Teresa Maria Włosowicz University of Social Sciences in Łódź Faculty of Applied Sciences in Cracow Translanguaging as the Mobilisation of Linguistic Resources by Learners of Spanish as a Third or Additional Language A b s t r a c t The present study investigates the comprehension and production of Spanish as a third or additional language (De Angelis’s (2007) term), paying special attention to the use of code-switching and translanguaging. Following Lewis, Jones, and Baker (2012, p. 655), it is assumed that translanguaging involves the mobilisation of all of learners’ linguistic resources “to maximise understanding and achievement,” so the learners’ use of languages other than Spanish (especially English, but also e.g. French, Italian, etc.) in the tasks could be assumed to be an example of translanguaging too. Simultaneously, the use of words from languages other than Spanish for lack of a Spanish word could be argued to be more precisely classified as code-switching. Multilingual repertoires are highly complex and, according to Otheguy, García, and Reid (2018), multilingual competence is unitary rather than divided into several distinct languages, so, in their view, words are selected from a single lexicon. However, as shown by Williams and Hammarberg (1998), the different languages in multilingual reper- toires perform various functions, which gives rise to different types of switches. The study was carried out with English Philology and Romance Philology students studying Spanish as a third or additional language. As the results show, even though the Romance Philology students were generally more skilled at translanguaging, viewed as the use of all their lin- guistic resources, they avoided switches into other Romance languages, probably to mini- mise interference. By contrast, the English Philology students, who had lower proficiency in Spanish, were less capable of using their multilingual resources, including English, to provide the missing words, possibly also due to problems with the comprehension of the Spanish sentences. Keywords: multilingual repertoires, linguistic resources, translanguaging, production strategies Theory and Practice of Second Language Acquisition vol. 6 (1), 2020, pp. 65–106 https://doi.org/10.31261/TAPSLA.7770 https://orcid.org/0000-0001-8767-9332 https://creativecommons.org/licenses/by-sa/4.0/deed.en https://creativecommons.org/licenses/by-sa/4.0/deed.en Teresa Maria Włosowicz66 Introduction The purpose of the present study has been an investigation of the written production and comprehension of Spanish as a third or additional language (De Angelis, 2007, p. 11), taking into consideration the use of code-switching and translanguaging as a reflection of multilingual linguistic and strategic competence. The term ‘third or additional language’ is used here on purpose, as the participants’ language repertoires were varied and Spanish was not necessarily their L3, but could also be their L4 or even L5. However, ac- cording to Williams and Hammarberg (1998, p. 296), an L2 can be defined as a previously learnt foreign language, while the L3 is the language cur- rently being studied; in their view, therefore, a learner can have more than one L2 and more than one L3 at a time. In other words, while Spanish was, chronologically, for example, a student’s L4, in Williams and Hammarberg’s (1998) terminology it could still be regarded as an L3, so the participants’ language repertoires could be supposed to be sufficiently comparable to allow the realisation of the study. Despite the various definitions and approaches to translanguaging (see Section The Phenomenon of Translanguaging), it can generally be assumed to be the mobilisation of a learner’s linguistic resources in their entirety “to maximise understanding and achievement” (Lewis, Jones, & Baker, 2012, p. 655), so the use of English and, possibly, other languages (especial- ly other Romance languages, such as French, Italian or Portuguese) could also be classified as translanguaging. Similarly, while some researchers on translanguaging (e.g., Otheguy, García, & Reid, 2018) regard multilingual competence as unitary, thus rejecting code-switching as switching between different languages (Otheguy et al., 2018, p. 16), the approach followed here is more traditional, regarding code-switching as a type of translanguaging (García, 2009a, p. 140; MacSwan, 2017, p. 191), for example, as a commu- nication strategy applied when the target Spanish word cannot be retrieved. Indeed, this approach seems the most appropriate in the context in which the study was conducted. Unlike bilingual immigrant children, the English and Romance Philology students were studying Spanish as a foreign language in formal university contexts and, as it turned out (see Results and Conclusions), translanguaging was not their natural way of using Spanish, but rather, it had to be explicitly encouraged by the research design. Otherwise, it can be assumed, the avoidance rates would have been higher (in fact, they were relatively high, especially among the English Philology students, see Tables 1 and 2), as foreign language classes generally focus on the target language and do not involve the mobilisation of all linguistic resources, so students are not used to translanguaging. Translanguaging as the Mobilisation of Linguistic Resources… 67 However, on the basis of the students’ use of resources belonging to languages other than Spanish, it is attempted to draw some conclusions re- garding multilingual repertoires, the place of Spanish as a third or additional language in them, and the use of translanguaging as a comprehension (some prompts in English are also used; similarly, learners tend to translate into their L1 to improve and consolidate understanding, Kern, 1994) and produc- tion strategy. The use of translanguaging will be analysed in two different groups: English Philology students, for whom Spanish is just an additional foreign language, unrelated to the other languages they study, and Romance Philology students, for whom Spanish is a more important part of their curriculum and is related to the other Romance languages, which may also strengthen the links between Spanish and the other language(s) they study (cf. Singleton, 2001; Herwig, 2001). Multilingual Language Processing and Use Multilingual Competence In general, multilingual systems (often called repertoires, especially in studies related to the choice of one language or another, or of code-switching, in particular sociolinguistic contexts, cf. García & Otheguy, 2019; Li, 2018) are not sums of clearly delineated and separate language systems, but rather networks within which the different languages constitute interconnected and interdependent subsystems. Following Grosjean (1985, p. 467), who empha- sised that a bilingual was not the sum of two monolinguals, De Angelis and Selinker (2001, p. 45) observe that “a multilingual is neither the sum of three or more monolinguals, nor a bilingual with an additional language,” but rather “a speaker of three or more languages with unique linguistic configurations, often depending on individual history.” Indeed, as Cieślicka (2000) has shown, the links between L1 and L2 words in the bilingual lexicon vary from one speaker to another and depend on such factors as the language learning context, learning strategies, proficiency, etc. In the case of three or more languages, the system is even more complex, as the words of L3, L4, etc. can become attached, for example, to their L2 rather than L1 equivalents, if the L3 and the L2 are typologically closer (Herwig, 2001, p. 117, Singleton, 2001). Certainly, the languages within multilingual systems are interconnected not only at the lexical, but also at the grammatical level, which led Cook (1991, n.p., as cited in Cook, 2016, p. 2) to propose the notion of multi-com- petence as “the compound state of a mind with two grammars.” However, Teresa Maria Włosowicz68 as he later admitted (Cook, 2016, p. 2), this definition may be misleading, as it may suggest that multi-competence refers only to syntax, even though his original definition was based on the Chomskyan idea of grammar as linguistic knowledge in general. Therefore, the current definition of multi- competence postulates that it is “the overall system of a mind or a commu- nity that uses more than one language” (Cook, 2016, p. 3). As he explains, the term “system” is more neutral than “knowledge,” which might be regarded as static, and the definition “does not confine multi-competence to language alone, brings in language use and implies that language is not separate from the rest of the mind” (Cook, 2016, p. 3). Therefore, it can be assumed that, if translanguaging involves a learner’s multi-competence, it combines the use of his or her different languages with extralinguistic knowledge, cultural competence, strategic competence, etc. In general, multilingual (or, as Coste, Moore, & Zarate (1997) call it, “plurilingual”) competence, is varied and cannot really be “balanced,” how- ever advanced the learner is in all his or her languages. Coste, Moore, and Zarate (1997, p. 12) define plurilingual and pluricultural competence as the linguistic and cultural communicative competence possessed by a person who has different levels of proficiency in several languages and different degrees of experience with several cultures, but who is able to manage his or her linguistic and cultural capital. In their view, it is not a juxtaposition of dis- tinct competences, but rather a whole complex and heterogeneous repertoire composed of singular, even partial competences, available to the language user (Coste el al., 1997, p. 12). In this sense, translanguaging can be regarded as a fully legitimate use of one’s multilingual repertoire to communicate the intended meanings. As mentioned above, the languages in a multilingual repertoire are to some extent interconnected. For example, as illustrated by Herwig’s (2001) model of the multilingual mental lexicon, there are links between the words (or, more precisely, between the nodes where different components of lexical knowledge are stored in a distributed way) of the multiple languages at different levels (for example, semantic, phonological, orthographic, etc.) and the strength of those links depends on the similarity between the words. For example, translation equivalents which differ in form may be connected at the semantic level, but cognates can be connected at the semantic, phonological, and orthographic levels. As will be shown in more detail later in this article, the connections between the words of the different languages make possible both their strategic use (for example, code-switching as a communication strategy or an attempt to elicit the target language word) and interference errors. However, as Singleton (2003, p. 168) points out, “the existence of marked formal differences between languages,” such as phonological differences, constitutes an argument against full integration. Similarly, bilingual and multilingual speakers usually keep Translanguaging as the Mobilisation of Linguistic Resources… 69 their languages apart while speaking, and it is even possible that, if one does not expect to hear a particular language, its comprehension may be blocked (Singleton, 2003, p. 168). While connections between words can be observed more easily, for ex- ample, on the basis of lexical associations (e.g., Gabryś-Barker, 2005), at the grammatical level, the languages of a multilingual are not fully separated either. On the one hand, grammar is lexicalised and sentence structure de- pends on the syntactic properties of individual words, especially verbs, more than on general phrase-structure rules, such as S → NP + VP. For example, the difference between the sentences “She intends to eat chocolate tonight” and “She regrets eating chocolate tonight” is due to the different syntactic properties of the verbs “to intend” and “to regret” (Singleton, 2000, p. 17). Therefore, as lexical items belonging to different languages are connected, one might access the properties of a non-target language word. Indeed, as shown by Gibson and Hufeisen (2001), learners do use verbs with inappro- priate prepositions, for instance, “*sich konzentrieren an” (target preposition: “auf”; to concentrate on), “*sich fürchten von” (target preposition: “vor”; to be afraid of), etc. (Gibson & Hufeisen, 2001, p. 185). According to Hall and Ecke’s (2003) Parasitic Model, L2 learners copy into L2 lexical entries the properties of L1 words, which may lead to errors if those properties are dif- ferent, as in the case of “I like X” and “me gusta X,” where the subject is the person or object liked by the speaker (Hall & Ecke, 2003, p. 77). On the other hand, learning a second language also leads to the restructuring of L1 competence to some extent; thus bilingual English-French speakers’ gram- maticality judgements in L1 English were different from those of monolingual English speakers (Cook, 1996, p. 65). In summary, even though multilingual systems consist of several languages, which can be used separately, for example, while talking to a monolingual, they are to some extent interconnected and can therefore interact. In fact, fol- lowing Coste et al. (1997), it can be assumed that, as multilinguals’ language repertoires are available to them as wholes, the use of non-target language words or structures, for example, to convey a meaning for which the learner lacks the target language word, is a natural consequence of this interconnection and thus translanguaging can be regarded as a normal phenomenon in third or additional language acquisition. Multilingual Language Production and Comprehension Since the languages in multilingual repertoires are interconnected and inter- action between them is inevitable, the consultation of more than one language occurs in both comprehension and production. As shown by Grainger and Teresa Maria Włosowicz70 Beauvillain (1987), in visual word recognition, bilingual lexical access does not involve