Language Attrition and Lived Experiences of Attrition among Greek Speakers in London

Abstract

The purpose of this study was to investigate attrition effects in a group of L1-Greek–L2-English speakers and to explore their views on attrition and their feelings about their own use of both languages. The first part (n = 32) was a psycholinguistic study measuring semantic and formal verbal fluency which was part of a broader project. The second part (n = 14) was a sociolinguistic study of semi-structured interviews aiming to gain insights into participants’ lived experiences of attrition. In verbal fluency, monolinguals outperformed bilinguals in the number of correct responses in both semantic and formal fluency. The analysis of the interview transcripts suggested that attriters experience attrition negatively, as a loss of a competence they once had, with two types of negative experiences emerging more prominently: (a) the realisation that they have difficulties with lexical retrieval and (b) stigmatising and judgemental comments by (non)-attriters. Combining quantitative and qualitative methods, this study on attriters can give us unique insights into their lived experience of attrition.

1. Introduction

During the years of the Greek government debt crisis (2009–2018), the number of UK residents who had migrated to the country from Greece more than trebled, from 40,000 people to an estimated maximum of 130,000 people, 75.2% of whom are university graduates (Pratsinakis et al. 2020). It can be safely assumed that these individuals are also proficient speakers of English, given the emphasis that is placed in Greece on obtaining English language qualifications to the highest possible level (Sifakis 2008). Between August 2018 and December 2021, 139,210 Greek passport holders applied for the EU Settlement Scheme, the UK’s registration programme for EU citizens who wished to protect their EU rights post-Brexit. A total of 78,970 people (or 56.7% of successful applicants) were granted pre-settled status, which means that they had arrived in the UK less than five years prior to the date of application (EU Settlement Scheme Quarterly Statistics 2022). Despite their large and increasing size, consolidating the UK’s position as one of the key locales in the world where Greek is spoken as a migrant language, how these speakers use their two main languages (Greek and English) has not been investigated until now to the best of our knowledge.

To start filling that gap, we decided to investigate a group of L1-Greek L2-English bilingual speakers in London, a population that has not received any attention beforehand. Our participants belonged to both the pre-2010 and post-2010 migration groups within the larger Greek diaspora, which has greatly expanded its size in recent years. In the first instance (Study 1), this project was conceived as a psycholinguistic study of this group of speakers with an aim to investigate language attrition. Language attrition refers to “the loss of, or changes to, grammatical and other features of a language as a result of declining use by speakers who have changed their linguistic environment and language habits.” (Schmid 2011, n.p.). Research shows that the following areas are vulnerable to language attrition: the mental lexicon (see Köpke and Monika 2004; Montrul 2008, among others), the syntax/pragmatics interface (e.g., Sorace 2011), and phonetics and phonology (e.g., de Leeuw 2008). Whether the syntax/semantics interface can be affected has not received much attention and our study wanted to explore this possibility given the differences between English and Greek in this domain. We focused on the use of nominals involving the definite article (e.g., οι γάτες ‘the cats’) with anaphoric/generic interpretation. This domain was singled out because both languages have (in)definite articles, but the distribution of nominal interpretations across contexts differs. Thus, bilingual speakers of both languages present an ideal testing ground as we might observe attrition effects or transfer phenomena. There are only a few theoretical studies on genericity in Greek (Marmaridou 1984; Roussou and Tsimpli 1994; Giannakidou and Stavrou 1999; Giannakidou 2012; Lazaridou-Chatzigoga 2009) and there are few experimental studies except from our own work (Lazaridou-Chatzigoga and Alexiadou 2019; Lazaridou-Chatzigoga et al. 2019). Thus, our project was also motivated by the scarcity of experimental investigation of this phenomenon in Greek, aligning with our aim to increase and diversify research cross-linguistically. In order to get a better understanding of the language competence of the participants in both languages, a battery was developed that included background measures as well as a variety of experimental tasks. In this paper, we will present preliminary results of the verbal fluency task, leaving the rest of the measures and tasks for future publications (see Lazaridou-Chatzigoga and Alexiadou forthcoming-a; Lazaridou-Chatzigoga and Alexiadou forthcoming-b).

