Llanito: Neither language nor emerging variety

Full Text

1. Introduction

Since the 1960s and 1970s, studies on code-switching have increasingly occupied a prominent place within contrastive linguistics. Although a substantial number have focused on contact between English and Spanish in the United States, there is also no shortage of works addressing the mechanics of code-switching in societies where languages other than English and Spanish are intermingled. One of the areas that has attracted particular scholarly attention is Gibraltar, since, in addition to uniting English and Spanish, the literature has repeatedly asserted that the combination of the two languages has given rise to a new variety, known as Llanito, in homage to the popular demonym used to refer to Gibraltarians (Levey 2008).

Analyses of Llanito have steadily accumulated over the past few decades. From Moyer’s (1992) thesis to more recent articles — which approach Gibraltarian linguistic reality from discursive, pragmatic, and identity-based perspectives, almost all authors have argued that Llanito constitutes a linguistic variety (Lipski 1986, Levey 2008, among others). However, none of those defending this claim has systematically codified the linguistic structure of such a variety, nor provided a set of phonetic, grammatical, or lexical features by which it may be characterised. What seems to exist is a repertoire of words considered specific of such a region, together with the repeated assertion that Llanito alternates between Spanish and English, is restricted to informal contexts, and, in recent times, is evidently in decline.

Despite the assumption that Llanito represents a linguistic variety, the aim of this article is to argue and demonstrate the opposite. In other words, the so-called Llanito is nothing more than the label applied to the process of code-switching attested on the Rock, which has not resulted in the emergence of a new variety. To this end, the following section provides a brief overview of previous research on Gibraltar and Llanito, before outlining the theoretical framework, corpus, and methodology of the present study. I then present the data from my research, followed by an analysis and discussion which support my hypothesis that Llanito does not, in fact, constitute a variety. Finally, I conclude with a summary of my findings and a bibliography.

2. Gibraltar and Llanito

The enclave of Gibraltar, situated on a small peninsula to the east of the province of Cádiz (Spain), in the Bay of Algeciras, has been a British Overseas Territory since the Treaty of Utrecht (1713), following the War of the Spanish Succession. From that time onwards—and despite various military and diplomatic attempts to recover it—English rose to the status of the prestigious language, relegating Spanish to a diglossic position, associated with informal, colloquial, oral contexts, or with a social class distinct from that of the administrative, military, and social elite of the colony. Nonetheless, this assertion requires nuance or, at the very least, appropriate historical contextualisation.

According to Moyer (1992), the establishment of English as the official and administrative language remained precisely that: it did not entail the widespread proficiency and usage of English among the Gibraltarian population, who continued to employ Spanish as their vehicular language, in contrast to the elite, who adopted English as a marker of dominant social class. Population exchanges across the frontier were continuous, such that—aside from a small group of Britons arriving from Great Britain or descending from the ruling classes (army, government, civil service)—the majority of the population were, above all, Spanish-speaking.

The two centuries following the transfer of Gibraltar into the British were homogeneous in this respect, yet the situation began to shift after the Second World War. The post-war period, extending to the present day, established compulsory education well into adolescence, delivered in the official language and promoted as a pathway to social mobility. Every Gibraltarian, regardless of status, received their academic instruction in English as the single option available.

Moyer (1992) and Kramer (1986) identify this turning point as the beginning of Spanish decline on the Rock and the true point of departure for Llanito as a linguistic reality. The latter half of the twentieth century was characterised by a population predominantly bilingual in English and Spanish, with English acquired through schooling, except among those classes for which it had always been a native tongue. Indeed, this stratum remains monolingual in English, whereas the rest of the population command both languages to a comparable degree—although this balance has gradually decreased over the years in favour of an increasingly monolingual English-speaking population, or one with only passive competence in Spanish (Lipski 1986, Kellermann 2001, Levey 2008, Feijóo Rodríguez 2015). The closure of the border also contributed to the decline of Spanish, as argued by Moyer (1992) and Mariscal Ríos (2014), since it further reinforced the rise of English as a political response to that event.

Kramer (1986) contends that demographic cycles and educational policies have been decisive in determining the predominance of one language over another. Thus, in his view, British control of Gibraltar did not entail the triumph of English over Spanish, which persisted as the majority language despite the influx of populations from various parts of the Mediterranean. With regard to social perception, Moyer (1992) rejects the notion that Spanish was under diglossia, maintaining instead that both Spanish and English could occur in numerous social functions of a similar nature. English was restricted solely to the administrative sphere, while even the mixture of English and Spanish was positively evaluated as an expression of local identity. However, Moyer did find class-based distinctions: the lower the social status, the greater the likelihood of using Spanish or engaging in code-switching. Register also played a role, with Spanish more likely to appear in informal contexts, while code-switching seemed to enjoy a form of covert prestige.