pre-selective search; rather, language-specific orthographic cues point to the activation of words in a particular language (Grainger & Beauvillain, 1987, pp. 314–315). At the phonological level, there is competition between phono- logically similar L1 and L2 words, which supports “the hypothesis of parallel activation of both languages” (Marian & Spivey, 2003, p. 104). As Marian and Spivey (2003, p. 109) conclude, “bilinguals can and do experience competition from both languages and into both languages, although the magnitude of the effect changes under different circumstances.” In comprehension, as pointed out by Green (1993, p. 260), all the lexical items consistent with the input are activated, not only those which belong to the target language, and similarity between L1 and L2 items can delay com- prehension. In production, as Herwig (2001, p. 128) concludes on the basis of the results of her study, “lexical selection in situations of non-accessibility of an item in demand involves both automatic and deliberate consultation of several languages.” In fact, as De Angelis (2005) has shown, the control mechanism is not perfect and a word from a non-target language can be se- lected and regarded by the speaker as a target language word, a phenomenon which De Angelis (2005, pp. 10–11) calls a system shift. The factors which contribute to system shifts are “perception of correctness” and “association of foreignness” (De Angelis, 2005, p. 11). According to De Angelis (2005, p. 11, her emphasis), “[p]erception of correctness refers to multilinguals’ resistance to incorporating L1 linguistic knowledge into interlanguage production when other information is available for them to use.” In other words, it is “learners’ ability to successfully monitor their production and identify what is correct or incorrect target language output” (De Angelis, 2005, p. 11). By contrast, as- sociation of foreignness is the perception of foreign languages as closer to one another, which results in greater acceptance of non-native words, even if they come from a non-target foreign language (De Angelis, 2005, p. 12). An example might be the use of Spanish and Italian words in the written production of Portuguese, for instance, “Quando o sol tramontava” instead of “Quando o sol se punha” (When the sun was setting), where the Italian verb “tramontare” was activated and accepted by three participants as the target Portuguese word (Włosowicz, 2016, p. 79). In other words, the students mobilised their multilingual repertoires in the attempt to retrieve the target Portuguese items, but the perceived similarity between the languages was too great to allow effective control. Another factor which influences the probability of choosing a non-target language word is the level of activation of each language in the multilingual repertoire. According to Green’s (1986) Inhibitory Control Model, a language can be selected, active or dormant. A selected language is the one that is cur- rently being used, an active language is not being used, but it remains acti- Translanguaging as the Mobilisation of Linguistic Resources… 71 vated and participates in processing, whereas a dormant language remains in long-term memory and does not have any effect on processing (Green, 1986, p. 215). For example, in a bilingual speaker, if L1 is selected, L2 is externally suppressed and thus the phonological assembly of L2 words is inhibited (Green, 1986, p. 217). The fact that bilingual and multilingual speakers can use a single lan- guage in particular contexts indicates that they can indeed inhibit the non- target language(s), for example, while talking to a monolingual. On the other hand, a conversation among bilinguals can contain elements of both languages (Grosjean, 2001, p. 5). Thus, Grosjean (2001, p. 3) speaks about a continuum of language modes, ranging from the monolingual to the bilingual language mode. He defines a language mode as “the state of activation of the bilingual’s languages and language processing mechanisms at a given point in time” (Grosjean, 2001, p. 3). In the bilingual mode, the dominant language is called the base language, while the less active one is the other language or the guest language. In fact, as Grosjean (2001, p. 7) admits, “it is proposed that the other language is probably never totally deactivated at the monolingual end and that it very rarely reaches the same level of activation as the base language at the bilingual end.” However, according to Dijkstra and Van Hell (2001), the amount of control over the activation of languages is limited and lexical candidates belonging to different languages are activated on the basis of the input word rather than the target language. By contrast, the results obtained by Dunn and Fox Tree (2014) show that language mode activation to some extent depends on language dominance. What is also interesting is the fact that, unlike bilinguals, trilinguals possess better control and regulation mechanisms, which Cedden and Saǧın Şimşek (2014, p. 566) summarise as follows: “a third language system represented in the mind has the effect of promoting experience or regulation costs of the executive control system which might lead to the development of a more sophisticated and balanced language system.” Undoubtedly, knowing that the interlocutor knows both languages allows bilinguals to switch more freely without being afraid that the interlocutor will fail to understand. Code-switching can also be used as a communication strat- egy, for example, if the speaker does not know the target word or a non-target language word is more available (Poulisse & Bongaerts, 1994, p. 36). Among interlingual communication strategies, Faerch and Kasper (1983, pp. 46–47) mention code-switching (ranging from single words to complete turns, though single-word switching is also referred to as borrowing) interlingual transfer (a combination of linguistic features from the native language and the inter- language, also called “foreignizing” if it involves adjusting the L1 morphol- ogy or phonology, and “literal translation” in the case of the word-for-word translation of idioms and compounds, p. 47), and inter-/intralingual transfer, where the generalisation of an interlanguage rule is influenced by the L1 Teresa Maria Włosowicz72 rule (for example, an irregular L2 word may have a regular L1 equivalent). It can thus be seen that code-switching cannot be dismissed as a form of interference, but it performs a particular role in bilingual and multilingual communication. As shown by Morytz (2017), though Polish learners of Italian use a va- riety of communication strategies, the dominant strategies are the use of gestures, that of an electronic translator and transfer from another language (Morytz, 2017, p. 203), especially English (p. 199). In her view (Morytz, 2017, p. 203–204), foreign language teaching should involve more strategy training and metalinguistic awareness raising. In particular, students should be taught to replace unknown words with synonyms, hyperonyms, paraphrases, etc., rather than resorting to extralinguistic strategies, which do not contribute to the development of language skills (Morytz, 2017, p. 204). She also observes that, in the Polish context, students who lack an Italian word tend to replace it with a Polish one, because it is going to be understood anyway (Morytz, 2017, p. 197). Moreover, as shown by Williams and Hammarberg (1998), code-switching in multilinguals can be of different types in which the languages perform a variety of functions. For example, L1 (English in their study) predominantly serves such functions as META (comments on one’s own performance, requests for help, etc.), EDIT (self-repair, facilitating interaction, etc.), and INSERT (in- serting a word or phrase, for example, to elicit a Swedish L3 word, Williams & Hammarberg, 1998, pp. 306–309). On the other hand, L2 German occurred mainly in WIPP switches (“Without Identified Pragmatic Purpose,” p. 308), which were non-intentional and, as they were often followed by self-repair, they did not serve to elicit Swedish words, which the learner already knew. On the basis of these results, Williams and Hammarberg (1998) proposed a polyglot speaking model, assigning roles to the different languages: L1 English has an instrumental role, while L2 German is called a “default supplier” (Williams & Hammarberg, 1998, p. 318), or a language that is co-activated all the time and influences target language lexical planning and structures. Williams and Hammarberg (1998, p. 322, their emphasis) suggest that “the assignment of default supplier role may be the result of interplay between four factors, namely proficiency, typology, recency, and L2 status.” In other words, a default supplier is a foreign rather than the native language, it is typologically close to the target language, it has recently been used (so it remains active) and one has a fairly high level of proficiency in it. It can thus be concluded that multilinguals do use their language repertoires to communicate the intended meanings, in a way that involves their multiple languages, some of which (if not all) remain active and participate in processing. In fact, even in the monolingual mode the non-target languages are not fully deactivated and the amount of control a speaker can exert is limited. Hence, Translanguaging as the Mobilisation of Linguistic Resources… 73 it may be assumed that the mobilisation of one’s whole linguistic repertoire is a normal phenomenon and can, at least in some cases, be capitalised on rather than suppressed. Translanguaging The Phenomenon of Translanguaging By and large, translanguaging has been defined by Baker (2011, p. 288, as cited in Lewis, Jones, & Baker, 2012, p. 655) as “the process of making meaning, shaping experiences, gaining understanding and knowledge through the use of two languages.” More particularly, in the classroom context, “trans- languaging tries to draw on all the linguistic resources of the child to max- imise understanding and achievement” (Lewis et al., 2012, p. 655). Certainly, this does not necessarily have to apply to children as such, as adult learn- ers can also draw upon all their linguistic resources. In a similar vein, García (2009a, p. 140) defines translanguaging as “the act performed by bilinguals of accessing different linguistic features or various modes of what are described as autonomous languages, in order to maximize communicative potential.” Therefore, as García (2009a, p. 140) admits, translanguaging “goes beyond what has been termed codeswitching, although it includes it.” In the present article, all the above definitions of translanguaging will be relied on, regarding it as the use of multilingual resources to make communication more efficient, but also to facilitate learning, and code-switching will be considered one of the ways of facilitating understanding and communication. Even though some more recent studies on translanguaging (e.g., Otheguy, García, & Reid, 2018) reject the boundaries between languages in multilingual repertoires and thus also code-switching as the act of switching between languages, there is still evidence in favour of the psychological reality of code-switching and some rules governing it (MacSwan, 2017; Toribio, 2001). Certainly, there are situ- ations where the term “code-switching” becomes irrelevant because speakers mix several languages, none of them being dominant, so “translanguaging” is the only appropriate term, as in the case of the multilingual community in Singapore described by Li (2018, pp. 13–14). In a single dialogue, the speakers use Hokkien, Teochew, Mandarin, Malay, Cantonese, Singlish, and English, and, as Li (2018, p. 14) concludes, “[a] classic code-switching approach would assume switching back and forward to a single language default,” which is impossible in that community. However, the participants in the present study Teresa Maria Włosowicz74 do not belong to such a community and are more conscious of code-switching between their different languages. As Lewis et al. (2012, p. 659) observe, “[t]here is clearly much overlap between code-switching and translanguaging,” and the difference is mainly related to the fields the terms originate from: while “code-switching” is de- rived from the linguistic analysis of bilingual speech, “translanguaging” is ap- plicable to situated language use in sociolinguistics. According to Juffermans, Blommaert, Kroon, and Li (2014, p. 49, their emphasis, as cited in Treffers- Daller, 2018, p. 13), the difference between code-switching and translanguag- ing is not phenomenological but theoretical, as “codeswitching grosso modo takes a structural perspective on bilingual text or talk whereas translanguag- ing focuses primarily on what speakers actually do and achieve by drawing on elements from their repertoires in situated contexts.” In other words, the learners’ language production may actually look the same in both cases, but, while a code-switching approach would analyse the grammatical elements that can be switched and those which cannot (e.g. Toribio, 2001), a translanguaging approach would consider the use of multilingual repertoires to facilitate commu- nication. It must thus be remembered that code-switching and translanguaging can actually be very similar, but that translanguaging takes a broader outlook on multilingual communication, including the use of other semiotic means, such as gestures. As Li (2018, p. 20, his emphasis) concludes, “[l]anguage, then, is a multisensory and multimodal semiotic system interconnected with other identifiable but inseparable cognitive systems.” He then moves on to define translanguaging as “transcending the traditional divides between lin- guistic and non-linguistic cognitive and semiotic systems” (Li, 2018, p. 20). Another difference between code-switching and translanguaging is that trans- languaging is “a more systematic and strategic process that allows the speaker to make meaning and to foster the affective side of language use in such a way that