Accommodation and Language Contact

4.1 Accommodation in L2 and heritage speakers

The first five papers focus on accommodation in L2 and heritage speakers, a relatively understudied aspect of linguistic accommodation that has been vastly investigated in dialogic situations between partners sharing the same language variety. As discussed by Grace Wenling Cao in her paper Phonetic Dissimilarity and Category Formation in L2 Accommodation, L2 accommodation differs from L1 accommodation in motivation and patterns of convergence (which can be affected by different factors such as novelty, atypicality, recency, distraction and gender), and crucially faces three important challenges. The first one is the need to cope with patterns of accommodation that are highly heterogeneous: besides huge inter-speaker and intra-speaker variations in L2 speakers’ accents, the degree of phonetic convergence can be affected by phonetic talent and personality (Lewandowski & Jilka, 2019) as well as the L2 experience of the learners. The second one is represented by the intertwining of L2 accommodation with the issue of acquiring a new L2 phonological category, a scenario that speakers do not have to face when accommodating in their L1. The third challenge is the lack of L2-related theories of accommodation in the field of speech accommodation. Models that are often used to explain the accommodation of native speakers (CAT, IAM, some exemplar-based models, such as Johnson, 1997, 2006; Pierrehumbert, 2001, 2002) might be able to explain L2 accommodation of a similar category for high-proficiency L2 speakers, but they might not be able to account for L2 accommodation of a new category which involves phonological acquisition difficulties.

Cao recognises the need to distinguish between accommodation to similar or new categories and attempts to answer the question of whether patterns of L2 accommodation can be predicted by the L2 category formation hypothesis of the revised Speech Learning Model (Flege & Bohn, 2021). A strictly related question addressed in the paper refers to the role of cross-L2 dialect (dis)similarity in L2 accommodation, specifically whether a larger cross-L2 dialect phonetic dissimilarity would predict a stronger degree of L2 accommodation. The paper tested bilingual Hong Kong English (HKE) speakers’ short-term accommodation towards Received Pronunciation (RP) and General American English (GenAmE). Hong Kong speakers have difficulty acquiring English fricatives /z/ and /th/, which are not present in the phonological system of their native Cantonese. Such phonological difference allows testing the patterns of L2 accommodation to a new category. On the other hand, the role of phonetic (dis)similarity in L2 accommodation is assessed by analysing the patterns of accommodation in the vowels of the THOUGHT and PATH sets, which are phonetically similar in Hong Kong English and RP but different in Hong Kong English and GenAmE. By taking into account both the direction and the distance of shifting, Cao shows that for both investigated vowels, HKE speakers converged to RP and GenAmE interlocutors similarly, not supporting the hypothesis that HKE speakers would converge more towards the language variety that is globally more different from the native one (GenAmE). However, when looking at the local phonetic dissimilarity, HKE speakers accommodate more towards the interlocutor the larger the phonetic distance between their vowels. Results on the accommodation for English fricatives show that the patterns of accommodation crucially depend on whether the bilingual HKE speakers had already formed the relative L2 categories or not, a result which is unexpected based on the revised Speech Learning Model. Results only partially confirm the role of phonetic dissimilarity in convergence and fail to support the prediction that only fully or partially formed L2 categories can induce convergence. To explain L2 accommodation, Cao proposes a new exemplar model that integrates L2 learning models and exemplar-based theories and takes into account the important role of speech perception and attention weight adjustment.

