impersonal, i.e., generic or arbitrary, interpretation. Impersonal subjects that potentially include the speaker – equivalent to one or you in English – are considered ‘generic’, while those excluding the speaker – equivalent to people or they in English – are considered ‘arbitrary’ (Holmberg et al. 2009, 63–4). Specifically, in the examples provided, since the speaker seems to be excluded from the referential scope of the null subjects, these subjects should be classified as arbitrary: in (33) the informant is not in Haiti at the time of making the statement; in (34), the informant is not – and has never been – in Spain. The source structure for these examples is very common in Brazilian Portuguese (for better comparability, consider the following example, which also involves the verb ‘to speak’):

Interestingly, in Brazilian Portuguese, it is the younger and middle-aged generations – thus, those with whom the interviewed Haitian migrants (recall, aged between 20 and 45 years) are most likely in more contact – who most frequently use generic and arbitrary null subjects (Duarte and Marins 2021, 11–4).^[18]

Future research will need to determine whether this phenomenon has also made its way into the local HC used in Chapecó (HC being a language inclined toward overt subject pronouns that licences some null subjects – cf. DeGraff 1993 – but seems to avoid generic/arbitrary null subjects).

On the other hand, I frequently find the formation of responses to yes/no questions by repeating the main verb of the interrogative sentence, a feature characteristic of Portuguese (Santos 2009) but not traditionally of French and Spanish (example (38) deviates even further from the canonical pattern, as the speaker spontaneously repeats the verb without altering its ending to the 1SG morpheme -o):

Generally speaking, the presence of Portuguese elements in Spanish is particularly intense. In this regard, I highlight another recurrent syntactic phenomenon in my data: the existential use of the possessive verb tener, following the pattern of Brazilian Port. possessive/existential verb ter. This type of linguistic change is relatively common worldwide. However, within Spanish dialectology, it seems to be activated primarily in contexts of linguistic contact with languages that constructionalize existence and predicative possession together (this has been demonstrated before for the case of Misiones Spanish, specifically due to language contact with Brazilian Portuguese: Cerno et al. 2021):

In fact, the insertion of Portuguese elements into Spanish discourse is consistent, and it might even appear that transitioning from one language to another is a manifestation of congruent lexicalization (Muysken 2000). However, this notion is complicated by the fact that speakers may deliberately employ certain linguistic features to differentiate themselves from Portuguese and signal their ability to speak Spanish (using yo instead of eu, hablar instead of falar, [we] instead of [ɔ], etc.). Hence, it can be argued that some of these discourses in Spanish utilize a Portuguese base interspersed with distinct Spanish markers.