bilinguals use the whole linguistic and semiotic repertoire at their disposal to shape their experiences and create meaning” (Doiz & Lasagabaster, 2017, p. 160). In fact, not only does translanguaging transcend the boundaries between languages, but some researchers actually assume unitary competence, without any boundaries within multilingual repertoires. For example, Otheguy et al. (2018, p. 2) claim that “[t]he myriad linguistic features mastered by bilinguals (phonemes, words, constructions, rules, etc.) occupy a single, undifferentiated cognitive terrain that is not fenced off into anything like the two areas sug- gested by the two socially named languages.” In their view, the division of languages into separate entities, such as English, Spanish, etc., “is anchored in sociocultural beliefs, not in psycholinguistic properties of the underlying system” (Otheguy et al., 2018, p. 4, their emphasis). As a result, even though the existence of some internal differentiation is obvious, the claim that the dif- Translanguaging as the Mobilisation of Linguistic Resources… 75 ferentiation is specific to the separate languages present in the system is based on the social division between them (Otheguy et al., 2018, p. 8). According to García and Otheguy (2019, p. 10), translanguaging involves “the deployment of features that are most appropriate to communicate a message to a listener,” but those features do not have to belong to a single language and they do not even have to be linguistic, as communication can also include gestures, gaze, posture, etc. By contrast, MacSwan (2017, p. 169, his emphasis) offers “a multilingual perspective on translanguaging,” which recognises the existence of discrete languages within multilingual repertoires and which regards code-switching as a rule-governed activity. In contrast to the Dual Competence Model, which as- sumes the existence of fully discrete language systems, and the Unitary Model of Multilingualism, based on the assumption of a single, undifferentiated system (represented, for example, by Otheguy et al., 2018), MacSwan (2017, p. 179) proposed the Integrated Multilingual Model, according to which “bilinguals have a single system with many shared grammatical resources but with some internal language-specific differentiation as well” (MacSwan, 2017, p. 179). MacSwan (2017, p. 179) adds that it is not monolingualism but multilingual- ism that is universal, and that monolingualism is a social construct, because some internal differentiation can be observed even in so-called monolingual systems (MacSwan, 2017, p. 185). Following Grosjean (1982), MacSwan (2017, p. 190) recognises that “[a] bilingual is a uniquely situated language user who functions bilingually, drawing on whatever language resources are appropriate, and is not the sum of two monolinguals.” However, what is regarded by speakers as appropriate is also grammati- cally based. For example, MacSwan (2017, p. 181, his emphasis) observes that Spanish-English bilinguals may say “the white house, la casa blanca and the white casa, but not the house white, la blanca casa, or the casa white,” which he attributes to “structured and internally organized differentiation of some kind” (MacSwan, 2017, p. 181) and to bilinguals’ sensitivity to that complex system of rules (p. 190). Some evidence of the rule-governed character of code- switching is provided by Toribio’s (2001) study, in which bilinguals at differ- ent levels of proficiency made judgements on the grammaticality of different switches. She concludes that advanced bilinguals’ code-switching behaviour reflects the constraints of Universal Grammar (Toribio, 2001, pp. 226–227), but she also observes that “the rule-governed nature of code-switching is upheld by even the non-fluent bilinguals in the sample, whose behaviour suggests at least enough incipient competence in the second language to switch codes” (Toribio, 2001, p. 225). According to Otheguy et al. (2018, p. 15), grammatical- ity judgements require specifying the language system in which an utterance is grammatical or not, and bilingual informants who believe that they speak two languages. Teresa Maria Włosowicz76 Even so, it can be assumed that learners who grew up in a basically mono- lingual community (although, following MacSwan, 2017, it can be admitted that monolingualism is a social construct, so no speaker or community is fully monolingual; rather, it is a theoretical oversimplification based on the fact that they speak one socially named language at home, they learn another socially named language at school, and another at university, etc.) and learnt their L2, L3, etc. as foreign languages, do distinguish between them. Certainly, there are communities where the boundaries between the languages are blurred, as in the multilingual community in Singapore (Li, 2018), but in the present study, the participants can be assumed to be aware of the internal differentiation of their multilingual repertoires and to use code-switching and other forms of translanguaging as conscious strategies rather than their natural language production. Undoubtedly, in the Polish context, one cannot assume unitary multilingual competence, given the existence of monolingual Polish speakers (at least func- tionally monolingual, because their knowledge of foreign languages, especially Russian, which was taught in Polish schools for several decades, is dormant), who would not understand a mixture of Polish, English, and— in the case of the participants in the present study—Spanish and, possibly, French, Italian, and Portuguese. Some translanguaging in possible in the case of regional lan- guages, such as Silesian, for example, Arabski (2002, p. 211) mentions the use of Silesian words as keywords in memorising English and German vocabulary. However, unlike Kashubian, Silesian has the status of a dialect and, in spite of considerable debate (Myśliwiec, 2013), the Polish Parliament has not recognised it as a language (TVS, 2019). Moreover, while Polish and Silesian are similar enough to make such translanguaging comprehensible, a mixture of Polish and English would be comprehensible to Polish speakers of English and a mixture of Polish, English, and Spanish would require of the recipient knowledge of all these languages. In Morytz’s (2017) study cited above, Polish learners of Italian tend to resort to English as a better-known language and, as she observes (Morytz, 2017, p. 201), English has now become a lingua franca, so it is likely to be understood, also by an Italian speaker. In fact, mixing Polish with English as a lingua franca occurs in the Polish branches of international companies, but even such an informal, mixed language variety has its own rules, which might challenge Otheguy et al.’s (2018) claim that separate languages are only social constructs. As shown by Włosowicz (2013), mixing Polish, English, and French in an international company is subject to certain constraints (for example, some English and French words are only used as terminology), so even in a community using all three languages certain usages are considered acceptable, while others are not. Translanguaging as the Mobilisation of Linguistic Resources… 77 Translanguaging as a Pedagogical Practice Translanguaging originated from a pedagogical practice used in Welsh schools. The term was first used by Williams (1994, as cited in MacSwan, 2017, p. 170) in reference to “the planned and systematic use of two languages inside the same lesson” (Baker, 2011, p. 288, as cited in MacSwan, 2017, p. 170). As Lewis et al. (2012, p. 657) remark, historically, translanguaging is related to classroom code-switching. However, while translanguaging is often spontaneous, initiated by the learners themselves, who use “both their languages to maximise understanding and performance” (Lewis et al., 2012, p. 658), “responsible” code-switching (García, 2009b, as cited in Lewis et al., 2012, p. 658) is used by the teacher to clarify the L2 material, to develop the learners’ metalinguistic understanding and to increase their metacognitive awareness. As Duarte (2018, p. 3) points out, there is considerable evidence of the advantages of translanguaging “at different levels of school performance and for both migrant and minority languages.” In her article, Duarte (2018) presents two school contexts in which translanguaging is used: at a kindergarten in Luxembourg and at a primary school in the Netherlands. As she remarks (Duarte, 2018, p. 12), her examples are instances of “official translanguaging,” planned by the teachers and systematically applied. She distinguishes three functions of official translanguaging: the symbolic function (acknowledgement of the pupils’ native languages), the scaffolding function (building bridges be- tween the languages and attributing equal value to them) and the epistemologi- cal function, or the use of translanguaging for content and language learning, which require of the teacher different levels of competence in the languages involved, from no proficiency at all, except in the instruction language, to proficiency in both (or more) languages (Duarte, 2018, p. 13). As Duarte (2018, p. 14) concludes, translanguaging allows “pupils to actively use their dynamic plurilingual practices for learning” and her typology may help teachers to develop their own translanguaging practices in the future. However, multilingualism has long been ignored in education. In general, language teaching has followed what Howatt (1984, as cited in Cummins, 2008, p. 65) called “the monolingual principle,” according to which foreign language teaching was supposed to take place in the target language, excluding the learners’ L1, while immersion programs were designed to keep both languages separate (Cummins, 2008, p. 65). However, Cummins (2008) advocates the abandonment of reliance on such monolingual approaches and emphasises the role of promoting cross-language transfer. Following Donovan and Bransford (2005, p. 4, their emphasis, as cited in Cummins, 2008, p. 67), who observe that “new understandings are constructed on a foundation of existing understand- ings and experiences,” Cummins (2008, p. 68) claims that foreign language Teresa Maria Włosowicz78 teaching “should explicitly attempt to activate students’ prior knowledge and build relevant background knowledge as necessary.” According to the inter- dependence hypothesis (Cummins, 1981, as cited in Cummins, 2008, p. 68), proficiency in Lx can be transferred to Ly if there is both adequate exposure to Ly and sufficient motivation to learn it. Such transfer can apply to conceptual and linguistic elements as well as to metacognitive and metalinguistic strate- gies, pragmatic aspects of language use and phonological awareness (Cummins, 2008, p. 69). Therefore, it can be said that reference to learners’ prior knowledge (for example, L1 and L2 proficiency in L3 learning) can facilitate learning and should be used responsibly rather than avoided. However, as shown by Doiz and Lasagabaster (2017), teachers do use trans- languaging in English-medium instruction (EMI), even though there is some discrepancy between their beliefs and practices. Even though they claim to believe that the L2 should always be used, they use the L1 in teaching, for ex- ample, to explain specialised vocabulary, or at least for organisational purposes, such as making announcements (Doiz & Lasagabaster, 2017, p. 169). However, while translanguaging is common outside the classroom, for example, during office hours, “it is not generally accepted in classroom interactions and assess- ment tasks” (Doiz & Lasagabaster, 2017, p. 173). As Doiz and Lasagabaster (2017, p. 174) conclude, teachers should, first, focus on both language and content so that the students profit more from EMI and, second, they may need training in translanguaging in order to “break away from the monolingual view of language codes” (Doiz & Lasagabaster, 2017, p. 174). Still, the monolingual view seems to be represented not only by teachers, but also by students and, to some extent, it is actually good to practice the target language as much as possible, but the advantages of positive transfer and language awareness should not be neglected either. Last but not least, in the context of the present study, it is worth pointing out that different criteria should be applied to students’ proficiency and lan- guage use in different study programmes and at different levels of education. While minority and immigrant children may be encouraged to use all their language repertoires so as to feel that their home languages are valued and to develop additive bilingualism, philology students should, arguably, be taught using a different approach. Though some reference to languages other than the target language should be used, for example, to raise metalinguistic aware- ness, as future teachers, translators, etc., they should obey certain monolingual norms as well. After all, a translation, say, from Polish into English, contain- ing a mixture of Polish and English words (other than proper names, culture- specific terms explained by the translator, etc.), would be incomprehensible to English recipients. Translanguaging as the Mobilisation of Linguistic Resources… 79 The Study Participants The study was carried out with twenty-six participants learning Spanish as a third or additional language, seven of whom were English Philology students following a Spanish language course at the branch of Ignatianum University in Mysłowice and at the University of Social Sciences in Cracow, and nineteen were Romance Philology students studying different Romance lan- guages (Spanish, Portuguese, French, Italian or even Romanian) at Jagiellonian University and at the University of Silesia. Their Spanish was basic (A1/A2) or intermediate (B1) at most, at least by their own admission. As time limitations made it impossible to carry out placement tests in English and Spanish, their proficiency levels had to be estimated, on the one hand, on the basis of the questionnaires and, on the