The demographic and linguistic-policy shifts outlined above underpin the arguments of Goria (2021) and Rodríguez García (2024), who claim that Gibraltar moved from a system of diglossia—lasting until the Second World War—to one of dilalia (a type of weak diglossia), following the frameworks of Auer (2005) and Berruto (2005). Thus, during the period from the British conquest of the Rock until the mid-twentieth century, Gibraltarian society developed a diglossic dynamic in which British English functioned as the prestigious variety and Spanish as the informal variety. Yet English was confined largely to extremely formal and administrative domains, and to a very small elite of Britons from Britain. At present, according to these authors, Gibraltar exhibits a situation of dilalia, whereby British English remains the most prestigious variety, but Gibraltarian English and Spanish occupy an intermediate position, beneath which Llanito emerges. The main distinction between dilalia and diglossia lies in the difficulty, in the former, of determining which contexts favour each option or of establishing a clear functional and structural separation between the most prestigious variety and the others. In the absence of certainty regarding the pragmatic constraints when selecting a variety, code-switching emerges naturally, fostered by the looseness of rules, the interlocutor, the topic, or the degree of communicative informality.

The situation described by Moyer (1992) has led scholars to conclude that Gibraltar possesses a new linguistic variety known as Llanito: a manifestation of Spanish–English bilingualism that also functions as an identity marker for those who take pride in local culture and its linguistic diversity, as is reflected in the various types of press publications that highlight this distinctive feature (García Caba 2022). Nevertheless, the growing hegemony of English has prompted the emergence of associations advocating for the preservation of Llanito, to the extent that Gibraltareños por una Sociedad Multilingüe has even proposed a specific orthography. These groups also draw attention to the decline of Spanish.

3. Theoretical framework

The emergence of a new linguistic variety is a complex process, resulting from several possible scenarios (e.g. Kerswill 2010, Proshina 2026), among which are pidginisation, creolisation, and koineisation. Code-switching situations, however, fall outside this contingency, although they may occur at certain preliminary stages prior to the establishment of new varieties. In what follows, I shall briefly describe the first three, before focusing primarily on the dynamics of code-switching in order to understand what is taking place in Gibraltar.

3.1. Pidgin, creole and koiné

The emergence of new languages or varieties generally derives from three distinct linguistic processes: pidginisation, creolisation, and koineisation. In the first case, it deals with a language created ex novo for fulfilling the communicative needs of speakers of mutually unintelligible languages when interacting. Creoles, by contrast, may be explained by two factors: either they are the offspring of pidgin speakers, who stabilise the enormous variability inherent in pidgin, or they are varieties in which a hegemonic colonial language provides the lexical base, while the local language supplies the grammatical and phonetic strategies. Whereas pidgins do not contain established linguistic rules—displaying, as emphasised, high variability according to their speakers—creoles do exhibit systematicity and linguistic regularity (Matras 2009, Holm 2010).

A different process is koineisation, which results from the contact of speakers of mutually intelligible varieties who ultimately generate a new variety, again involving simplification. This process may also stem from migratory dynamics, but in either case Kerswill (2010) argues that three stages are required for its completion: first, a phase of great variability; second, a stage dominated by the following generation, which begins to systematise some usages over others; and third, a subsequent generation, which generalises certain usages, eliminates others, and thus brings to an end the high variability observed in earlier stages.

In the scenarios outlined, the emergence of a new variety always entails—except in the case of pidgin—a stage of extreme variability, chaos, and arbitrariness in the selection of linguistic forms over others, even within the same speaker. This circumstance is also observable in code-switching; hence, in what follows, I shall examine what distinguishes the three processes described from the phenomenon of code-switching.

3.2. Code-switching

Code-switching refers to the alternation, within a single utterance, of at least two languages (this is the definition I will follow throughout). At first sight, this phenomenon arises in a speaker whose command of those languages may be considered bilingual; however, the development of code-switching within an individual or a society can be understood as encompassing four phases, which are not necessarily chronological. In other words, code-switching may manifest itself in four distinct forms: insertion, congruent lexicalisation, alternation, and back-flagging (Muysken 2020).

The first involves the occasional inclusion of lexical items from one language within an utterance produced in the other. The subsequent two forms entail the exchange of more than one juxtaposed constituent between the languages involved across the sentence. Back-flagging, by contrast, is restricted to the incorporation of discourse markers, interjections, or connectors from one language into spontaneous speech in the other. Muysken (2013) has systematised the functioning of code-switching from a theoretical perspective, proposing a series of rules or tendencies to help determine which phase of the phenomenon is at play, or to distinguish between code-switching and simple borrowing from one language into another. This author also accounts for these phases in relation to a series of social parameters, which I summarise in Table 1.