One aspect of L2 accommodation that emerged in the study by Cao and was also observed in other studies (e.g., Pardo et al., 2017) is the high individual variability of the phenomenon. The paper Phonetic Accommodation on the Segmental and the Suprasegmental Level of Speech in Native–Non-Native Collaborative Tasks by Christiane Ulbrich investigates individual differences in L2 accommodation by relating them to the more general question of whether speech accommodation is driven by social motivations that make it a conscious voluntary mechanism (as in CAT) or is the product of an involuntary mechanism (as in IAM). She analyses accommodation in quasi-spontaneous conversations, in which dyads of speakers are engaged in the completion of a map task where a participant takes the role of a tourist and a control speaker that of a guide or vice versa. Crucially, the participants had different levels of L2 proficiency: speakers of Castilian Spanish with high or low proficiency in L2 German were paired up with three control speakers who were either native speakers of Northern Standard German or Spanish speakers with high or low proficiency in the same variety of German. Granted that the aim of non-native speakers engaged in conversations with native ones is to achieve a high level of intelligibility, the study focuses on the role of L2 proficiency and asks whether “insufficient proficiency impedes the application of accommodation strategies to affiliate with native speakers of their L2 and for automatisms to kick in.” The variables taken into consideration are of segmental and prosodic nature, two domains in which Spanish and German differ. Results show that both segmental and suprasegmental characteristics are accommodated by non-native speakers of German in conversation with both native and non-native interlocutors, but not homogeneously. The proficiency of both the speaker and the interlocutor matters: highly proficient Spanish speakers accommodate to both the native German control and the high-proficiency Spanish control on both segmental and prosodic features, while low-proficiency Spanish speakers accommodate only to high-proficiency Spanish control and only for prosodic features. High-proficiency speakers, though, also accommodate to low proficiency ones by producing fewer cases of segments’ reductions, thus lowering their articulation rate, in a conscious effort to maintain intelligibility and facilitate the conversation. Overall, Ulbrich’s findings point to a balance of conscious accommodation driven by similarity affiliation effects, and the extra demands that L2 production imposes on the L2 speakers, which can block possible alignment.

The problem of individual variability is tested in both the L1 and L2 of the same speaker in the paper The language-specificity of phonetic adaptation to talkers by Anne Cutler, Ann Burchfield, and Mark Antoniou. Their paper addresses another important topic in L2 accommodation: that is, if it is language-specific. In most instances, coping with new talkers seems quite unproblematic, and human listeners adapt rapidly to talkers they have never encountered before. One of the ways in which this adaptation is achieved is the deployment of lexical knowledge to resolve a phonemic ambiguity, a process known as lexically-guided perceptual learning. The specific question Cutler et al. try to answer is whether lexically-guided perceptual learning occurs as freely in L2 as in L1. In their experimental paradigm, L1 and L2 perceptual learning is compared in the same listeners, with the advantage of controlling for the individual differences that have been observed in the process of learning about individual talkers. Adaptation was tested in both the L1 and L2 of two groups of bilingual speakers: a group with L1 Mandarin and L2 English, and a group with L1 English and L2 Mandarin. English and Mandarin were selected as they are very dissimilar languages differing in phoneme repertoire size, syllable structure, and morphology, as well as in the presence of lexical tone contrasts. In three experiments, an ambiguous sound midway between /f/ and /s/ replacing either /f/ or /s/ in Mandarin and English words was presented for lexical decision to both groups of bilingual listeners. Results show that the English-native group displayed learning in both L1 and L2, while the Mandarin-native one displayed learning only in L1 but not in L2. The authors conclude that perceptual accommodation to the speech patterns of individual talkers can be language-specific, as the listener’s ability to adapt rapidly in one language does not necessarily imply that the same will be true of every language in the same listener’s repertoire.