other hand, on the knowledge second-year Romance Philology students could be assumed to have and, in the case of the present author’s English Philology students, the syllabuses and their performance dur- ing the classes. Their native language (L1) was Polish, except for one English Philology student who indicated Russian and one Romance Philology student who indi- cated both Polish and Italian (i.e., she had been raised bilingually). Their L2 was predominantly English, as it is the most frequently taught foreign lan- guage in Poland and, generally, in Europe (the L1 Russian participant is from Belarus). However, two Romance Philology students indicated Spanish as their L2, one indicated Italian and one – Romanian. It might thus be questionable whether the L2 Spanish students can be included in the analysis. Actually, it can be assumed that they can because, first, English is more likely to be their L2 (it is taught from an earlier age and it is unlikely that they started Spanish earlier or that they achieved a higher level of proficiency in Spanish in a much shorter time), second, with parallel language learning, the acquisi- tion order can be established only approximately (Cenoz, 2000), and third, they may not necessarily have indicated their language sequences correctly. For example, five English Philology students did not indicate Spanish at all, even though they were studying it. In fact, at secondary school all the English Philology students had also had a foreign language other than English, that is, German or French (Włosowicz, in preparation), so Spanish was actu- ally their L4, even though only the L5 Spanish student indicated French as L3 and German as L4, whereas the other students did not mention their L3 German or French at all. The L1 Russian student indicated English as L2 and Polish as L3, while Spanish as L4 could be assumed because she followed the Spanish course. Teresa Maria Włosowicz80 The Romance Philology students had more varied language repertoires. Apart from Polish (in one case, Polish and Italian) as L1 and English (14 par- ticipants), Spanish (two participants, though the L2 status is questionable, see above), Italian (1) and Romanian (1) as L2, they had Spanish (12), German (2), French (2) and English (2) as L3, Spanish (4), French (3), German (2), English (1) and Portuguese (1) as L4, and Portuguese (5) as L5. Even though two of the Romance Philology students did not mention English in the questionnaires, they can be assumed to have studied it, or else they would not have been able to do the tasks and complete the questionnaire. It can thus be seen that the participants’ language repertoires are varied and the languages can be assumed to have been acquired largely simultaneously rather than consecutively, which complicates the establishment of acquisition orders even further. Indeed, as observed by Van Gelderen et al. (2003, p. 23), L3 learner populations are usually more heterogeneous than the research de- sign would require. At the same time, as Cenoz (2000, p. 40) pointed out, in multilingual acquisition, it is no longer possible to mark simultaneously acquired languages as L1, L2, L3, etc., but rather as Lx, Ly, etc. Hence, in simultaneous L3 and L4 acquisition, the sequence would be L1 → L2 → Lx/Ly. Therefore, the participants’ repertoires and acquisition orders were heterogeneous and could be determined only approximately, but it can be as- sumed that Spanish has the status of a third or additional language, whether as L3, L4 or L5, or as Lx in the case of simultaneously studied Lx, Ly, and even Lz. Method The research tool used in the study was a written test, followed by a questionnaire. The test was designed in such a way as to co-activate the participants’ languages and to provoke some translanguaging and, in particu- lar, code-switching. As translanguaging mobilises all of a learner’s language resources, it was assumed that the use of other languages, especially English, in both comprehension and production, could be regarded as translanguaging (for example, some expressions were prompted in English and the participants were supposed to provide the Spanish equivalents). The test consisted of three tasks: first, a gap-filling task consisting of two dialogues, one between a shop assistant and a customer and one between a tourist asking for directions and a receptionist, second, a gap-filling task consisting of ten independent sentences which had to be completed with words or expressions, and third, an error correction task with ten Spanish sentences to be classified as correct and incorrect, with a correction and an explanation in the case of the incorrect sentences. In the first task, some of Translanguaging as the Mobilisation of Linguistic Resources… 81 the possible Spanish target words were prompted in a box below the dia- logues, but no translation was provided, so the participants were supposed to recognise them in order to use them. However, some non-target Spanish words which had also been studied at the learners’ level, such as “calcetín” (sock) and “falda” (skirt), were also given in the box for the task not to be too easy. Alternatively, the students were encouraged to use words from other languages (English, as was written in the instructions, originally designed for English Philology students, but also French, Portuguese, Italian, etc., which the present author told the Romance Philology students while explaining the tasks to them) to show that they understood the sentences and knew the meanings of the target words. In other words, they were supposed to use code-switching as a strategy. The second task involved translanguaging in the form of L2 English language comprehension, which was supposed to facili- tate the choice of the Spanish expressions, as leaving gaps without any cues might not have prompted the intended meanings to them, while cues in Polish would have triggered lexical-level connections (Kroll & Stewart, 1994), but they might not have mobilised more of the participants’ language resources. The third task involved translanguaging in a different sense: rather than using words or expressions from other languages to facilitate the task, the partici- pants were supposed to judge the grammaticality of Spanish sentences, some of which involved negative transfer from Polish, English, or both, so mobilis- ing the language repertoires was rather meant to identify the rules underlying the stimulus sentences and find the correct ones. Thus, unitary competence not specifying the divisions between the particular languages (Otheguy et al., 2018) would have rendered the error correction task more difficult, while integrated multilingual competence (MacSwan, 2017) would have permitted the identification of sentences correct in Spanish or, on the contrary, based on negative transfer from Polish, English or both. The tasks were followed by a questionnaire on the participants’ language repertoires (it even contained an example of a similarity between Spanish and Russian, for the Russian- speaking students of the University of Social Sciences) and the difficulties and cross-linguistic interaction (Herdina & Jessner’s (2002) term) they had encountered while studying Spanish and performing the tasks. The tasks are presented in Appendix 1 and the questionnaire – in Appendix 2 at the end of the article. The research questions were as follows: First, to what extent did the participants in each group (English Philology and Romance Philology) use translanguaging to solve the tasks? In particular, to what extent did they use code-switching as a strategy to fill in the gaps with words from languages other than Spanish? Second, what do the results reveal about their multilingual repertoires and the place of Spanish in them, as well as about the character (unitary or differentiated) of their multilingual competence? Teresa Maria Włosowicz82 Results As the results show, in Task 1 relatively few switches into English were observed (thirteen in the English Philology group, nine of which were correct (8) or contextually acceptable (1) and four were contextually unacceptable, and two in the Romance Philology group, both correct) and none into Portuguese, French or Italian. The participants preferred to rely on their knowledge of Spanish, producing either correct, contextually acceptable or contextually un- acceptable responses, but avoidance was also quite frequent, especially in the English Philology group. Correct responses were considered to be words which fitted well in the context of the dialogue (there was often more than one pos- sible correct answer), contextually acceptable ones sounded slightly odd, but they still could be thought of, for example, “ochocientos gramos de pimienta” (eight hundred grams of pepper (as a spice, not a vegetable)), and contextually unacceptable answers did not fit in the context at all, for example, “¿Tiene pi- mienta?” “No, tiene que ir a la carnicería” (“Do you have pepper?” “No, you have to go to the butcher’s”), where the target word had to be some kind of meat, or even “ochocientos gramos de calcetín” (eight hundred grams of sock). Similarly, English words provided instead of Spanish ones could be correct, contextually acceptable or contextually unacceptable. As a translanguaging approach was adopted, their correctness was regarded as the same as that of their Spanish equivalents. At this point, it must be stressed that it was not a traditional error analy- sis, judging the students’ performance in reference to strict grammatical and semantic rules, but rather an evaluation of the communicative potential of their responses. That is why the use of English words for lack of Spanish ones was not rejected, but even encouraged as a communication strategy. Following Morytz (2017, p. 199), it could be assumed that the participants would resort to English in case of difficulty in finding contextually appropriate Spanish words. However, even the most liberal approach to translanguaging, rejecting all boundaries between “socially named languages” (Otheguy at al., 2018, p. 2) and all language-specific rules, should, arguably, take into consideration the use of such sentences in communication. Thus, telling the shop assistant at the grocer’s or at the greengrocer’s that you want to buy 800 grams of sock does not seem communicative at all, that is why such responses were classified as “contextually unacceptable.” The percentages of the English Philology students’ responses to Task 1, based on their acceptability and the languages used, are given in Table 1 (the dialogue in the shop) and Table 2 (the dialogue between the tourist and the receptionist), and those of the Romance Philology students’ responses are given in Tables 3 and 4 respectively. Translanguaging as the Mobilisation of Linguistic Resources… 83 Table 1 Percentages of the English Philology students’ responses provided in Task 1, dialogue 1 Gap in the text Correct Spanish word Correct English word Contextually acceptable Spanish word Contextually acceptable English word Contextually unacceptable Spanish word Contextually unacceptable English word Avoidance [%] ¿Qué le ? 0 0 14.29 0 14.29 42.86 28.57 dos 28.57 14.29 28.57 0 0 0 28.57 un kilo de 28.57 14.29 57.14 0 0 0 0 diez 42.86 14.29 14.29 0 0 0 28.57 doscientos gramos de 28.57 0 14.29 14.29 0 0 42.86 ¿Algo ? 28.57 14.29 0 0 0 0 57.14 ¿tiene ? 28.57 0 28.57 0 0 0 42.86 ¿Tiene también ? 28.57 0 28.57 0 14.29 0 28.57 ochocientos gramos de 28.57 0 14.29 0 14.29 0 42.86 una 0 0 57.14 0 14.29 0 28.57 ¿Algo ? 28.57 14.29 0 0 14.29 0 42.86 unos 14.29 0 14.29 0 0 0 71.43 ¿Tiene ? 28.57 0 0 0 14.29 0 57.14 ¿Algo ? 28.57 14.29 0 0 0 0 57.14 Tengo , 14.29 0 28.57 0 0 0 57.14 , 28.57 0 14.29 0 0 0 57.14 … 42.86 0 0 0 0 0 57.14 de cincuenta euros 0 28.57 0 0 28.57 0 42.86 Teresa Maria Włosowicz84 Table 2 Percentages of the English Philology students’ responses provided in Task 1, dialogue 2 Gap in the text Correct Spanish word Correct English word Contextually acceptable Spanish word Contextually acceptable English word Contextually unacceptable Spanish word Contextually unacceptable English word Avoidance [%] ¿Cómo 0 0 0 0 57.14 0 42.86 muy 0 0 0 0 28.57 14.29 57.14 Al del hotel 0 0 0 0 14.29 0 85.71 tiene que 14.29 0 0 0 14.29 0 71.43 todo 42.86 0 0 0 14.29 0 42.86 la Mayor 28.57 0 0 0 28.57 0 42.86 Atraviese la 28.57 0 0 0 28.57 0 42.86 después 0 0 0 0 28.57 0 71.43 la primera 42.86 0 0 0 0 0 57.14 ¿Y decirme 0 0 0 0 28.57 0 71.43 del Arte 14.29 0 0 0 14.29 0 71.43 bastante 0 0 0 0 42.86 0 57.14 el metro 0 0 0 0 42.86 0 57.14 Tiene que 0 0 0 0 14.29 0 85.71 cinco 14.29 0 14.29 0 14.29 0 57.14 Jardín Zoológico 0 0 0 0 14.29 0 85.71 a Goya 14.29 0 0 0 0 0 85.71 tres 0 0 0 0 14.29 0 85.71 Aeropuerto 0 0 0 0 14.29 0 85.71 a la tercera 0 0 14.29 0 0 0 85.71 que se llama 14.29 0 14.29 0 0 0 71.43 los 0 0 14.29 0 0 0 85.71 Hay 0 0 0 0 14.29 0 85.71 todas las 0 0 0 0 14.29 0 85.71 la más cercana 0 0 0 0 14.29 0 85.71 dolor de 0 0 0 0 14.29 0 85.71 de Flores 57.14 0 0 0 0 0 42.86 la primera 0 0 0 0 14.29 0 85.71 a la izquierda 42.86 0 14.29 0 14.29 0 28.57 todo 28.57 0 0 0 14.29 0 57.14 Translanguaging as the Mobilisation of Linguistic Resources… 85 Table 3 Percentages of the Romance Philology students’ responses provided in Task 1, dialogue 1 Gap in the text Correct Spanish word Correct En/Fr/ Pt/It word Contextually acceptable Spanish word Contextually acceptable En/Fr/Pt/It word Contextually unacceptable Spanish word Contextually unacceptable En/Fr/Pt/It word Avoidance [%] ¿Qué le ? 31.58 0 10.53 0 52.63 0 5.26 dos 94.74 0 5.26 0 0 0 0 un kilo de 89.47 0 10.53 0 0 0 0 diez 94.74 0 5.26 0 0 0 0 doscientos gramos de 100.00 0 0 0 0 0 0 ¿Algo ? 100.00 0 0 0 0 0 0 ¿Tiene ? 94.74 0 0 0 0 0 5.26 ¿Tiene también ? 89.47 0 5.26 0 0 0 5.26 ochocientos gramos de 73.68 0 15.79 0 5.26 0 5.26 una 89.47 0 5.26 0 0 0 5.26 ¿Algo ? 94.74 0 0 0 0 0 5.26 unos 57.89 0 26.32 0 0 0 15.79 ¿Tiene ? 57.89 0 36.84 0 0 0 5.26 ¿Algo ? 94.74 0 0 0 0 0 5.26 Tengo , 68.42 0 15.79 0 10.53 0 5.26 , 73.68 0 10.53 0 5.26 0 10.53 ... 