Table 1. Phases of code-switching (Muysken 2013) 

Based on Table 1, it is possible to observe that, depending on the linguistic, historical, or social characteristics of the society in which two languages are used simultaneously, a different outcome arises in the dynamics of code-switching.

The case of Gibraltar refers to that of a bilingual territory. Nonetheless, as Romaine (2010) points out, bilingual societies range from those that develop relatively suddenly, through those that are dynamically bilingual until one language eventually disappears, to those that may be considered stable. In the latter, the coexistence of two languages can result in a situation where each is assigned different functions within society. Thus, one may encounter a diglossic arrangement, in which one language is considered prestigious and the other less so. In other cases, where bilingualism is subject to sociolinguistic factors, language choice may depend upon the type of interlocutor, the age, or the gender of the speaker.

Unstable bilingual societies are those that, having had a widely used language, transition to a different one that subsequently becomes the new dominant language. During such a transformation, there is an intermediate phase in which speakers are bilingual in both, but, as the model predicts, this period does not extend over time as it does in stable societies; it simply emerges as a transitory stage in the shift from one dominant language to another. Questions surrounding this process have centred on whether there is always an interval of code-switching, which groups accelerate the shift from one language to another, whether it entails the reduction of structures in the receding language, and to what extent the changes are socially evaluated within that community.

This trajectory may culminate in the disappearance and eventual death of the receding language in favour of the other. This last circumstance—gradual language death—entails a period in which speakers are bilingual in the two competing languages. However, according to Romaine (2010), this bilingual phase is always temporary and constitutes a preliminary stage leading to the triumph of a new monolingualism, represented by the language that has prevailed.

4. Data and methodology

The data obtained for this research are the result of fieldwork conducted in Gibraltar between November 2023 and May 2024. Oral and spontaneous extracts were collected by means of three combined methods: semi-structured interviews, free conversation, and the reception of spontaneous audio messages from WhatsApp, thereby ensuring the informants’ spontaneity. Altogether, I compiled 22 recordings, amounting to twelve hours, from three different generations, both sexes and varying educational levels.

All the recordings were transcribed and, in order to preserve the anonymity of the informants, the examples provided are labelled with an alphanumeric code designed to reveal the speaker’s complete sociolinguistic profile. The system uses the following categories: M (male), F (female), –30 (young), 30–60 (middle-aged), +60 (elderly), L (low educational background) and H (high educational background). The concatenation of these labels makes it possible to identify the profile of the speaker responsible for a given utterance. For instance, if an example is tagged M30–60H, this indicates that the utterance was produced by a middle-aged man with a high educational background.

5. Results of the analysis

The fieldwork data reveal divergent phases in the linguistic mixture that takes place between Spanish and English. Three clearly defined stages may be distinguished, depending on the speaker’s generation, each of which reflects a different type of code-switching: insertion, alternation together with congruent lexicalisation, and back-flagging. That segment of the population born prior to the closure of the frontier resorts to insertion, whereas those born during Gibraltar’s period of isolation tend to favour alternation and congruent lexicalisation. Individuals born in the 1990s and, mainly, in the 2000s, display a preference for back-flagging.

Furthermore, the generation characterised by insertion has Spanish as their L1, while their acquisition of English corresponds to the learning of an L2. By contrast, their descendants possess both languages as L1 in most cases, while those who favour back-flagging have reversed the situation of previous decades, with English as their L1 and Spanish emerging as an L2, or as a language with only passive or very limited competence. Naturally, the sociolinguistic factor is also relevant: the higher the social class, the lower the probability of retaining Spanish as part of the linguistic repertoire across any of the generations.

(1) Y como habéis visto, esto Gibraltar, es lo que llamamos a melting pot, ¿no? Gente de todas partes que han venido, oleadas de gente. (M60L)

[And as you have seen, this is Gibraltar, what we call a melting pot, no? People from everywhere, many people.]

(2) Y viene con un acento muy, muy, muy pijo. Posh, very posh. Y entonces, en una ocasión dijo “yo soy del pueblo, aunque hablo así en inglés, pero soy llanito, soy del pueblo”. El pobre, quería identificarse con el pueblo, pero con ese acento no… (M60L)

[And he comes with a very, very, very posh accent. Posh, very posh. And then, on one occasione, he said “I am from the people, alhough I speak English like this, but I am Llanito, from the people”. Poor man, he wanted to identify himself with the people, but with that accent…]

(3) Hay unas listas de espera de… cerca de 800 y pico de personas. Entonces, pues… Han hecho este sistema, que le llamo fifty-fifty. (M60L)

[Here there is a waiting list… nearly 800 people or so. Then, they have created tis system, I call it fifty-fifty]

(4) Mira, ¿tú tienes el número de teléfono de [NP], by any chance? H60N

[Hi, look, do you have the phone number of [proper name], by any chance?]