The way the sound systems accommodate each other within an individual is discussed with reference to the variable susceptibility of different segments to phonetic convergence by Khia Johnson and Molly Babel in the paper entitled Language Contact Within the Speaker: Phonetic Variation and Crosslinguistic Influence. The authors address the issue of the asymmetry in the crosslinguistic influence in early bilingual speakers by bringing together well-known concepts in the investigation of bilingual speech and predictions from proposals on sound change (specifically Harrington et al., 2018; Harrington & Schiel, 2017). They posit into question if the direction of change in bilingual speech relates to the distribution of variation in the languages that are in contact with the linguistic system of the individuals, specifically, whether a less variable language constraints variation in a more variable language. The investigation regards obstruents (/b d g/ and /t k/) produced in English by early Cantonese–English bilinguals. Obstruent voicing and release differ in Cantonese and English, but the segments involved are linked (e.g., the voiced/unaspirated stops are linked by their similarities in onset position, where in Cantonese they can both surface as short-lag stops). Crucially, in both stop voicing and unvoiced stop release, Cantonese exhibits relative invariance in comparison to English. The paper reports on two studies on phonetic variation in stop categories in English: the first regards sub-phonemic variation in the realisation of voiced stops, whereas the second investigates the release in voiceless final stops. Both studies are carried out by analysing three spontaneous speech corpora: The SpiCE corpus of bilingual speech in Cantonese and English, two monolingual corpora involving speakers that do not have an early multilingual experience (a collection of Interviews of British Columbia English speakers, and the Buckeye Corpus of Conversational Speech). In the first study, the analysis of bilingual semi-spontaneous speech showed that English /b d g/ in intervocalic position were produced with less voicing by Cantonese–English bilinguals than by English monolinguals. The second study showed that word-final English /t k/were more likely to be unreleased in bilingual productions in comparison with the productions in two English monolingual corpora. Overall, results provide evidence that the phonetic variation in cross-linguistically linked categories in early bilinguals is affected by the distribution of phonetic variation, in that Cantonese (less variable language) was shown to constrain the variation in English (more variable) as spoken by early Cantonese–English bilinguals. Based on these results, the authors propose a “mechanistic account” for the observed differences in the susceptibility of segments to cross-language influence. Noticeably, some speakers in the SpiCE corpus may be considered as heritage speakers, as they were raised in an English-dominant society and primarily used Cantonese within a family context.

Accommodation in heritage speakers is explicitly addressed in the paper by Chiara Celata and Naomi Nagy, Sociophonetic variation and change in heritage languages: Lexical effects in Heritage Italian aspiration of voiceless stops. The paper is a follow-up of a previous study in which the authors studied the aspiration of voiceless stops in Heritage Calabrian regional Italian and English across three generations of Italians living in Toronto. Both the heritage and the host language exhibit long-lag VOT voiceless stops, but they differ both in the phonological distribution of the feature and in their sociophonetic value: only in Calabrian Italian stop aspiration indexes information about speaker characteristics and sociocultural features. Results showed that the transmission of sociophonetically relevant features across generations differs from the transmission of phonetic features that are not markers of sociolinguistic identity in the homeland variety. Specifically, the first one decreases cross-generationally, but this change was not influenced by external (sociolinguistic orientation or sex) and internal variables (vowel height and consonant place of articulation) in a straightforward way.

In the present paper, Celata and Nagy go further in investigating whether lexical frequency has a role in shaping cross-generational changes. They hypothesise that the two types of aspiration could differ at the lexical level: English-like aspiration, showing a more predictable influence of language-internal and language-external variables, may also be subject to typical frequency effects, with a direct correlation between lexical frequency and rate of aspiration. Calabrian-like aspiration would not be as much due to the progressively more limited speakers’ exposure to the heritage language and its indexical value. Results show that both types of aspiration changed across generations in terms of their lexical distribution but in different ways. The diffusion of a majority language’s phonetic feature into a heritage language is substantially frequency-driven, with a consistent increase across generations in the number of distinct lemmas that are the targets of the phonetic change, while the maintenance or loss of the sociophonetically relevant stop aspiration in the speech of heritage speakers is not: it shows a wider and more even distribution across lemmas and its decrease across generations is not influenced by lexical frequency. The paper thus shows that the consequences of long-lasting internal contact depend on the type of the phonetic variable in contact. Contact-induced phonetic changes that do not affect socio-indexical variables are similar to lexically-gradual sound changes, while those changes that do affect socio-indexical variables do not depend on a process of progressive lexicalisation due to entrenchment into a small set of high-frequency words.

4.2 Accommodation in long-lasting community contact

The other four papers focus on the time dimension of dialect or language contact and show the effects of long-lasting contact on a linguistic code resulting either in sound changes in progress, in mixed languages or the stable integration of segmental or prosodic features in the linguistic system.