68.42 0 15.79 0 5.26 0 10.53 de cincuenta euros 57.89 0 15.79 0 15.79 0 10.53 Teresa Maria Włosowicz86 Table 4 Percentages of the Romance Philology students’ responses provided in Task 1, dialogue 2 Gap in the text Correct Spanish word Correct En/Fr/Pt/ It word Contextually acceptable Spanish word Contextually acceptable En/Fr/Pt/It word Contextually unacceptable Spanish word Contextually unacceptable En/Fr/Pt/It word Avoidance [%] ¿Cómo 89.47 0 0 0 5.26 0 5.26 muy 63.16 5.26 0 0 26.32 0 5.26 Al del hotel 63.16 0 10.53 0 10.53 0 15.79 tiene que 78.95 0 10.53 0 5.26 0 5.26 todo 94.74 0 0 0 0 0 5.26 la Mayor 89.47 0 0 0 5.26 0 5.26 Atraviese la 84.21 0 5.26 0 5.26 0 5.26 después 15.79 0 78.95 0 0 0 5.26 la primera 94.74 0 0 0 0 0 5.26 ¿Y decirme 84.21 0 10.53 0 0 0 5.26 del Arte 89.47 0 0 0 5.26 0 5.26 bastante 78.95 0 10.53 0 5.26 0 5.26 el metro 89.47 0 5.26 0 0 0 5.26 Tiene que 73.68 0 10.53 0 0 0 15.79 cinco 68.42 0 10.53 0 10.53 0 10.53 Jardín Zoológico 63.16 0 5.26 0 10.53 0 21.05 a Goya 47.37 0 10.53 0 10.53 0 31.58 tres 68.42 0 10.53 0 5.26 0 15.79 Aeropuerto 31.58 0 31.58 0 5.26 0 31.58 a la tercera 15.79 0 68.42 0 0 0 15.79 que se llama 68.42 0 0 0 26.32 0 5.26 los 89.47 0 0 0 5.26 0 5.26 Hay 52.63 5.26 15.79 0 5.26 0 21.05 todas las 57.89 0 26.32 0 0 0 15.79 Translanguaging as the Mobilisation of Linguistic Resources… 87 la más cercana 78.95 0 10.53 0 5.26 0 5.26 dolor de 94.74 0 0 0 0 0 5.26 de Flores 94.74 0 0 0 0 0 5.26 la primera 73.68 0 21.05 0 0 0 5.26 a la izquierda 94.74 0 0 0 0 0 5.26 todo 94.74 0 0 0 0 0 5.26 It can thus be seen that the percentages of correct answers are generally higher in the Romance Philology group. However, even the Romance Philology students had problems with some items, especially “y después la primera” and “ a la tercera” (15.79% of correct answers in each case), “ Aeropuerto,” “¿Qué le ?” (31.58% of correct answers in each case) and “ a Goya” (47.37% of correct answers). In the case of “y después la primera ” (and then the first ), the target words were “tome” or, possi- bly, “coja” (take) and “calle” (street). 78.95% of the answers were contextually acceptable, for example, “coge” or “toma”; even though the verbs were cho- sen correctly, the forms of the less polite imperative (correct while talking to a friend, but not to a tourist who is a customer at the hotel) were used. The English Philology students provided no correct answer and the only attempt at filling in the gap was a grammatically incorrect Spanish word (not only was the verb form informal, but its syntactic properties did not make it fit in the con- text either): “y después vas la primera calle” (and then you go the first street). The word “calle” proved much easier, as it was correctly given by 94.74% of the Romance Philology students and also by 42.86% of the English Philology students. With “ a la tercera,” the target word was “baje” (descend at the third [station]), which the Romance Philology students mostly provided in the informal imperative form, for example, “baja” or “sale” (instead of “salga,” from “salir” – get off/out; 68.42% of the answers were classified as contextually acceptable) and the English Philology group mostly avoided this item (85.71% of avoidance), only one person provided the contextually acceptable answer “Ir a la tercera calle” (go (infinitive) to the third street); in fact, the other missing word was not “calle,” but “estación” (station) or “parada” (stop). In the case of “ Aeropuerto,” the target answer was “dirección Aeropuerto” ([in the] direction [of the] Airport), which, apparently, was not so obvious, even though the context indicated the final destination of the under- ground line. Other possible answers included “al Aeropuerto” (to the airport; accepted as correct), “hacia Aeropuerto” (towards the Airport; contextually ac- Teresa Maria Włosowicz88 ceptable), etc. The English Philology students mostly avoided this item (85.71%), except one person who wrote “lejos Aeropuerto” (far away the airport), which was both semantically and syntactically odd (hence, contextually unacceptable). The Romance Philology students, apart from the 31.58% of correct answers mentioned above, produced 31.58% of contextually acceptable answers and one contextually unacceptable answer (5.26%), “tres metros del Aueropuerto” (three metres from the Airport); there was also considerable avoidance in comparison with the other items (31.58%). With “ a Goya,” one English Philology student provided a correct answer, “ir” (“go”; in the context “tiene que […] ir a Goya,” “you have to […] go to Goya,” it was fully acceptable) and the remaining six left a gap. The Romance Philology students were not very sure, as the avoidance rate (31.58%) shows, only 47.37% of the responses were cor- rect (e.g., “bajar a Goya”), 10.53% were contextually acceptable, and 10.53% were contextually unacceptable, for example, “arena a Goya” (sand at Goya; the student may have confused two Spanish words, but the source of the error is impossible to identify). The shop dialogue was generally easier, but it is surprising that the expres- sion “¿Qué le pongo?” (“What would you like?,” literally: “What shall I give you?,” used by shop assistants and present in Spanish language textbooks) caused the participants so much difficulty. In the English Philology group, no correct answer was provided, and in the Romance Philology group, only 31.58% of the responses were correct, while 52.63% were contextually unacceptable and 10.53% were contextually acceptable. Interestingly enough, one English Philology student provided a contextually acceptable Spanish word, “¿Qué le gustaría?” (“What would you like?”, but rather in the sense of “What would please you?” than “What would you like to buy?”), which seems to be a case of translanguaging and mobilising all of one’s language resources, including English, and producing a calque of the English phrase “What would you like?”. Three English Philology students used code-switching, but the English words did not fit in the context (“¿Qué le want?” and “¿Qué le need?”). The Romance Philology students mainly produced contextually unacceptable Spanish words, such as “¿Qué le necesita?” (the intended answer was “¿Qué necesita?” (What do you need?), but the pronoun “le” required a verb with other syntactic prop- erties), and, similarly, “¿Qué le ayuda?” (What helps you?) and “¿Qué le puedo ayudar?” (literally: What can I help you?), probably under the influence of the English “How can I help you?”. The idiomatic, frequently used phrase “¿Qué le pasa?” (What is happening to you?, i.e., Are you O.K.?) was also observed. In general, the names of the products were provided fairly well, also because there were many acceptable possibilities, such as different names of fruits and vegetables. However, some of the students, especially in the English Philology group, provided answers which were either contextually incompatible and thus unacceptable (e.g., “ochocientos gramos de calcetín y una oranja grande,” where Translanguaging as the Mobilisation of Linguistic Resources… 89 “oranja” instead of “naranja” (orange) is due to interference from English, or “tengo taquilla, churros, coliflor…“ (I have ticket office, churros, cauliflower…); possibly, the student wanted to write “tequila”), or contextually acceptable (se- mantically possible in the context), but containing a grammatical gender error, e.g., “una chocolate grande” (a big chocolate), where “el chocolate” is masculine. By contrast, the sentence “¿Tiene de cincuenta euros?”, where the target word was “cambio” (“change,” as the customer asked: Do you have change for fifty euros?; but “vuelta” was also possible), proved more difficult. In the English Philology group, no correct answer was provided in Spanish, but two students (28.57%) provided correct English answers: “¿Tiene change de cin- cuenta euros?,” a switch to English which revealed their correct comprehension. However, avoidance was also frequent (42.86%) and two students gave contex- tually unacceptable answers in Spanish: “¿Tiene más de cincuenta euros?” (Do you have more than fifty euros?) and “¿Tiene 98 de cincuenta euros?”. In the Romance Philology group, 57.89% of the answers were correct Spanish words, no-one switched to English or any other language, there were also 15.79% of contextually acceptable answers (e.g., ¿Tiene dar la vuelta de cincuenta euros?, which was comprehensible but syntactically odd), 15.79% of contextually unac- ceptable ones (e.g., ¿Tiene coger de cincuenta euros? (approximately: “Do you have to take fifty euros?,” but syntactically odd)), as well as 10.53% of avoidance. The numbers of correct, contextually acceptable and contextually unac- ceptable answers in Spanish and other languages, as well as avoidance, in both groups were compared by means of a chi-square test for each dialogue. For the shop dialogue, the difference between the groups was statistically significant at p < 0.001 (df = 6) and, for the dialogue between the tourist and the receptionist, the difference was also statistically significant at p < 0.001 (df = 6). On the one hand, the Romance Philology students produced more correct answers and less avoidance, and on the other, their higher proficiency in Spanish permitted them to use it more often, rather than to rely on code-switching or to resort to avoidance. In general, translanguaging was not used very often, but switches to English were more frequent in the English Philology group, probably because of a higher level of proficiency in English and a lower one in Spanish; possibly, English was a kind of default supplier for the students, especially because they had been explicitly encouraged to use code-switching as a strategy. However, in both groups some indirect influence of English could be observed, as in the examples “¿Qué le ayude?”, “¿Qué le ayuda?”, and “¿Qué le puedo ayudar?”, based on “How can I help you?”. This indicates that the students mobilised their multilingual resources, rather than limiting themselves to Spanish, and it is possible that translanguaging as the consultation of one’s multilingual resources was inevitable, as the other languages are never fully deactivated. Moreover, some of the English Philology students’ answers contain “wild guesses” which suggest that they did not understand the Spanish sentences Teresa Maria Włosowicz90 very well, for example: “Hay mantequilla en todas las sandías” (There is but- ter in all the watermelons). Yet, even the Romance Philology students did not always control the contextual compatibility of their answers, for example: “No, usted tiene que ir a la carnicería. ¿Algo diferente? Tengo pollo, pescado, carne de cerdo…” (No, you have to go to the butcher’s. Anything different? I have chicken, fish, pork…). As the shop assistant did not have any meat and told the customer she had to go to the butcher’s, offering her chicken and pork in the next sentence was incompatible in the context, which suggests that the student processed the text sentence by sentence and, in order not to overburden her working memory, she did not keep focused on a larger context. As mentioned in Section Method, even though Task 2 also involved trans- languaging, it was not in the form of switching to a language other than Spanish, but rather in the form of consulting the English mental lexicon and mental translation from English into Spanish. The English expressions in brack- ets served to some extent as prompts (e.g.,, that Ana goes on holiday to the seaside and not e.g. that she goes with pleasure to the seaside), but they could also provoke some interference errors if the English and Spanish expressions differed to some extent (e.g.,, “to go on holiday” and “ir de vacaciones” rather than “ir en vacaciones”). Correct responses were the target idiomatic expres- sions, partly correct ones could be regarded as acceptable in the context (e.g., “también” (too, in the sense of “also”) as a translation of “too” as a degree adverb: “This blouse is too large and too long”), and incorrect ones contained some errors, for example, “no problemo” instead of “no me importa.” Failure to provide an answer was broadly classified as “avoidance.” The results of Task 2 are presented in Tables 5 and 6. Here it was not enough for a response to be semantically related to the context, even despite some grammatical error (as in the case of “una chocolate grande” above), but the target form had to be retrieved. However, the mobilisation of all linguistic resources was expected to guide the participants to the target expressions (e.g., “tampoco” is like “neither” in English, but not like “też” (too) in Polish). As can be seen, there are more correct answers and less avoidance in the Romance Philology group. The most errors in both groups (71.43% in the English Philology group and 36.84% in the Romance Philology one) were produced in response to the expression “on holiday.” While Spanish uses the preposition “de” (de vacaciones), the Polish and English prepositions are equiva- lent to each other (“na wakacje” and “on holiday” respectively). Errors included, for example, “en vacaciones” (possible influence of English, but in the case of a student of Italian, also of the Italian “andare in vacanza”); “a vacaciones” and “a las vacaciones” were overgeneralisations of the preposition “a” (to) in Spanish (ir a casa (to go home), a la playa (to the beach), etc.), “on vacaciones” was apparently a system shift (a switch to English, perhaps without noticing it) and even the form “vacaciones” without a preposition was observed. Translanguaging as the Mobilisation of Linguistic Resources… 91 Table 5 Percentages of the English Philology students’ responses provided in Task 2 Word/phrase Correct Partly correct Incorrect Avoidance [%] [%] [%] [%] on holiday 14.29 0 71.43 14.29 I adore it 14.29 28.57 14.29 42.86 I don’t mind 14.29 0 14.29 71.43 too 14.29 28.57 0 57.14 long 0 14.29 14.29 71.43 olives 0 0 14.29 85.71 shoes 14.29 14.29 14.29 57.14 has written 14.29 28.57 14.29 42.86 Neither do I 14.29 14.29 0 71.43 to watch 42.86 