(5) Estaban todas muy, muy laid out. (M60L)

[They were all very, very laid out.]

(6) Son dos, son pictures. Son las primeras que hizo [NP] en color. (M60L)

[These are two, two pictures. They are the first ones [proper name] took in colour.]

(7) Está con una costilla que la tiene fractured de un golpe que se dio, y está con la pierna también, que está fastidiada con la pierna. Eh… Tú… si tú puedes this week, esta semana creo que no va a ser posible, a ver si podemos… yo le pregunto luego a ella. (F60H)

[She has a rib, a fractured rib because she beat herself, and her leg is also working badly. If you can this week, this week I think it is not possible, let’s see, I’ll ask her later.]

Examples (1–7) demonstrate that the base language and L1 is Spanish, into which English words are occasionally inserted. These shifts are not conditioned by grammatical category or by the absence of a corresponding lexical item in one of the languages. Rather, they are mere oscillations prompted by the speaker’s linguistic competence in the two languages coexisting in their repertoire. Since Spanish is the dominant language and English was either acquired at a very late stage or retained only as a passive variety, the insertions are generally readjusted to the phonological system of the L1. Although this cannot be verified in writing, all English usages produced by these informants were articulated with the characteristic pronunciation of Western Andalusian. Thus, for example, a melting pot was pronounced with rhotacism (merting) and with aspiration and loss of final consonants /t/ and /g/: a mertin po.

These same informants also display the strongest Hispanophone features when speaking English, since they acquired it as an L2. However, it is possible to identify certain borrowings from English into Spanish, which are not to be confused with the insertions occurring during code-switching. Such borrowings are typically lexical in nature, for instance panqueque instead of tortita (from pancake) or escarte instead of rodapié (from skirting). By contrast, no influence of English is observed on the morphosyntax of the Spanish they speak.

The generation born between the 1960s and 1980s, on the contrary, behaves in line with the rules established for alternation in code-switching. Switching from one language to the other is constant, not limited to lexical insertions, and, in addition to observing semantic shifts in Spanish through contact with English, I have also detected modifications in syntax, attributable to the process of congruent lexicalisation described by Muysken (2013).

(8) ¿Qué pasa [NP]? Soy el [NP]. ¿Cómo estás llevando el confinement, brother? Yo lo llevo really bad, man, se me está haciendo very heavy. Esta ha sío mi muy primera vez que me he tenío que queá tanto tiempo at home. Yo no quiero salí pa na porque estoy very scared, pero esta mañana he hecho wake up más temprano de lo que suele ser y estaba supuesto de ir a Morrison’s pa hacer (M30-60H)

[What’s up, [proper name]? It’s [proper name]. How are you coping with the confinement, brother? I’m dealing with it really bad, man, it’s becoming very heavy. This has been my very first time I have had to be so long at home. I don’t want to go out at all, because I’m very scared, but this morning I woke up earlier than usual and I was supposed to go to Morrison’s to do some shopping.]

(9) No sé si vuelvo three o no, [NP], pero el redline era… Border a… Poland. Tos los países alrededor de Ukraine eran borderline red. Si algo pasaba across the borderline red o on the borderline red, the rest of the world would have to join in. (M30-60H)

[I don’t know if I’m flying back three or not, [proper name], but the redline was… Border a… Poland. All the countries around Ukraine were borderline red. If something happened across the borderline red or on the borderline red, the rest of the world would have to join in.]

(10) What a beautiful day in Gibraltar. Gracias a Dios por esta playa que tenemos tan bonita. And I’m looking out here, you see, this is where, this is where the proposed marina will start, you see, from there, donde está el massive development, and then it goes all the way around una manga, ahí, right?, y ahí va a estar el new hotel […]. Venga, Gibraltar, join Catalan Bay y los caleteños who we love. No to the marina (F30-60L0

[What a beautiful day in Gibraltar. Thanks God this beautiful beach we have. And I’m looking out here, you see, this is where, this is where the proposed marina will start, you see, from there, where there is the massive development, and then it goes all the way around a land extension there, right?, and there will be the new hotel […]. So, Gibraltar, join Catalan Bay and caleteños who we love. No to the marina.]