In their paper, entitled Fast-Speech-Induced Hypoarticulation Does Not Considerably Affect the Diachronic Reversal of Complementary Length in Central Bavarian, Markus Jochim and Felicitas Kleber discuss a sound change in progress in the Central Bavarian dialect, which is spoken in the south-east of Germany and most parts of Austria. In Central Bavarian, the complementary length between a consonant and the preceding vowel (i.e., phonetically long vowels only occur before lenis-short stops and phonetically short vowels only before fortis-long stops) seems to be gradually replaced by the unrestricted combination of long and short vowels followed by either phonologically longer fortis and shorter lenis stops. The trend goes in the direction of what is observed in Standard German, which does not restrict the combination of vowels and stops, even though in this variety the vowel length contrast is phonemic and the contrast in stops fortis-lenis, which is mainly a length contrast in Central Bavarian, involves also correlates other than duration, such as VOT differences. The authors aim to check (1) if long vowel plus fortis stop sequences emerge in the Western Central Bavarian dialect, (2) if the change extends to short vowel plus lenis stop sequences, and (3) if, besides the role of contact with Standard German, fast-speech-induced hypoarticulation may be an internal driver of the change. Furthermore, they aim to understand the way different cues are affected, that is, how the change is taking place. Change is checked within the apparent-time paradigm, as 10 young and 10 old dialectal speakers are analysed, besides 10 young (20–30 years old) Standard German speakers. Speakers read the target sentences at a faster and slower rate in alternating speech rate blocks. Various measures of a duration-based rather than spectrum-based contrast were considered, involving vowel and consonant duration, with specific measures involving the closure and the aspiration phase in the target word’s stop. Results allow the author to discuss five main findings. First of all, the combination of long vowel plus fortis consonant (VːCː) is also legal in older dialect speakers, in line with previous observations on the regional accent of the area; second, in younger dialectal speakers, the combination of short vowel plus lenis consonant (VC) has become legal. This suggests that in the system of Central Bavarian, the VːCː sequence must have emerged prior to the VC sequence. As for the third goal, the acoustic analysis did not support the hypothesis that fast-speech-induced hypoarticulation could trigger the observed sound change. Further findings concern the VOT, which in younger dialect speakers shows values intermediate between that of older dialect speakers and younger standard speakers, and the variation (both between- and within-speaker) observed in some words more than in others. The results on the VOT suggest that this cue to the lenis–fortis contrast in stops is becoming stronger, while the variation observed in the case of some words suggests that the gradual sound changes may also spread irregularly via lexical diffusion. Investigating word frequency and lexical diffusion, together with the phonological features of the segments involved, is indeed suggested to be a promising follow-up of the investigation. Thus, the sound change in progress investigated in the paper is finally discussed as corresponding to a gradual convergence towards a standard which is directly triggered by the contact situation, rather than also having an internal motivation in speech rate variation.

As discussed by Sky Onosson and Jesse Stewart in the paper The Effects of Language Contact on Non-Native Vowel Sequences in Lexical Borrowings: The Case of Media Lengua, the contact between languages may lead to the emergence of what is considered a new, mixed language. Mixed languages, which are created for expressive purposes rather than for communicative needs by speakers who are proficient bilinguals in the source languages (see also fn.3), usually show systematic splits between the elements of the languages involved. In the case of phonology, however, the assimilation to the phonology of the ancestral language is observed, with the introduction of elements that allow to characterise a new sound system whose features are very difficult to predict. The case under investigation is Media Lengua, spoken by about 2000 people in the northern Ecuadorian province of Imbabura and formed from Quichua and Spanish in the early twenteith century. Previous studies showed that the system may include up to eight vowels that occupy an acoustic space reflecting a Quichua three-vowel system. The acoustic investigation described in the paper regards vowel sequences, which are vowel-like articulations containing two target positions/qualities that may correspond to distinct phonological vowels or diphthongs. The study aims to investigate the hypothesised integration and acoustic arrangement in Media Lengua of vowel sequences of Quichua and Spanish origin. The assumption is that the analysis of phonetic correlates, specifically formant trajectories, of such sequences in Media Lengua, Quichua, and Spanish may help in identifying processes such as assimilation but also gradient differences that may suggest if borrowed sounds conform to the phonological constraints of a new language. The comparison of vowel sequences produced in Media Lengua, Quichua, and Spanish through generalized additive mixed effect models shows that the vowel system of Media Lengua integrates the vowel sequences deriving from Spanish into what is the Quichua phonology. In particular, Spanish-derived Media Lengua vowel sequences differ significantly from their Spanish counterparts and frequently resemble Quichua sequences, occupying a more central region of the vowel space, similar to Quichua, and showing reduced trajectories in comparison to Spanish (and in some cases, they seem to be no longer analysable as sequences of independent vowels). Even in the one single case in which a Media Lengua sequence significantly differs from its Quichua counterpart, the difference from Spanish is greater than that from Quichua. Finally, the authors underline that phonological arrangements in mixed languages are complex (e.g., do not involve clear-cut cases of assimilation) and substantially affected by various factors, such as the age of acquisition of the source languages.