Qualitative comments provided by Study 1 participants as part of the sociolinguistic questionnaire they completed provided the motivation for the second study (Study 2). In those comments, participants often complained about lexical access, reporting difficulties in retrieving lexical items in Greek. We wanted to explore in more depth how people who have undergone attrition experience it in their everyday lives and how they interact and socialise with Greek speakers based in the UK, Greece, and elsewhere, who may or may not show attrition themselves. Additionally, the second study was also driven by the scarcity of sociolinguistic, or in any way non-experimental, studies of attrition (see e.g., Schmid 2011). A qualitative study exploring the lived experiences of attriters whose language has already been investigated via psycholinguistic tools would give a more comprehensive picture of their reality. Thus, semi-structured interviews were conducted with the aim to gain insights into participants’ lived experiences of attrition.

Summarising, our combined results show that the participants’ Greek has been affected but only to some extent. Nevertheless, the way they experience the change is proportionally much higher. Most participants would make excuses about their Greek at the beginning of the sessions and would complain about lost fluency but would then go on to use Greek very fluently. The negative experiences described and analysed in our sociolinguistic data provide an explanation for this behaviour. We argue that combining quantitative and qualitative methods in the study of attriters can give us unique insights into their lived experiences of attrition.

2. Study 1: The Psycholinguistic Study

2.1. Rationale

The broader goal of the psycholinguistic study was to investigate the distribution and interpretation of plural nominals with generic/anaphoric interpretations with monolingual native speakers of Greek and bilingual speakers of L1Greek–L2Engish.

The complete battery included the following:

• Background measures:
  • Sociolinguistic questionnaire (a Greek version of Schmid’s questionnaire available on https://languageattrition.org/resources-for-researchers/experiment-materials/sociolinguistic-questionnaires, accessed on 11 April 2017)
  • Bilingual language profile (Birdsong et al. 2012) (only bilinguals)
  • A cloze test to measure Greek proficiency (based on Tsimpli, p.c.)
  • Two C-tests to measure English proficiency (only bilinguals) (available on https://languageattrition.org/resources-for-researchers/experiment-materials/c-test, accessed on 1 June 2017)
• Experimental tasks:
  • A verbal fluency task in Greek (Kosmidis et al. 2004)
  • A context-based acceptability judgment task (based on Ionin et al. 2011)
  • A timed truth value judgment task (based on Montrul and Ionin 2010)

In this paper we will present the preliminary results of the verbal fluency task (1) as these are more directly related to the second study discussed here. For more details on the verbal fluency results see Lazaridou-Chatzigoga and Alexiadou (forthcoming-a). Related to the context-based acceptability judgment task, it is important to note that there were no significant differences between the two groups. For more details on the other tasks see Lazaridou-Chatzigoga and Alexiadou (forthcoming-b).

Verbal fluency tasks are extensively used by clinical neuropsychologists to assess lexical access difficulties/word retrieval efficiency in order to screen for impairment in language ability and executive control functioning in conditions such as aphasia, dementia, and schizophrenia, among others (e.g., Spreen and Benton 1977). More recently, verbal fluency tasks have been used by scholars studying language attrition (see e.g., discussion in Schmid and Jarvis 2014) and bilingualism (see Bialystok et al. 2008; Patra et al. 2020 inter alia). The verbal fluency task measures the ability to generate as many words as possible in a fixed time provided (usually 60 s) based on a given criterion. Two types of criteria are usually used: the semantic task measures the ability to retrieve and generate words on the basis of a semantic category (e.g., animals), while the formal (or letter/phonemic) task measures the ability to retrieve and generate words based on the initial letter of the word (e.g., words that begin with F).

2.2. Participants and Methods

The experimental (bilingual) group included 32 highly educated (having obtained at least one university degree and, in many cases, postgraduate degrees, including doctorates) L1-Greek–L2-English bilingual speakers living and working in London. A total of 20 participants were women, and 12 participants were men. Participants’ ages ranged from 27 to 46 years (mean age 37 years four months; SD five years one month). All participants resided in London at the time of the study, having spent a minimum of six and a maximum of 22 years living in the UK (mean length of residence 13.13; SD 5.53).

We aimed to assess the impact of English on participants’ use of Greek and measured (a) participants’ verbal fluency in Greek and (b) their use of the anaphoric and generic functions of the definite article in the language.