The three excerpts above, drawn from three different informants aged between 40 and 55, perfectly illustrate the two stages of code-switching that commonly overlap: congruent lexicalisation and alternation. It is pertinent to recall that the former entails changes in the syntax of one language under the influence of the other, or the morphological hybridisation of the two languages in contact. This is precisely what occurs in (8), where the speaker produces an English syntactic construction while speaking Spanish: estaba supuesto de ir a Morrison’s (literally ‘I was supposed to go to Morrison’s’). Instead of opting for the impersonal construction with reflexive se (se suponía que iba a ir a Morrison’s: lit. ‘it was supposed that I was going to Morrison’s’), the speaker has adopted the Anglo-Saxon passive form, which is ungrammatical in any Spanish variety. Similarly, the same informant reproduces the analytic superlative modelled on English, producing mi muy primera vez (literally ‘my very first time’) instead of the normative strategy by means of the suffix –ísimo attached to the lexeme.

The remainder of the utterances (from this speaker as well as from the other two) abound in alternation, with Spanish as the dominant language—except in the last case, where English predominates. Switching may occur at any point, whether at the beginning of a sentence (main or subordinate), within a phrase; or even inside a phrase between tightly coalesced elements.

This section of the population, from which the passages are taken, is bilingual in Spanish and English, though it is more precise to state that the vast majority have Spanish as their L1, acquired at home from birth, and English as their L2, learned spontaneously in the street and once integrated into the educational system. By contrast, the younger generation displays a different type of code-switching, more in keeping with what is termed back-flagging, embodying precisely the opposite behaviour of the older speakers. In this group, the hegemonic language is English and, from time to time, elements of Spanish are added. Unlike insertion, however, the resources introduced tend to be collocations, discourse markers, conjunctions, and similar items.

(11) This is porridge. This is a well-known brand that actually sells very good porridge,

[This is porridge. This is a well-known brand that actually sells very good porridge, ok?]

(12) Por supuesto, [NP]. You’re one step ahead. (F-30H)

[Of course, [proper name]. You’re one step ahead.]

(13) I’m sick, no estoy muy Catholic (F-30H)

[I’m sick, I’m not very Catholic today [meaning to be ill]

(14) We loved a lot about it, pero I didn’t mean it. (F-30H)

[We loved a lot about it, but I didn’t mean it.]

Examples (11–14) attest the stage known as back-flagging, since the informants base their discourse on English, into which they occasionally introduce conjunctions, as in (14), discourse markers and connectors, as in (11–12), and even collocations—although in (13) part of the collocations is already in English, with the adjective Catholic pronounced in Received Pronunciation (RP).

This younger generation, however, is to some extent divisible into two. Whereas those born in the 1990s possess only a passive knowledge of Spanish and rarely produce it (unless they belong to the lower social classes), those born in the present century have not acquired Spanish at all and, according to the testimony of teachers and compatriots, may even ask at secondary school how to say words such as car in Spanish.

The occurrences discussed above clearly bear witness to the trend towards which Spanish appears to be heading, while also revealing the arbitrariness involved in language choice at each instance of code-switching. In the following section, I shall argue why Llanito cannot be regarded as a linguistic variety and the extent to which phenomena attested in other geographical contexts are replicated in Gibraltar.

6. Discussion

The excerpts presented in the previous section are further evidence of code-switching, as exemplified by other scholars in their respective studies on the Gibraltarian linguistic situation. When the utterances recorded by Moyer (1992) are compared to those analysed by other academics, as well as with the data supplied in this article, it becomes clear that the so-called Llanito can only be classified as code-switching, and not even as an incipient creole or independent variety.

Firstly, it is impossible to determine the syntactic structures, morphology, or lexicon of which it consists. In none of the stages described is it possible to systematise which option corresponds to which reality, or which linguistic configuration is appropriate for conveying a specific grammatical function. In other words, it is impossible to know whether the entity corresponding to the lexeme table, for example, would in Llanito be realised by the Spanish word mesa, the English option, or some hybrid of the two. Nor is it possible to predict whether a noun phrase would be constructed with a Spanish article and an English head noun, or vice versa. The arbitrariness of speaker choice prevents from treating this as a concrete variety with its own grammatical rules.

Likewise, each informant displays diverse linguistic behaviour, since the hybridisation typical of congruent lexicalisation—for instance, estar supuesto de (‘to be supposed to’) plus infinitive—is not shared by all individuals, nor is it systematic. Even local associations that advocate for the promotion of Llanito cannot agree on how it should be represented orthographically. As a matter of fact, the lexicographical works that exist are devoted to cataloguing a number of fossilised words that are supposed to be only used on the Rock, with no equivalent in English or Spanish. Such is the case with meblis (‘marbles’), panqueques (‘pancakes’), and similar examples, which merely indicate borrowings rather than code-switching. Yet those same lexicographical sources cannot determine, for instance, which term corresponds to the entity table in Llanito.