Turning to contributions addressing prosodic issues, the paper Echoes of Past Contact: Venetian Influence on Cretan Greek Intonation by Mary Baltazani, John Coleman, Elisa Passoni and Joanna Przedlacka offers an interesting insight into the impact that language contact may have on intonation systems and the persistence of the influence after three and a half centuries. The authors investigate intonation contours and peak alignment in broad focus declaratives as well as information-seeking polar questions in Cretan Greek, in the dialect and the variety of Italian spoken in the area of Venice, and Athenian (standard) Greek, considered as control. Cretan Greek and the Venetian dialect were in contact for about four centuries and a half, until approximately the second half of the seventeenth century, and the goal of the authors is to look for traces of such contact in speech material collected in the present day. The analysis is based on natural speech corpora collected in the languages/varieties considered, including both spontaneous (e.g., television and radio interviews) and semi-spontaneous speech (e.g., map-task dialogues) and a variety of speech styles (e.g., short reading passages and sentence lists). The functional data analysis of the F0 contours and measures of F0 peak alignment to relevant segmental landmarks show that the Cretan declarative contour is a fall from an H peak aligned with a prenuclear vowel and followed by a low F0 plateau, similar to what found in the varieties spoken in the Venetian area; while the Athenian declaratives show a fall-rise-fall pitch contour, with the H aligned with the nuclear vowel. In addition, the L tone representing the end of the fall is typically aligned after the end of the nuclear vowel in Athenian and with the nuclear vowel in Cretan and Venetian. Even stronger similarities between Cretan and Venetian are found in the analysis of polar questions. Specifically, no significant difference between Cretan and Venetian is found in the alignment of the H peak and (three of the four) shape coefficients, whereas Athenian differs for both the H peak alignment and the same shape coefficients. As the authors underline, similarities between the intonation patterns of Cretan Greek and Venetian declarative and polar question tunes attest the robustness of contact effects almost three and a half centuries after regular contact ceased, and show that in Cretan Greek the influence of Venetian is found besides characteristics that resisted to change. The authors observe that, besides the major contact period, when the populations cohabited on the island, various ties between Venetians and Cretans were maintained even later, playing a role in preserving the contact outcome. Thus, in the discussion section, the authors interestingly observe that their results support the idea of periods of language changes that “are punctuated with periods of stability, as change and development usually come about in specific historical and social circumstances” (see also Nicholls, 2003).