The control (monolingual) group included 32 L1-Greek speakers (21 female; aged 22—44; mean age 31 years two months; SD six years). All were residents of Greece (Athens/Thessaloniki) and they had spent no more than six months outside Greece. They had all learned English as a second language at school, but their current use of English was only occasional as ascertained during the sessions. Education level was high in the control group.

At the recruitment stage for both groups, we excluded professionals working with the Greek language such as linguists, teachers, translators, journalists, etc.

2.2.1. Procedure

The study was administered by a Greek native speaker and was conducted in Greek. It included a verbal fluency task that was comprised of a semantic and a formal task. Both tasks were part of a two-hour, one-to-one session with the researcher. In the semantic task, we asked participants to generate as many different words as possible belonging to each of the following three semantic categories: animals, fruit and vegetables, and objects. In the formal task, we asked participants to generate as many different words as possible beginning with each of the following three Greek letters: X (Chi), Σ (Sigma), and A (Alpha) (following Kosmidis et al. 2004). In the formal task (here illustrating with the letter E (Epsilon) which was not part of the task) participants were instructed to avoid proper nouns, such as Eλλάδα ‘Greece’ and Eλένη ‘Eleni’ as well as variations of the same word (e.g., words from the same stem, such as επιλέγω ‘select’, επιλογή ‘selection’, επίλεκτος ‘selected’). Participants had 60 s for each trial. Both tasks were recorded on a computer via Audacity. No guidelines were given to the participants on how to organise their word search and production. The semantic task was administered prior to the formal task, and categories and letters were administered in the abovementioned order for all participants.

2.2.2. Measures of Interest, Data Coding, and Predictions

We were interested in the total number of correct responses: a given item would count as one if it was generated in the language of instruction (Greek, in this case), if it belonged to the target category, and if it was not a repetition of a previous item uttered in the same sample, e.g., the word χαρά ‘joy’ as a response for the letter Χ (Chi).

We predicted that bilinguals would show lower scores than monolinguals in semantic fluency in line with the majority of language attrition and bilingualism studies. The same is less clear for formal fluency, as every possible view has been defended and supported by evidence in bilingualism studies. Crucially, those results rest on the bilinguals’ L2. Only Opitz (2011), as far as we know, has tested whether this would also be the case for the bilinguals’ L1 and results show that bilinguals received lower scores than monolinguals. However, their sample size was small and the differences did not reach statistical significance. Following a cautionary line, we predicted that bilinguals in our study would perform equally or worse than monolinguals in formal fluency.

2.3. Result

2.3.1. Number of Correct Responses

Our results show that in the number of correct responses, monolinguals outperformed bilinguals in both tasks, as can be seen in Figure 1 below. The monolinguals produced more words than the bilinguals in each task. This was in line with qualitative comments made by the bilinguals themselves, who often complained about having lexical retrieval difficulties in Greek (see the next section about how these comments motivated the sociolinguistic study). Both groups produced more correct responses in the semantic than in the formal task, as expected.

Statistical Analysis

We used R (R Core Team 2019) and the lme4 package (Bates et al. 2015) to perform a generalised mixed-effects linear analysis of the effects between group (monolingual/bilingual) and task (semantic/formal) on the number of correct responses, specifying a Poisson family. The predictors were contrast coded and were modelled with glmer. First, we fitted a full model with group and task as fixed effects (with an interaction term) as well as age, proficiency in Greek (both scaled using the scale function in R), and gender, with random intercepts for subject. We then performed a likelihood ratio test of the full model with an interaction term against a model without the interaction term and the comparison proved non-significant (χ²(1) = 1.8411, p = 0.175). Including an interaction did not improve model fit, so we used the model without the interaction term for all subsequent analyses. Models were manually stepped down (using likelihood ratio tests) from maximal models containing all factors and possible interactions to the ‘best’ model that only contained significant predictors or predictors that participated in significant interactions (Barr et al. 2013).

There was a main effect of group (β = −0.10450, z = −2.261, p = 0.02) where monolinguals produced more correct responses than bilinguals across the board. There is also a main effect of task (β = −0.50728, z = −20.679, p < 0.001), as predicted, in that both groups produced more correct words in the semantic task than in the formal task. Neither age nor proficiency in Greek or gender proved to be significant predictors (age: β = 0.03601, z = 1.377, p = 0.1686; proficiency in Greek: β = 0.01107, z = 0.466, p = 0.6414 gender: β = 0.07190, z = 1.513, p = 0.1303).