The high degree of variability observed could be related to one of the stages in the creation of a new variety, as discussed in earlier sections. However, this possibility cannot in itself be classified as a variety; rather, it represents a preliminary stage before a new variety becomes established. Historical developments interrupted the expected outcome of such variability, since successive generations gradually shifted towards monolingualism favouring English, which became hegemonic and, at times, exclusive. The process attested in Gibraltar is typical of a contact zone, where one language enjoys greater prestige than another and thereby enters into competition.

Table 1, reproduced from Muysken (2013), is fully borne out, since each phase of code-switching corresponds to a sociolinguistic factor that mirrors Gibraltarian reality. Thus, in the stage of insertion, it is possible to observe that, apart from the local English-speaking elite, the wider population had Spanish as their L1, with English competence being passive, limited, or functioning as an L2, within a (post)colonial framework. In the subsequent two stages, there are long contact, relaxed norms, close-knit social networks, and bilingual competence, sharpening the tension between the vehicular language (Spanish) and the minority but prestigious one (English). In the later stage of back-flagging, a reversal in the competing languages is witnessed: English becomes the L1, while Spanish shifts to L2 status and is used by most speakers only passively or in a very limited way. At this point, it deals wih a change fostered by the second or third generation, which leave unstable bilingualism in favour of monolingualism (in this case, English).

In short, the most characteristic phase of code-switching in Gibraltar—the one involving alternation and congruent lexicalisation, and which has underpinned claims that Llanito constitutes a new linguistic variety born of English–Spanish mixing, in the manner of Spanglish—is in fact simply a stage of great variability or volatility in speakers’ utterances, as a consequence of their high competence in both languages. It could indeed represent a prototypical stage in the development of a koine, for the generation engaging in such apparently chaotic and arbitrary alternations typically precedes that which consolidates certain usages, dismisses others, and thereby stabilises a new variety, often characterised by paradigmatic simplification. It is this latter case that does systematise lexicon, morphology, pronunciation, and syntactic structures—so only once evolution reaches this point can we speak of a new or distinct variety.

In Gibraltar, however, historical events disrupted this trajectory and instead aligned with the classic development of a code-switching situation. Here there is an unstable bilingual societiy, characterised by a period of linguistic contact and bilingualism of limited duration, before tipping into monolingualism, invariably in the language of prestige. Unlike stable bilingual societies, where both languages coexist—though often each becomes specialised for different domains—unstable societies eventually abandon bilingualism altogether after several generations.

This distinction is crucial, transcending the identity-related feelings and attitudes that speakers themselves may express about their bilingualism, since the (un)conscious preference for monolingualism responds to the prestige of the selected language, or to the social mobility associated with it. In this sense, it must be emphasised that, in the specific case of Gibraltar, political disputes were not the decisive factor undermining the survival of Spanish, although they may have accompanied the process. Rather, the key factor was the sweeping social transformation brought about in Western countries by the end of the Second World War.

On the one hand, had the political climate been the trigger for the increasingly homogeneous rejection of Spanish, this would have occurred much earlier in the Rock’s history, since the conflict dates back to the cession of the territory to the British Crown. Furthermore, had Franco’s dictatorship and its policy of isolating the colony through the closure of the border generated a negative attitude towards speaking Spanish, we would expect to find evidence of such rejection during that period. Yet the generations born throughout the years of friction between Gibraltar and the dictatorship are highly competent in both languages, with Spanish still serving in many cases as the L1. In fact, it is only from the 1990s onwards that a marked decline in Spanish competence becomes apparent—a trend that varies with the speaker’s social class, but which, broadly speaking, is attested across the region.

I therefore insist that the true reason—or at least the central motive—for the linguistic paradigm shift must be attributed to social mobility (one of the reasons, together with the identification of lingua franca, that lies behind the expanding usage of English in Russia, Korea, Japan or China (see Kirkpatrick 2020, Lovtsevich & Sokolov 2020, Proshina & Nelson 2020 among others). The introduction of compulsory education (exclusively in English) and the possibility of economic and professional advancement through it constituted a profound social transformation, since mastery of the prestigious language enabled access to universities in Britain and to highly qualified positions. Testimonies from my informants support this claim, as corroborated by the following excerpts.

(15) H2: Sí, sí, sí. Hay un dicho que dice “accent defines the man”. (M60L)

[Yes, yes, yes. There is a saying which states “accent defines the man”.]

H1: ¿Es un dicho de inglés?

[Is that an English saying?]

H2: Sí, sí. Abres la boca y ya te ponen, te definen. (M60L)

[Yes, yes, you open your mouth and they define you.]