The contribution by Andrea Pešková, entitled Intonational and Syntactic Innovations in a Language Contact Situation: An Explorative Study of Yes/No Questions in Paraguayan Guarani–Spanish Bilinguals, considers the impact of a long-lasting contact situation (about five centuries, since 1537) on the intonation system and its interfaces with syntax, and discusses the case of the ongoing contact and the consequent bilingualism. The bilingualism under investigation is due to the contact between the indigenous Guarani language and Spanish in Paraguay, where nowadays half of the population is bilingual, and most speakers use Guarani or the mixture of Paraguayan Spanish and Guarani in informal settings, and Spanish in formal ones. The paper regards the variety of Paraguayan Spanish and its features in bilingual speakers, who are either Guarani-dominant bilinguals or Spanish-dominant bilinguals. Guarani and Spanish differ in the way questions are realised (question particles only in the former), and the paper aims at understanding whether yes-no questions in the Paraguayan variety of Spanish present Guarani intonational or syntactic features; further, it aims at examining possible differences between Guarani-dominant, Spanish-dominant bilinguals, and Paraguayan Spanish monolinguals. A Discourse Completion Task, including 17 situations that lead to neutral or biased yes/no questions, was performed by Spanish dominant, Guarani dominant, and Spanish monolinguals and was analysed both intonationally and syntactically. Overall, results show that Paraguayan Spanish presents a rather small inventory of nuclear configurations, probably as a result of the overgeneralisation of some F0 patterns. Further, both monolingual and bilingual speakers mainly realized yes/no questions, especially neutral questions, with F0 rising (L*+ H + L*LH%) and less with F0 falling contours. Most important, the former pattern shows a form very similar to a Guarani contour. Various patterns are found in biased questions, besides the rising one. One of them, found in “emphatic” yes/no questions, shows a Spanish-like focal nuclear pitch accent and a low boundary tone which was found in counter-expectational and echo yes/no questions. As for syntactic features, the subject position is similar to that found in Spanish, whereas bilinguals display Guarani lexical elements, such as the question particles, in biased questions and informal contexts (whereas Spanish-speaking monolinguals did not produce any particles); further both monolingual and bilingual groups use the construction Sera que. . .? (which is often used in Spanish questions or declaratives to express future probability) at the beginning of both neutral and biased yes/no questions. The author suggests that Sera que is becoming reanalyzed as a question particle, thanks to the existence of question particles in Guarani. Furthermore, the presence of particles could influence the F0 contour as, in the case a particle is realised, a smaller final LH% rise is found in comparison to cases where no particle is not produced. As for the difference within bilinguals, in biased questions Guarani dominant bilinguals showed more variation compared with the two other groups. However, despite these differences, both bilingual groups and monolinguals showed the same intonational tendencies, which points towards the convergence of intonational systems due to the long period of contact and the use of Guarani particles. Besides pointing to the possible difference in the way contact affects different levels of grammar, here intonation and syntax, the paper argues in favour of language change due to intensive contact and causing intonational and syntactic “innovations” in Paraguayan, though it is reasonable to assume that any interference between Guarani and Spanish must have been bi-directional.

5.1 Factors affecting the impact of accommodation

Within this perspective, time is crucial in various respects. First of all, it is crucial because accommodation is a (gradual) process that exists only in time and, thus, time is preliminary to the observation of the process itself, its impact and other factors affecting it. Second, time differentiates the duration of accommodation (to start with the distinction between short-term and long-term), and, together with other factors, its effects and their persistence.

However, several factors may play a relevant role in promoting, or restraining, the accommodation process and its output (bottom level in Figure 1, where dots indicate that the lists are not intended to be exhaustive). As Figure 1 suggests, the main relevant factors, besides time, range from linguistic to extra-linguistic ⁹ and sociolinguistic ones; they further range from factors that seem to be more specific for situations involving individuals and groups of bilinguals (L2 learners, L2 speaking immigrants and heritage speakers; see the grey background cells in the bottom table) to factors that are expected to be at play also when entire communities undergo a language change (white background in the bottom table).

Linguistic factors that may be at play in all contact situations considered here (white background in the table) include systemic factors, such as phonetic and phonological aspects (e.g., from the type of segment to its phonological role in the system and the maintenance of phonological contrast; see, a.o., the contributions of Jochim & Kleber, Ulbrich, Celata & Nagi, Cutler et al., or Johnson & Babel in this Special Issue, where the variability of the systems involved are considered); they also include characteristics involving other levels of grammar and interface phenomena (see Celata & Nagy’s paper for the relevance of lexical frequency and Pešková’s paper for the interface with syntax). As for other levels in grammar, think also of works on the contact between intonation systems, showing that interrogatives favour accommodation because of the salience of their phonetic form (Simonet, 2011; Romera & Elordieta, 2013).