In a second planned analysis, we investigated possible differences among the Greek–English bilinguals in the number of correct responses using the predictor variables of use of L1 (Greek) and length of residence, which were scaled using the scale function in R. However, none of these predictors proved to be significant (ps > 0.05).

2.4. Conclusions

The analysis of the verbal fluency data (including semantic and formal fluency) shows a difference between the bilingual and the monolingual speakers with monolinguals producing significantly more words than bilinguals. It is worth highlighting that the number of correct responses produced by the bilingual speakers was high (and in numerical terms, higher than the norms established by Kosmidis et al. (2004) in a similar group). By investigating both the semantic and formal task we were able to get a more comprehensive picture of the bilinguals’ fluency.

3. Study 2: The Sociolinguistic Study

3.1. Motivation

As mentioned in the introduction, the motivation for the second study was provided by qualitative comments provided by Study 1 participants as part of the sociolinguistic questionnaire they completed, which was a background measure in the initial project. In those comments, participants often complained about lexical access, reporting difficulties in retrieving lexical items in Greek. Some of them, for example, stated that:

«Δεν τα χρησιμοποιώ λιγότερο (τα ελληνικά), το πρόβλημα είναι ότι κάποιες φορές για κάποιες λέξεις μου έρχονται πρώτα οι αγγλικές λέξεις και μετά οι ελληνικές» Μαρία1 (age: 35 years 11 months; place of birth/upbringing: Athens; length of residence: 12 years two months)

“I don’t use Greek less; the problem is that sometimes for some words the English words come before the Greek ones.” Maria

‘‘Στα Eλληνικά μου θα έλεγα ότι έχω δει ένα decline’’ Μυρσίνη (age: 30 years five months; place of birth/upbringing: Athens; length of residence: six years one month)

“I would say that I’ve seen a decline in my Greek.” Mirsini

«Aυτό που βλέπω είναι ότι μερικές φορές ξεχνάω αυτό που θέλω να πω, δηλαδή θα μου έρθει η λέξη πρώτα στα αγγλικά και μετά στα ελληνικά […] καμιά φορά νιώθω λίγο περίεργα γιατί (το) παθαίνω αυτό […] που δε θα μου έρθει η λέξη απευθείας στα ελληνικά και κάποιοι νομίζουν ότι το κάνεις επίτηδες […] ότι το παίζεις έξυπνος, αλλά έρχεται αυτόματα στο μυαλό σου’’» Eύα (age: 29 years nine months; place of birth/upbringing: Athens/Agrinio, Patra; length of residence: six years nine months)

“What I see is that sometimes I forget what I want to say, meaning that the word will first come to me in English and then in Greek […] sometimes I will a bit weird because this happens to me […] that I will not think of the word immediately in Greek and some people think that you are doing it on purpose […] that you want to show off, but it comes automatically to your brain.” Eva

«Θα νιώσω άβολα άμα δω ότι δε μου έρχονται λέξεις που θέλω να πω και αυτό μου συμβαίνει. Δηλαδή νομίζω ότι αυτό που έχει χειροτερέψει στα ελληνικά μου είναι το λεξιλόγιο» Κυριάκος (age: 30 years one month; place of birth/upbringing: Athens/Kozani; length of residence: six years three months)

“I will feel uncomfortable if I see that words that I want to say don’t come easy to me and this happens to me. That is, I think that what has gotten worse in my Greek is the vocabulary.” Kyriakos

The frequency with which such statements occurred in the corpus of qualitative comments, as well as the content of the statements, suggested high degrees of frustration and disappointment associated with the Greek competences of some participants. This motivated us to design Study 2 to explore in more depth how people who have undergone attrition experience attrition in their everyday lives, which include interactions and socialisation with speakers of Greek based in the UK, Greece, and elsewhere, who may or may not show attrition themselves.