 (16) M1: Y en el patio, si hablabais en español ¿os reñían?

[At the courtyard, if you spoke in Spanish, did they tell you off]

H2: Nos regañaban, sí, nos reñían, sí. Porque lo que querían era que los niños aprendieran inglés, el español estaba… era secundario, ¿no? (M60H)

[They told us off, they scolded us, yes. Because they wanted kids to learn English, Spanish was secondary.]

(17) H1: No se ha hecho a… a cosa hecha, ha evolucionado así [...] Pero, siempre se hablaba el español en casa […] Y… después de la II Guerra mundial los que tenían carrera, los que habían ido a la universidad a estudiar, eran los hijos de papá, los hijos de los que estaban bien adinerados. […] A partir de los años ochenta […] todos los estudiantes que… aprobaban sus estudios de bachillerato y les ofrecían una plaza para estudiar cualquier asignatura, cualquier licenciatura en las universidades de Inglaterra y del norte de Irlanda con una beca. […] Y, entonces, se ha hablado más inglés. (M60H)

[It wasn’t on purpose, it has evolved like this [...] But, at home, you always spoke in Spanish […] And after WWII, those with a degree, with a university degree, were rich […] From the 1980s […] all the students that passed high school were offered a fellowship to study any degree in England and northern Ireland. […] And then, from that moment English has been spoken more and more.]

(18) H2: Ah, vale, vale, vale. Entonces, entonces, [NP], la gente como tu, como tu… tus nietas, eh… cuando van a España, eh, simplemente tienen una competencia pasiva, ¿no? Entienden, pero, salvo en algún caso, no pueden hablarlo como lo hablas tú o lo hablo yo.

[Ah, ok, ok, ok, then, then, [proper name], people like you, like you, your granddaughters, when they go to Spain, they simply have a passive competence, don’t they? They understand, but they cannot speak it as you can or I can.]’

The well-known Spanglish, to which the literature has made continuous reference by coining a term to designate the linguistic hybridisations characteristic of Hispanics in the United States, is no more a linguistic variety than Llanito—that is to say, neither can be considered one. In both cases, what we encounter is a process of code-switching which, depending on the generation, is situated at one stage or another. This is corroborated by studies carried out in various parts of the United States with a strong Hispanic presence (see for instance Solzhenitsyna (2014)), since the degree of hybridisation—and therefore of so-called Spanglish—varies based upon the same parameters. For instance, if we turn to the linguistic reality of Miami, we observe that the base language for recent arrivals or for groups with a low level of English proficiency is Spanish, into which English words are sporadically inserted. However, such vocabulary does not replace the equivalent Spanish words, but rather alternates with them. Moreover, it may affect any type of lexical-content word (De Jongh 1990).

(19) Las conoci a ellas ahi en el bildin

‘[I met them there at the building.]

(20) Estoy en estanbai

[I’m on standby.]

Examples (17–18), taken from De Jongh (1990), replicate what has been observed in Gibraltar: insertions phonetically adapted to Spanish, which serves as the L1. The descendants of these generations, fully educated within the United States system, are bilingual in both languages, with such a high level of competence that each may be regarded as native. This group produces alternation and congruent lexicalisation, shifting between Spanish and English across entire sentences, without the phonetic influence of Spanish on their English (a feature that could still be found among those who engaged solely in insertion) (De Jongh 1983, 1990).

(21) El piloto, I don’t know if you got any guy que es, pues gran piloto. Tú sabes, que ... who would hacer la cosa así, con un piloto. Pero la cosa es lo siguiente: El vuelo sale desde el punto que uno quiere que salga. It comes con un flight plan, entra-pam, pam. Aterriza en [...], pasa Customs, y de ahí sale pa’ donde uno (unintelligible). Vaya, a typical flight

‘The pilot, I don’t know if you got any guy who is, well a good pilot. You know that... who would do this so, with a pilot. But the thing is this: the flight departs from the place one wants it to depart. It comes with a flight plan, enters-pam, pam. It lands on [...], passes Customs, and then you go where (unintelligible). So, a typical flight’

(22)  El avión está lisiado

‘The plane is leased’

Examples (19–20), again taken from De Jongh (1990), confirm code-switching in a speaker highly proficient in both English and Spanish, together with congruent lexicalisation manifested in the morphological adaptation of lease (‘to rent’) to a Spanish structure.