One of the linguistic factors that seem to be more specific to contexts involving individuals or groups of L2 learners, heritage speakers, and other bilinguals such as L2-speaking immigrants (grey background in the table) is the quality of input. The input’s quality varies if the speaker is exposed to the mother tongue or L2 speech (see discussion in, e.g., Cao’s paper in this Special Issue) and, in the case of heritage communities (Celata & Nagy’s paper), the input quality differs if heritage speakers are exposed to the majority language orthe heritage speakers’ speech. Another linguistic factor is the individual linguistic condition, meaning the number of languages that individuals know and their proficiency level (see, e.g., Ulbrich in this Special Issue).

In all contact situations considered here, accommodation may be also affected by extra- and sociolinguistic factors, such as the socio-cultural environment (white background cells in the table). Both a language change and the output of, for example, an L2 learning process are indeed possibly affected by partially intertwining factors, such as those related to characteristics of the family environment (to begin with the attitude of the whole family towards the use of different languages), the social class, the instruction level, the social network, and by dominance relations (in this Special Issue, see the paper by Pešková) and social distance between communities and speakers (as for the role of the specific social net the individual belongs to, see Allard & Landry, 1994).

Other factors at play may be gender and age, including the age of L2 learning. For instance, even in the case specific units are found to be particularly susceptible to accommodation and change, differences are often reported to depend on specific groups involved in the contact, such as groups sharing the same age or gender (Simonet, 2011; Romera & Elordieta, 2013), as well as groups sharing the length of residence in an L2-speaking country.

More specific to contexts involving individuals or groups of L2 learners and heritage speakers, and other bilinguals such as L2 speaking immigrants (grey background cells in the table) is the quantity of input, that is the amount of exposure to linguistic stimuli. ¹⁰ The quantity of input may differ in learning settings (immersion in L2 speaking community vs. formal setting within a society where L2 is not the primary language; as for the possible impact of formal instruction, see Butler, 2013); it may change in heritage speaker settings (where exposure to heritage-speaker L1 language may be highly variable, see Schmid & Köpke, 2019), as well as in bilingual L2 speaking immigrants (as the amount of L2 input depends on the amount of interaction with the local mother tongue-primary language speakers). On the other hand, in the case of language change, the quantity of input may probably be included under the duration of contact.

Finally, individual cognitive and socio-psychological traits affecting accommodation may be considered, such as the need for social approval, the attitude towards integration (e.g., the need for inclusion vs. the need to maintain or match group identity), the learning attitude, the attention capability, the empathy or, to make a final example, the autistic traits. ¹¹ Hypothetically, the lack of homogeneity in the results of investigations on various contact situations may be ascribed to differences in the perception of linguistic and social identity (see Trudgill’s salient features; in this Special Issue, see Celata & Nagi’s paper regarding the heritage speaker setting) and the related issue concerning the attitude towards integration (which, for instance in Simonet, 2011, Romera & Elordieta, 2013, seems to play a more important role than the intensity or the time extension of contact). Individual factors are numerous and potentially very strong, such as learning aptitude and personality traits as the level of empathy (Orrico & D’Imperio, 2020; Esteve-Gibert et al., 2020; see also Cao’s work in this Special Issue). However, studies in the literature increasingly consider the role of cognitive factors and their impact on language use, such as attentional abilities (Kong & Lee, 2018), working memory capacity (Yu et al., 2011), and even autistic-like traits (in a non-clinical population, Jun & Bishop, 2015; Yu, et al., 2011). Interestingly, having an impact on language use, these traits can easily play a role in favouring or disfavouring accommodation.

The overview given in the paper is ambitious, and various topics have been mentioned by referring to relevant studies. Of course, the references included do not aim at being exhaustive. Rather they are meant to offer examples, to start with contributions to the Special Issue, supporting the view of both temporary and long-term changes in various types of contact situations as due to the same process, that is accommodation. The different aspects related to the numerous issues discussed in this paper may be deeply understood thanks to references found in the works mentioned in the paper.