0 0 57.14 the baker’s 0 14.29 0 85.71 Table 6 Percentages of the Romance Philology students’ responses provided in Task 2 Word/phrase Correct Partly correct Incorrect Avoidance [%] [%] [%] [%] on holiday 57.89 0 36.84 5.26 I adore it 52.63 26.32 15.79 5.26 I don’t mind 63.16 5.26 15.79 15.79 too 68.42 26.32 0 5.26 long 73.68 5.26 10.53 10.53 olives 68.42 10.53 5.26 15.79 shoes 89.47 0 5.26 5.26 has written 84.21 5.26 5.26 5.26 Neither do I 63.16 0 26.32 10.53 to watch 89.47 0 5.26 5.26 the baker’s 73.68 0 15.79 10.53 Another fairly difficult expression was “A mí tampoco” (Neither do I), where the verb “gustar” required a prepositional phrase (see Hall & Ecke, 2003, above), as shown by the very high avoidance rate (71.43%) and only one correct answer (14.29%) in the English Philology group, and 26.32% of incor- rect answers in the Romance Philology group. Although “tampoco” is used like “neither” in English, one Romance Philology student used a calque from Polish (“a mí también,” like “ja też”); possibly, English was not sufficiently Teresa Maria Włosowicz92 active and thus available as a linguistic resource for her. Four students (one in English Philology and three in Romance Philology) used the wrong syntactic structure (“Yo tampoco,” which would have been possible with a verb other than “gustar”), due to negative transfer from English and/or Polish. One student produced the form “Ni yo” (Me neither). Other sources of difficulty were the expressions “me encanta” (I adore it) and “no me importa” (I don’t mind). In the first case, there was only one correct answer (14.29%) and as much as 42.86% of avoidance among the English Philology students, and only 52.63% of correct answers in the Romance Philology group. “Me encanta” is idiomatic, but like “me gusta,” it has unusual syntactic properties (the subject is the thing that is adored and the person who adores it is the indirect object) and, moreover, the verb “adorar” also exists. Incorrect answers included, for example, “Yo amo” (“I love,” with the direct object pronoun missing), and “Me le gusta!” (syntactically odd, as it contains two indirect objects, which might be literally translated as “*It pleases me him”). Such answers as “Me gusta” (“I like it”; possible but weaker than “I adore it”), “Lo amo” (I love it) and “Lo adoro” (I adore it) were classified as “partly correct,” and partly correct answers were relatively frequent in both groups (26.32% and 28.57%), in comparison to other items. “No me importa” uses the same structure: while in English the subject is the person who does not mind something, in Spanish the subject is the activity and the person is the indirect object. However, the response “no me molesta” ([it] does not disturb me) was also accepted. In the English Philology group, there was a high avoidance rate (71.43%), only one correct answer and an incorrect one (“No problemo”; possibly a communication strategy based on the English “No problem,” but formulated as if it were a verb: “No problemo cuidar a tus niños” – “*I don’t problem to look after your children”). In the Romance Philology group, the number of correct responses was quite high (63.16%), but the avoidance rate was relatively high for that group (15.79%) and there were also three errors (10.53%): “Ni pienso” (I have no intention, I do not even think of; possibly due to a problem with understanding the English expression), “No tengo ganas de” (I do not feel like [doing]), and “No pienso en” (I do not think of…), which may have been similarly motivated, as well as one partly correct answer: “Puedo” (I can), which signalled agreement and could thus be accepted as similar in meaning. The results obtained by both groups were compared by means of a chi- square test. The difference was statistically significant at p < 0.001 (df = 3), which confirms that the results depended on the philology studied and that the Romance Philology group performed significantly better. Finally, Task 3 consisted in correcting errors in Spanish sentences or indi- cating that a sentence was correct. A correct answer involved both finding the error and correcting it, or justifying why the sentence was incorrect. An answer Translanguaging as the Mobilisation of Linguistic Resources… 93 was classified as partly correct if the error was found (the student marked the sentence as incorrect) but no correction or justification was provided, or the justification was not fully correct (i.e., the student intuitively knew that the sentence was incorrect, but had difficulty identifying the rule). An incorrect answer was either the acceptance of an incorrect sentence as correct, or the rejection of a correct sentence as incorrect. Here, unlike in the previous tasks, where translanguaging as the mobilisation of one’s whole linguistic resources was meant to facilitate the task, the co-activation of several languages was meant to lead to errors, as the erroneous sentences contained negative transfer from English, Polish, or both. The percentages of both groups’ responses to Task 3 are given in Tables 7 and 8. Table 7 Percentages of the English Philology students’ responses provided in Task 3 Sentence Correct Partly correct Incorrect Avoidance [%] [%] [%] [%] 1 14.29 14.29 28.57 42.86 2 28.57 0 28.57 42.86 3 28.57 0 28.57 42.86 4 14.29 0 71.43 14.29 5 42.86 0 14.29 42.86 6 42.86 0 0 57.14 7 28.57 0 28.57 42.86 8 42.86 0 14.29 42.86 9 71.43 0 0 28.57 10 42.86 0 14.29 42.86 Table 8 Percentages of the Romance Philology students’ responses provided in Task 3 Sentence Correct Partly correct Incorrect Avoidance [%] [%] [%] [%] 1 42.11 15.79 31.58 10.53 2 73.68 5.26 15.79 5.26 3 21.05 5.26 68.42 5.26 4 63.16 0 31.58 5.26 5 68.42 0 26.32 5.26 6 63.16 5.26 26.32 5.26 7 73.68 0 15.79 10.53 8 57.89 0 31.58 10.53 9 89.47 0 5.26 5.26 10 78.95 0 10.53 10.53 Teresa Maria Włosowicz94 As the results show, the easiest item for both groups was Sentence 9, “¿Hay un banco cerca de aquí?” (Is there a bank near here?), which was correct and was accepted as such by 89.47% of the Romance Philology and 71.43% of the English Philology students. By contrast, the most difficult item was Sentence 3, “*El sábado vi Susana con su novio” (On Saturday I saw Susan with her boyfriend), where the target structure required the preposition “a” because Susana is a human being (El sábado vi a Susana con su novio), a marked rule which is specific to Spanish. Therefore, translating the sentence into Polish or English to be sure of one’s comprehension, to relieve working memory, etc. (Kern, 1994), actually hindered the identification of the error. Similarly, Sentence 4 “*Barcelona es en el noreste de España” (Barcelona is in the North-East of Spain; for the locations of objects, including cities, the verb “estar” should be used, that is, “Barcelona está en el noreste de España”) involved a marked distinction that is specific to Spanish and to Portuguese (though in Portuguese the locations of cities are used with the verb “ser” and “estar” is used for less permanent locations, for example, objects that are in a room, so transfer from Portuguese would have been negative here), but which does not exist in Polish, English or French, and the distinction between “es- sere” and “stare” in Italian might be misleading, as they can sometimes be used interchangeably. Thus, if one translated the sentence into L1 or L2, the sentence would be perceived as correct, which was most probably the case, as the participants produced 71.43% of incorrect answers in the English Philology group and 31.58% of incorrect answers in the Romance Philology group. In Sentence 1, “*Es un cuarto después de las cinco” (literally: “It is a quarter past five,” where the target structure was: “Son las cinco y cuarto”), negative transfer from L1 Polish (Jest kwadrans po piątej) and, possibly, English, with a preposition instead of the conjunction “y” (and) is quite visible. Six Romance (31.58%) and two English Philology (28.57%) students accepted it as correct; however, in the English Philology group there was a fairly high avoidance rate (42.86%). However, it is possible that such avoidance was an example of negative transfer which occurs in a situation where no similarity is perceived, so reliance on L1 as a point of reference prevents learners from using the L2 structure (Ringbom, 1987, p. 50). Possibly, as the students were not sure if the Spanish structure was similar to the Polish and the English ones, they chose to avoid marking it as correct or incorrect. However, as a partly correct answer, a student justified her rejection of the sentence as follows: “Cinco (five) is singular (el cinco),” as if the preposition “después” were correct. In fact, hours, except one o’clock, are plural (“es la una”—it is one o’clock, but “son las dos”— it is two o’clock, etc.), so the rule had been incompletely acquired. Another item which proved relatively difficult was Sentence 8, “*A mí también no me gustan las salchichas” (I do not like sausages either; the er- ror was like the English “*I do not like sausages too”), which was similar to Translanguaging as the Mobilisation of Linguistic Resources… 95 Sentence 8 in Task 2, with the exception that Task 2 involved the translation of explicitly given English stimuli, while in Task 3 the stimuli were in Spanish and translanguaging, as the mobilisation of multilingual resources was more implicit at the level of processing rather than production. As in Task 2, the influence of L1 Polish was probably quite strong, and if the participants trans- lated the sentence into L1, they could not find the error (cf. Toribio, 2001, p. 226), so the activation of multilingual resources led to interference rather than facilitation. The results obtained by both groups were then compared by means of a chi-square test. At df = 3, p < 0.001, which indicates that the difference is statistically significant and that the students’ performance depended on their field of study. Last but not least, the responses to the questionnaire were analysed, from the point of view of the learners’ perception of the influence of English and their L1s on their Spanish, the difficulty of the tasks they had just completed, and cross-linguistic interaction during the tasks. In general, they did not perceive Spanish as very difficult (mean = 3.042, SD = 0.91), though it was more difficult for the English Philology (mean = 3.833, SD = 0.983) than for the Romance Philology students (mean = 2.778, SD = 0.732). As for the dif- ficulty of the tasks, they do not seem generally difficult (mean = 2.647), but the high standard deviation (1.367) shows considerable differences between the participants. Indeed, the tasks were perceived as more difficult by the English Philology (mean = 4.5, SD = 0.5774) than the Romance Philology students (mean = 2.077, SD = 0.954). As for the difficulty of the particular tasks, the same tendency can be observed. Task 1 was, on average, of medium difficulty (mean = 3.167, SD = 1.129), but it was more difficult for the English Philology students (mean = 4.286, SD = 0.756) than for the Romance Philology students (mean = 2.706, SD = 0.92), and, similarly, Task 2 (mean = 2.56, SD = 1.583 for both groups, 4.429, SD = 0.787 for the English Philology group and 1.833, SD = 1.15 for the Romance Philology group) and Task 3 (mean = 3.24, SD = 1.393 for both groups, 4.714, SD = 0.756 for the English Philology students, and 2.667, SD = 1.138 for the Romance Philology ones.) However, the participants’ answers concerning the subsystems of the Spanish language (grammar, vocabulary, pronunciation, idiomatic expressions, spelling and, possibly, something else) they found particularly difficult were compared by means of a chi-square test and the difference did not prove statistically significant ( p = 0.127 at df = 5). Thus, even though the Romance Philology group, being more advanced in Spanish, performed significantly better and found the tasks easier, there was no difference between the areas of difficulty perceived by both groups. As for the influence of L1 on Spanish, the responses of both groups (no influence, some similarities, misleading differences, a tendency to translate Teresa Maria Włosowicz96 literally, the use of both similarities and differences to facilitate learning and avoid errors, and “something else”) were compared, using a chi-square test. The difference was not statistically significant ( p = 0.0477, df = 5), so the influence of the native language on Spanish did not depend on the philology studied, nor on the level of proficiency in Spanish (which was higher in the Romance Philology group, as shown by their performance on the tasks). It might be argued that the participants’ native language was the same (with only two exceptions), but the influence of English on Spanish was also compared by means of a chi-square test and the difference was not significant either ( p = 0.0556, df = 5). Therefore, even though the influence of English might be supposed to be stronger in the case of the English Philology students, the difference is not statistically significant. Finally, the forms of cross-linguistic interaction (CLIN, an umbrella term for transfer, interference, borrowing, code-switching, etc., Herdina & Jessner, 2002, p. 29) perceived by the participants during the tasks were compared by means of a chi-square test. The types of CLIN taken into consideration were: negative transfer from L1, negative transfer from English, interference between L1 and Spanish, interference between English and Spanish, interfer- ence between another language and Spanish, problems with identifying the Spanish words (i.e., given in the box), problems with recalling Spanish words because of the activation of their English equivalents, difficulty finding errors because, as the sentences were translated into L1 or into English, they seemed correct, and “something else”, which covered any other possible difficulties, not included in the list. Again, the difference was not statistically significant ( p = 0.241, df = 9). Therefore, even though English Philology students