Later generations, though possessing Spanish at a more passive level than their predecessors, yet still demonstrating high competence, tend to opt for monolingual discourse in English, even in contexts where Spanish is understood and prompted. This group makes only minimal use of this language, limiting themselves to collocations or highly sporadic lexical items (Zurer-Pearson & McGee 1993, Portes & Schauffler 1996). However, the linguistic reality of Miami (and of other areas of the United States) displays an important peculiarity that distinguishes it from Gibraltar: the Floridian city continuously receives new Hispanic migration, so that the intergenerational development from insertion to back-flagging—with Spanish as L1 until its reversal with English as L1—is repeated with each new wave of Spanish-speaking arrivals settling in the United States (Carter & Lynch 2015). By contrast, Gibraltar does not receive continuous migration, since Spanish workers who cross the border do so daily and immediately return to Spain. The development of code-switching in Gibraltar has thus unfolded over several decades within the native population, culminating in a form of unstable bilingualism that has given way to monolingualism and the near extinction of Spanish. In Miami, however, each new arrival of Hispanic migrants initiates the evolution of code-switching (both linguistically and socially), causing the city never fully become a monolingual space nor crystallise into a separate variety.

I stress that neither Spanglish nor Llanito constitute new varieties generated from the combination of Spanish and English. Were this the case, they would have developed a specific grammatical structure, selected certain lexical items at the expense of others across the entire lexicon, and privileged some phonetic usages over others. Instead, in both cases, the alternation of words remains arbitrary; the grammatical structures of each language are respected—except during the stage of congruent lexicalisation (and even then only exceptionally and unsystematically); the pronunciation of each language is maintained (apart from the insertions characteristic of the first stage of code-switching); and, eventually, a tendency towards monolingualism emerges. In short, the ideal of Llanito and Spanglish often referred to by the bibliography corresponds to that attested in the generation with high bilingual competence and, therefore, to the stage of alternation and congruent lexicalisation. This is a short period, witnessed only in that generation, which disappears as the following one reaches adulthood, has children, and monolingualism becomes established—always, of course, favouring the prestigious language of the country of residence, namely English, both in the United States and in Gibraltar.

My fieldwork observations corroborate the argument put forth in these pages. On more than one occasion during spontaneous interactions, I have witnessed awkward moments in which a grandfather spoke to his grandson in Spanish and the child replied exclusively in English, without uttering a single word in the other language. When pressed to respond in Spanish, the grandson vehemently refused to comply with his relative’s demand. Likewise, when speaking with parents concerned about the decline of Spanish and of what they interpret as Llanito identity, insofar as it permitted the mixing of both languages, they reported that their children—despite visiting the other side of the frontier for the day, eating there, having a second home, or knowing that Spanish is one of the world’s most widely spoken languages—remain reluctant to learn it. Moreover, these parents asserted that, on social media, their children and grandchildren immediately dismiss videos that appear or are suggested in Spanish, choosing instead to focus exclusively on those in English.

The evidence presented suggests that Spanish is on the path to extinction in Gibraltar. Nonetheless, the findings have also served to displace clichés about the causes of its decline, for it is the language of prestige that has ultimately prevailed, following the same trajectory already documented in comparable contexts.

7. Conclusions

The so-called Llanito, so often regarded as Gibraltar’s own linguistic variety born of contact between Spanish and English, is in reality a case of code-switching within an unstable bilingual society transitioning towards English monolingualism, given that English represents both the prestigious language and the vehicle of social mobility. Although the considerable variability displayed in one of the most iconic phases of code-switching resembles the volatility observed in processes of koineisation or creolisation, the fact that it is impossible to establish the linguistic norms by which the so-called Llanito might be governed leads to the conclusion that it cannot be considered a linguistic variety. Furthermore, historical developments on the Rock have prevented the formation of a new variety, since the generations following those marked by such variability have progressively abandoned their parents’ linguistic practices, embracing English monolingualism with increasing determination.

As with Spanglish, linguistic hybridisation represents merely a preliminary phase before the triumph of one of the contact languages at the expense of the gradual disappearance of the other. This is precisely what has occurred among Spanish speakers in the United States, and what has also taken place in Gibraltar since the Second World War, when the introduction of compulsory education in the official language provided access to social mobility through English. Although instances of different tendencies can still be documented in the British colony, these correspond to divergent phases of code-switching, as different generations may still be observed within this process. Nonetheless, even if social class nuances are taken into account, the dominant trend on the Rock is towards English monolingualism, notwithstanding the identity-related sentiments speakers may express about the code-switching (but not linguistic variety) they refer to as Llanito.

The results contribute to disentangling the complexity of Spanglish in its different manifestations and to shedding more light to the study of linguistic variation.

About the authors

Víctor Lara Bermejo

Author for correspondence.
Email: victor.lara@uca.es
ORCID iD: 0000-0002-1068-8553

professor at Cadiz Univrity, Cadiz, Spain, and researcher specialised in Spanish dialectology, as well as in language contact between Portuguese and Spanish

References

Supplementary files