might be supposed to experience a stronger influence of English on their Spanish, the difference between the perception of CLIN in both groups was not statistically significant. It is possible that, if multilingual repertoires are activated in their entirety (except dormant languages, such as German in the participants’ case), all the active languages participate in the processing, even if they are not very closely typologically related; rather, the native language and a well-known L2, such as English, can influence a third or additional language quite strongly, both as a point of reference (translation into L1 and, possibly, consulting the corresponding L2 rule) and a supplier of lexical items. Conclusions To answer the research questions, it can be observed, first, that translan- guaging was used by both groups, but not always in an overt way. While it Translanguaging as the Mobilisation of Linguistic Resources… 97 was expected that code-switching, as a form of translanguaging, would be used in Task 1 as a communication strategy, it occurred less often than expected (thirteen switches in the English Philology group and two in the Romance Philology group) and did not always result in acceptable sentences (cf. “*¿Qué le want?”). In general, the Romance Philology students performed better, both because of a higher level of proficiency in Spanish and probably also because of higher metalinguistic awareness regarding the vocabulary and structure of the Romance languages. Both their switches into English were correct and may have resulted from the temporary unavailability of the target Spanish words: “Es muy simple” (It is very simple; target: Es muy sencillo/fácil) and “Hay ticket machines en todas las estaciones” (There are ticket machines at all the stations; the target word was either “taquilla” (ticket office) or “expendedor automático de billetes” (ticket machine)). It might be surprising that they avoided switches into other Romance languages, but it might be supposed that keeping the lan- guages separate was a conscious strategy in order to minimise interference, which suggests considerable language awareness on their part. However, their use of code-switching may also have been influenced by the instructions: they may have followed the written instructions, originally intended for the English Philology students, which encouraged switches into English, even though they had been explicitly told to switch into other Romance languages as well. In fact, both interpretations are possible: they may have decided to keep Spanish separate from the other Romance languages and, if necessary, to rely on English as a source of lexical items. On the other hand, more subtle forms of translanguaging can also be as- sumed to have taken place. In particular, translation into L1 and, possibly, also consulting the corresponding English (and maybe, e.g.,, Italian or Portuguese) rule to support the decision, could be regarded as a form of translanguag- ing, which can also be used to facilitate understanding (Lewis et al., 2012), although in the case of different rules, its result could be negative transfer and accepting an incorrect sentence as correct. In the English-Spanish translation task (Task 2), the English expressions served as prompts and, especially in the Romance Philology group, resulted in correct translations. In fact, judging by the amount of avoidance in the English Philology group, the students had dif- ficulty retrieving the Spanish words and expressions, so the prompts did not help them much, but, except in the case of “on holiday/de vacaciones,” they did not result in negative transfer either. Second, as for the participants’ multilingual repertoires and the place of Spanish in them, it can be stated that fully unitary competence cannot be as- sumed, as there was indeed some internal differentiation between the languages. On the one hand, several languages were co-activated and participated in the processing, including Polish, English, Spanish and, in the case of the Romance Philology students, possibly their other Romance languages, and the students Teresa Maria Włosowicz98 used their multilingual resources, so the Dual Competence Model, assuming “fully discrete, non-overlapping linguistic systems” (MacSwan, 2017, p. 180) cannot be confirmed. On the other hand, the relative avoidance of code- switching and reliance on more implicit forms of translanguaging, such as mental translation, indicates that the participants were aware of the boundaries between the “named languages” (Otheguy et al., 2018, p. 2). In contrast to members of multilingual societies where translanguaging is the norm (cf. Li, 2018, cited above), the students could to some extent control their production and avoid the use of certain languages, for example, to reduce interference, but at the same time, the impossibility of deactivating “active” (cf. Green, 1986) languages completely supports Grosjean’s (2001) observation that one is never in a fully monolingual mode. Certainly, the place of Spanish differs in the participants’ multilingual repertoires, depending, in particular, on the philology studied and the level of proficiency in Spanish, though these factors are largely interdependent, as the Romance Philology students had higher proficiency in Spanish and, possibly, also higher motivation for studying Spanish, as it was more closely related to their degree course. This resulted in a higher level of language awareness, also in the use of code-switching into English: they switched into English only twice, but both their switches were correct, while the English Philology stu- dents used switches into English which were either correct, for instance, “un kilo de (potatoes) ,” “¿Tiene (change) de cincuenta euros,” or “Es muy (close) ,” or acceptable: “¿Algo (anything else)?” (put in parentheses by the students to indicate the switches, as suggested by the instructions, though not all switches were put in parentheses), or syntactically incompatible (though to some extent comprehensible), for example: “¿Qué le (want)?” In fact, as mentioned above, some of the English Philology students’ responses indicate that they failed to understand the Spanish sentences. However, it can be stated that there are some connections between the languages in their multilingual repertoires, not only between Spanish and Polish (mostly L1), but also between Spanish and English. Consequently, translanguaging as the mobilisation of all of one’s language resources can be regarded as a natural phenomenon which should be capitalised on as a way of facilitating foreign language learning and communication, but at the same time, it should be combined with developing language awareness and strategic competence. Indeed, though contrary to “the monolingual principle” (Howatt, 1984, as cited in Cummins, 2008, p. 65) of language teaching, in some situations it can be a more effective communication strategy than a search for a target language word, ending in “message abandonment” (Faerch & Kasper’s (1983, p. 52) term) if the intended word is not found. Translanguaging as the Mobilisation of Linguistic Resources… 99 References Arabski, J. (2002). Learning strategies of L1, L2 and L3 lexis. In J. Arabski (Ed.), Time for words. Studies in foreign language vocabulary acquisition (pp. 209–215). Frankfurt am Main–Berlin–Bern–Bruxelles–New York–Oxford–Wien: Peter Lang. Baker, C. (2011). Foundations of bilingual education and bilingualism (5th ed.). Clevedon, UK: Multilingual Matters. Cedden, G., & Saǧin Şimşek, Ç. (2014). The impact of a third language on executive control processes. International Journal of Bilingualism, 18(6), 558–569. Cenoz, J. (2000). Research on multilingual acquisition. In J. Cenoz & U. Jessner (Eds.), English in Europe: The acquisition of a third language (pp. 39–53). Clevedon–Buffalo–Toronto– Sydney: Multilingual Matters. Cieślicka, A. (2000). The effect of language proficiency and L2 vocabulary learning strategies on patterns of bilingual lexical processing. Poznań Studies in Contemporary Linguistics, 36, 27–53. Cook, V. (1991). The poverty-of-the-stimulus argument and multi-competence. Second Language Research, 7(2), 103–117. Cook, V. (1996). Competence and multi-competence. In G. Brown, K. Malmkjaer, & J. Williams (Eds.), Performance and competence in second language acquisition (pp. 57–69). Oxford: Oxford University Press. Cook, V. (2016). Premises of multi-competence. In V. Cook & W. Li (Eds.), The Cambridge handbook of linguistic multi-competence (pp. 1–25). Cambridge: Cambridge University Press. Coste, D., Moore, D., & Zarate, G. (1997). Compétence plurilingue et pluriculturelle. Langues vivan- tes. Vers un Cadre Européen Commun de référence pour l’enseignement et l’apprentissage des langues vivantes : études préparatoires. Strasbourg : Éditions du Conseil de l’Europe. Cummins, J. (1981). The role of primary language development in promoting educational suc- cess for language minority students. In California State Department of Education (Ed.), Schooling and language minority students: A theoretical framework (pp. 3–49). Los Angeles, CA: Dissemination and Assessment Center California State University. Cummins, J. (2008). Teaching for transfer: Challenging the two solitudes assumption in bi- lingual education. In J. Cummins & N. H. Hornberger (Eds.), Encyclopedia of language and education, 2nd edition, Volume 5: Bilingual education (pp. 65–75). Boston: Springer Science + Business Media LLC. De Angelis, G. (2005). Multilingualism and non-native lexical transfer: An identification prob- lem. International Journal of Multilingualism, 2(1), 1–25. De Angelis, G. (2007). Third or additional language acquisition. Clevedon–Buffalo–Toronto: Multilingual Matters Ltd. De Angelis, G., & Selinker, L. (2001). Interlanguage transfer and competing linguistic systems in the multilingual mind. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), Cross-linguistic in- fluence in third language acquisition: Psycholinguistic perspectives (pp. 42–58). Clevedon: Multilingual Matters, Ltd. Dijkstra, T., & Van Hell, J. (2001). Testing the Language Mode hypothesis using trilinguals. Paper presented at the second International Conference on Third Language Acquisition and Trilingualism, Fryske Akademy, 13–15 September 2001 (CD-ROM publication). Doiz, A., & Lasagabaster, D. (2017). Teachers’ beliefs about translanguaging practices. In C. M. Mazak & K. S. Carroll (Eds.), Translanguaging in higher education. Beyond mono- lingual ideologies (pp. 157–176). Bristol–Blue Ridge Summit: Multilingual Matters. Teresa Maria Włosowicz100 Donovan, M. S., & Bransford, J. D. (Eds.). (2005). How students learn: History, mathematics and science in the classroom. Washington, D.C.: National Academy Press. Duarte, J. (2018). Translanguaging in the context of mainstream multilingual education. International Journal of Multilingualism, 1–16. doi: 10.1080/14790718.20181512607 Dunn, A.L. & Fox Tree, J. E. (2014). More on language mode. International Journal of Bilingualism, 18(6), 605–613. Faerch, C., & Kasper, G. (1983). Plans and strategies in foreign language communication. In C. Faerch & G. Kasper (Eds.), Strategies in interlanguage communication (pp. 20–60). London & New York: Longman. Gabryś-Barker, D. (2005). Aspects of multilingual storage, processing and retrieval. Katowice: Wydawnictwo Uniwersytetu Śląskiego. García, O. (2009a). Education, multilingualism and translanguaging in the 21st century. In A. Mohanty, M. Panda, R. Phillipson, & T. Skutnabb-Kangas. (Eds.), Multilingual educa- tion for social justice: Globalising the local (pp. 128–145). New Delhi: Orient Blackswan. García, O. (2009b). Bilingual education in the 21st century: A global perspective. Oxford, UK: Wiley-Blackwell. García, O., & Otheguy, R. (2019). Plurilingualism and translanguaging: Commonalities and differences. International Journal of Bilingual Education and Bilingualism, 1–19. doi: 10.1080/13670050.2019.1598932 Gibson, M., & Hufeisen, B. (2001). ‘Sich konzentrieren *an’: Prepositional verb errors by second language versus third language learners of German. In S. Björklund (Ed.), Language as a tool. Immersion research and practices (pp. 176–195). Vaasa: University of Vaasa. Grainger, J., & Beauvillain, C. (1987). Language blocking and lexical access in bilinguals. Quarterly Journal of Experimental Psychology: Human Experimental Psychology, 39A, 295–319. Green, D. W. (1986). Control, activation and resource. Brain and Language, 27, 210–223. Green, D. W. (1993). Towards a model of L2 comprehension and production. In R. Schreuder & B. Weltens (Eds.), The bilingual lexicon (pp. 249–277). Amsterdam–Philadelphia: John Benjamins Publishing Company. Grosjean, F. (1982). Life with two languages. An introduction to bilingualism. Cambridge, MA, & London, England: Harvard University Press. Grosjean, F. (1985). The bilingual as a competent but specific speaker-hearer. Journal of Multilingual and Multicultural Development, 6, 467–477. Grosjean, F. (2001). The bilingual’s language modes. In J. Nicol (Ed.), One mind, two languages: Bilingual language processing (pp. 1–25). Oxford: Blackwell. Hall, C. J., & Ecke, P. (2003). Parasitism as a default mechanism in L3 vocabulary acquisi- tion. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), The multilingual lexicon (pp. 71–85). Dordrecht–Boston–London: Kluwer Academic Publishers. Herdina, P., & Jessner, U. (2002). A dynamic model of multilingualism. Perspectives of change in psycholinguistics. Clevedon: Multilingual Matters Ltd. Herwig, A. (2001). Plurilingual lexical organisation: Evidence from lexical processing in L1– L2–L3–L4 translation. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), Cross-linguistic influ- ence in third language acquisition: Psycholinguistic perspectives (pp. 115–137). Clevedon: Multilingual Matters, Ltd. Howatt, A. (1984). A history of English language teaching. Oxford: Oxford University Press. Juffermans, K., Blommaert, J., Kroon, S., & Li, J. (2014). Dutch-Chinese repertoires and language ausbau in superdiversity: A view from digital media. Discourse, Context and Media, 4–5, 48–61. Kern, R. G. (1994). The role of mental translation in second language reading. Studies in Second Language Acquisition, 16, 441–461. Translanguaging as the Mobilisation of Linguistic Resources… 101 Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33, 149–174. Lewis, G., Jones, B., & Baker, C. (2012). Translanguaging: Developing its conceptualization and contextualisation. Educational Research and Evaluation, 18(7), 655–670. Li, W. (2018). Translanguaging as a practical theory of language. Applied Linguistics, 39(1), 9–30. MacSwan, J. (2017). A multilingual perspective on translanguaging. American Educational Research Journal, 54(1), 167–201. Marian, V., & Spivey, M. (2003). Competing activation in bilingual language processing: Within- and between-language competition. Bilingualism: Language and Cognition, 6(2), 97–115. Morytz, J. (2017). Strategie komunikacyjne w obliczu trudności językowych. Neofilolog, 49(2), 193–206. Myśliwiec, M. (2013). Ślōnskŏ godka – przaśny folklor czy język regionalny? Przegląd Prawa Konstytucyjnego, 3(15), 99–120. Otheguy, R., García, O., & Reid, W. (2018). A translanguaging view of the linguistic system of bilinguals. Applied Linguistics Review, 1–27. doi: 10.1515/applirev-1018-0020 Poulisse, N., & Bongaerts, T. (1994). First language use in second language production. Applied Linguistics, 15(1), 36–57. Ringbom, H. (1987). The role of the first language in foreign language learning. Clevedon: Multilingual Matters. Singleton, D. (2000). Language and the lexicon. An introduction. London: Arnold. Singleton, D. (2001). Cross-linguistic interactions in the multilingual lexicon. Paper presented at the second International Conference on Third Language Acquisition and Trilingualism, Fryske Akademy, 13–15 September 2001 (CD-ROM publication). Singleton, D. (2003). Perspectives on the multilingual lexicon: A critical synthesis. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), The multilingual lexicon (pp. 167–176). Dordrecht–Boston– London: Kluwer Academic Publishers. Toribio, A. J. (2001). On the emergence of bilingual code-switching competence. Bilingualism: Language and Cognition, 4(3), 203–231. Treffers-Daller, J. (2018). Code-switching and translanguaging: Exploring the creativity of mul- tilinguals. Retrieved from: https://www.researchgate.net/publication/322677346 TVS (2019). Język śląski nareszcie językiem regionalnym? Sejm właśnie zdecydował! 13 June. Retrieved from: https://tvs.pl/informacje/jezyk-slaski-nareszcie-jezykiem-regionalnym-se- jm-wlasnie-zdecydowal/ Van Gelderen, A., Schoonen, R., de Glopper, K., Hulstijn, J., Snellings, P., Simis, A., & Stevenson, M. (2003). Roles of linguistic knowledge, metacognitive knowledge and process- ing speed in L3, L2 and L1 reading comprehension. A structural equation modeling ap- proach. International Journal of Bilingualism, 7(1), 7–25. Williams, C. (1994). Arfaniad o dduliau dysgu ac addysgu yng nghyd-destun addysg uwchradd ddwyieithog [An evaluation of teaching and learning methods in the context of bilingual secondary education]. Unpublished doctoral dissertation, University of Wales, Bangor, UK. Williams, S., & Hammarberg, B. (1998). Language switches in L3 production: Implications for a polyglot speaking model. Applied Linguistics, 19(3), 295–333. Włosowicz, T. M. (2013). L’alternance codique dans les productions orales et écrites au sein d’une filiale polonaise d’une entreprise multinationale. Synergies Italie, 9, 129–140. Włosowicz, T. M. (2016). Multilingual processing phenomena in learners of Portuguese as a third or additional language. Theory and Practice of Second Language Acquisition, 2(1), 65–86. Włosowicz, T. M. (in preparation). The attrition and retention of German after long incubation periods. Teresa Maria Włosowicz102 A p p e n d i x 1 The Tasks Used in the Study Task 1: Fill in the gaps in the two dialogues. You can choose words from the box below the dialogues, or use some words of your own. Verbs in the infinitive may need to be put in the right form, and nouns may need to be changed to the plural or preceded by an article if they are in the singular. There are some words you do not have to use. Some words can be used more than once. Sometimes there is more than one possible word that fits the context. (If you really do not know which Spanish word to choose, you can fill in the gap with an English word that fits the context (in brackets), in order to show that you understand the sentence and you know what the target word should mean.) A) Dependienta: Buenos días, ¿qué le ? Clienta: Buenos días. Quería dos , una barra de pan, un kilo de , diez y doscientos gramos de . Dependienta: ¿Algo ? Clienta: Sí, ¿tiene ? Dependienta: Por supuesto, están dulcísimas y muy fresquitas. Clienta: Entonces, póngame un kilo, por favor. ¿Tiene también ? Dependienta: Sí, están aquí. Clienta: Quería medio kilo y también ochocientos gramos de y una grande. Dependienta: Aquí tiene. ¿Algo ? Clienta: Tal vez unos _________________, por favor. ¿Tiene ? Dependienta: No, usted tiene que ir a la carnicería. ¿Algo ? Tengo , , … Todos están riquísimos y fresquitos. Clienta: No, gracias. Eso es todo. Dependienta: Pues son veinticinco euros. Clienta: ¿Tiene de cincuenta euros? Dependienta: Por supuesto. Clienta: Muchas gracias, ¡adiós! Dependienta: Gracias, ¡adiós! B) Turista: Perdone, ¿cómo a la catedral? Recepcionista: Es muy . Al del hotel, tiene que a la derecha y seguir todo hasta la Mayor. Atraviese la Mayor y después la primera a la izquierda. La catedral está allí. Turista: Muchas gracias. ¿Y decirme cómo llegar al del Arte Moderno? Recepcionista: Está bastante , entonces usted tiene que el metro. Tiene que la cinco a la Jardín Zoológico, a Goya y cambiar a la tres Aeropuerto. a la tercera que se llama Picasso. Turista: Gracias. ¿Dónde se puede comprar los ? Recepcionista: Hay en todas las . Turista: ¿Y dónde está la más cercana? Necesito pastillas para el dolor de . Translanguaging as the Mobilisation of Linguistic Resources… 103 Recepcionista: Está en la de Flores. Tiene que la primera a la izquierda y seguir todo . Turista: Muchas gracias, ¡adiós! Recepcionista: De nada. ¡Adiós! atravesar – to cross el aeropuerto – airport Task 2: Translate the words and phrases given in English into Spanish: 1. Todos los años, Ana va (on holiday) al borde del mar. 2. ‘¿Qué tal este vestido?’ ‘¡ ! (I adore it!) Me lo llevo.’ 3. (I don’t mind) cuidar a tus niños. 4. Esta blusa es (too) ancha y (long). 5. Voy a poner (olives) en la ensalada. 6. Pruébate estos (shoes), Alicia. 7. Hoy Paula (has written) cinco cartas. 8. ‘A mí no me gusta la carne.’ (Neither do I.) 9. A Juana le gusta (to watch) televisión en el tiempo libre. 10. Ya no tenemos pan. Tengo que ir a la (the baker’s). Task 3: Are the sentences below correct or incorrect? If they are incorrect, please, explain why and/or suggest a correction. Sentence Correct Justification Yes No 1. Es un cuarto después de las cinco. 2. ¿Comiste alguna vez paella? 3. El sábado vi Susana con su novio. 4. Barcelona es en el noreste de España. 5. Elena le ha regalado a Ernesto una bicicleta. 6. ‘¿Has contado tu aventura a Luisa ?’ ‘No, no le la he contado.’ 7. Margarita acuesta a los niños a las nueve. 8. A mí también no me gustan las salchichas. 9. ¿Hay un banco cerca de aquí ? 10. ¿Qué te parece este vestido para mí? el noreste – the North East la bicicleta – bicycle sencillo pescado bajar ir poner huevo Plaza tomate plátanos arroz coger avena taquilla chorizo chocolate salchicha sandía lejos yogur salir tomar billete manzana piña uva calcetín colif lor fresa patata más pollo dirección Museo cambio pimiento pimienta Teresa Maria Włosowicz104 A p p e n d i x 2. The Questionnaire Used in the Study QUESTIONNAIRE Sex : F /M L1 (native language): L2: Level/time of study: L3: Level/time of study: L4: Level/time of study: What other languages have you studied? Please, indicate the levels. 1a) How difficult do you find the Spanish language? (1 – very easy, 5 – very difficult) 1 2 3 4 5 What do you find particularly difficult to learn? (You can choose more than one answer.) □ grammar □ vocabulary □ pronunciation □ idiomatic expressions □ spelling □ something else (please, specify) 1b) What is the influence of your English on your Spanish? (You can choose more than one answer.) □ I do not notice any inf luence at all. □ There are some similarities (e.g. the difference between the Present Perfect Tense and the Pretérito Perfecto on the one hand and the Past Simple Tense and the Pretérito Indefinido on the other) that make learning Spanish easier for me. □ There are a lot of misleading differences that make Spanish difficult for me. If so, please, give examples: □ I tend to translate sentences literally from English into Spanish. □ That depends: there are both similarities and differences and I use them both to facilitate learning and avoid errors. □ something else (please, specify) 1c) What is the influence of your native language on your Spanish? (You can choose more than one answer.) □ I do not notice any inf luence at all. □ There are structures that do not exist in my native language, that is why I find Spanish difficult to learn. □ There are some similarities between Spanish and my native language (e.g. desayunar in Spanish and завтракать in Russian) which make those Spanish words and structures easier to learn. □ There are a lot of misleading differences that make Spanish difficult for me. Translanguaging as the Mobilisation of Linguistic Resources… 105 If so, please, give examples: □ I tend to translate sentences literally from my native language into Spanish. □ That depends: there are both similarities and differences and I use them both to facilitate learning and avoid errors. □ something else (please, specify) 2a) How difficult do you find the task you have just completed? (1 – very easy, 5 – very difficult) 1 2 3 4 5 How difficult was Task 1 (filling the gaps)? 1 2 3 4 5 How difficult was Task 2 (translating words from English into Spanish)? 1 2 3 4 5 How difficult was Task 3 (error correction)? 1 2 3 4 5 Why? □ I had forgotten a lot of words during the summer holidays. □ I had forgotten a lot of grammar structures during the summer holidays. □ Interference from English was too strong. □ Interference from my native language was too strong. □ Some of the words or structures were new to me. □ I did not understand the sentences well. □ for another reason (please, specify) 2b) What forms of cross-linguistic interaction did you notice during the tasks? (You can choose more than one answer.) □ Negative transfer from my native language. (I relied e.g. on the literal translation of phrases and later I noticed it had led to errors.) □ Negative transfer from English. □ Negative transfer from another language (e.g. French). □ Interference between my native language and Spanish. (I made some mistakes and only later did I realise that they were due to my native language.) □ Interference between English and Spanish. □ Interference between another language and Spanish. □ I could not identify the right Spanish words, that is why I only gave English words in brackets. □ I could not recall the right Spanish words because their English equivalents were constantly on my mind. □ I had difficulty finding the errors in the Spanish sentences, because as I translated them into my native language, they seemed to be correct. □ I had difficulty finding the errors in the Spanish sentences, because as I translated them into English, they seemed to be correct. □ something else (please, specify) Thank you. Teresa Maria Włosowicz106 Teresa Maria Włosowicz Translingulismus als Mobilisierung von Sprachressourcen durch Personen, die Spanisch als dritte oder zusätzliche Sprache lernen Z u s a m m e n f a s s u n g Den Gegenstand des Artikels bildet die Untersuchung des Verstehens und der Sprachproduktion im Spanischen als dritte oder zusätzliche Sprache (der Begriff wurde von De Angelis im Jahre 2007 eingeführt), wobei besonderes Augenmerk auf die Verwendung von Code-Switching und Translingualismus (translanguaging) gelegt wird. In Anlehnung an Lewis, Jones und Baker (2012, S. 655) wird angenommen, dass der Translingualismus die Mobilisierung aller sprachlichen Ressourcen eines Lernenden umfasst, „um das Verstehen und die Leistungen zu maximieren“. Aus diesem Grund kann die Verwendung anderer Sprachen als Spanisch (insbesondere Englisch, aber auch z. B. Französisch, Italienisch u. ä.) durch die Teilnehmer in Aufgaben ebenfalls als Beispiel für Translingulismus angesehen werden. Der Gebrauch von Wörtern aus anderen Sprachen als Spanisch könnte gleichzeitig als Code- Switching eingestuft werden. Die mehrsprachigen Sprachrepertoires sind sehr komplex und die mehrsprachige Kompetenz ist – laut Otheguy, García und Reid (2018) – eher einheitlich und nicht in einige separate Sprachen unterteilt. Ihrer Meinung nach werden die Wörter aus einem mentalen Lexikon ausgewählt. Wie Williams und Hammarberg (1998) nachgewiesen haben, erfüllen verschiedene Sprachen unterschiedliche Funktionen in mehrsprachigen Repertoires, was zu unterschiedlichen Formen von Code-Switching führt. Die vorliegende Studie wurde unter Studierenden der englischen und romanischen Philologie durchgeführt, die Spanisch als dritte oder zusätzliche Sprache lernen. Wie die Ergebnisse zeigen, vermieden die Studierenden der Romanistik das Code-Switching in andere romanische Sprachen, um wahrscheinlich die Interferenz zu minimieren, obwohl sie im Allgemeinen besser im Translingualismus wa- ren, der als Nutzung all ihrer Sprachressourcen verstanden wurde. Andererseits waren die Studierenden der englischen Philologie, deren Kompetenzniveau im Spanischen niedriger war, weniger dazu bereit, mehrsprachige Ressourcen, auch im Englischen, zu verwenden, um feh- lende Wörter zu ergänzen, was möglicherweise auch aus Problemen mit dem Verstehen von Sätzen im Spanischen resultierte. Schlüsselwörter: mehrsprachige Repertoires, Sprachressourcen, Translingualismus, Strategien der Sprachproduktion