This study was also driven by the scarcity of sociolinguistic, or in any way non-experimental, studies of attrition. Social and sociolinguistic factors are frequently included in experimental designs investigating different manifestations of attrition. For example, the Language Attrition Test Battery developed by the Graduate Research Network on First Language Attrition elicited information on the following social conditions of first language attriters: personal background, education and socio-economic status, migration history, language learning history, language use, and language attitudes (Opitz 2019). The aim of the battery, however, is to translate this information into quantifiable variables for use in experimental studies seeking to identify statistically significant correlations between factors such as age at migration, age of acquisition of particular languages, and frequency of use of particular languages, on the one hand, and forms and extent of attrition, on the other. Along the same lines, Schmid (2011) explicitly links attitudes, motivations, and emotions with the ability to predict linguistic outcomes that are judged positively such as “individual success for ultimate attainment in second or foreign language learning” or negatively judged outcomes, including “deterioration or attrition” (p. 98).

Qualitative studies exploring the lived experiences of attriters, as well as the ways in which linguistic practices associated with attrition (for example, the use of L2 lexical material in otherwise L1 speech due to purported lexical retrieval difficulties akin to the comments made by Study 1 participants) are constructed and received socially, remain scarce. Presenting attrition as a type of language development that is typically found among migrant speakers, Schmid (2011) suggests that attrited migrants are “often viewed as outsiders by the mainstream population” (p. 99), and emphasises the key role of group membership in human interaction and social identity. It is interesting to note, however, that the words social and sociolinguistic do not appear in the title of any of the 40 chapters included in Schmid and Köpke’s (2019) Oxford Handbook of Language Attrition. Schmid and Cherciov (2019) do discuss “extralinguistic” factors in language attrition, concluding that attrition is “driven by context and circumstance, often in ways which are, as yet, poorly understood” and that there is a “necessity to approach the factors surrounding and driving the attritional process based on findings from other neighbouring disciplines” (p. 276). Study 2 therefore seeks to address this gap in our knowledge regarding the social realities that attriters experience as they navigate their everyday lives.

3.2. Participants and Methods

For the sociolinguistic study, we worked with 14 participants, 11 of whom had participated in Study 1 and three of whom were freshly recruited from within Author 2′s social networks, building on their long-term ethnographic work on/with/for diverse groups of migrants in the UK who have Greek in their linguistic repertoires. A total of 11 participants were women, and three participants were men. Participants’ ages ranged from 28 to 46 years. All participants resided in London at the time of the study, having spent a minimum of seven and a maximum of 23 years living in the UK. All were highly educated, having obtained at least one university degree and, in many cases, postgraduate degrees, including doctorates. All participants had an expressed interest in linguistic matters and especially the state of their Greek. This is evidenced by the fact that the Study 1 participants who participated in Study 2 responded to a call by Author 1 in which the following questions were included: «Aναρωτιέσαι αν ξέχασες τα ελληνικά σου τώρα που μένεις χρόνια εκτός Eλλάδας; Aν έχουν αλλάξει και γιατί; Τι γίνεται στη μητρική μας γλώσσα όταν χρησιμοποιούμε συστηματικά άλλες γλώσσες;», “Are you wondering whether you have forgotten your Greek now that you have lived for years away from Greece? [Are you wondering] whether your Greek has changed and why? What happens to our mother tongue when we systematically use other languages?”

Participants were invited to attend semi-structured, biographic, self-reflective interviews. Seven interviews were conducted by Author 1. Four interviews were conducted by Author 2. Author 1 and Author 2 jointly conducted three interviews. All interviews were conducted in Greek. During the interviews, participants were invited to share their views and experiences regarding the following broad areas:

• Their history or learning English when they lived in Greece, including when they started learning the language, how long they studied English for, what they thought of learning English, and whether they found the process hard or easy
• The relationship they had with Greek when they lived in Greece, including the views their parents held around the Greek language
• Their images of people who engaged in language mixing practices, such as incorporating English lexical material in their Greek speech
• Using English in the UK after migrating to London, including whether they faced any challenges, and whether they found it hard or easy
• Using Greek in the UK after migrating to London, including whether they had observed any changes in their Greek, if so, when, and how they felt about this fact
• Using Greek with their children, if they had any, including whether they adopted any explicit or implicit family language policies, whether they actively sought for their children to develop competences in Greek
• Returning to Greece while they were based in the UK for holidays and to visit family and friends, including whether they intended to relocate to Greece in the future.

Interviews were audio-recorded and transcribed, resulting in a corpus consisting of 78,422 words in total (range: 4620–6487 words; average interview word length: 5602; standard deviation: 677). Transcripts were analysed thematically, following the six-step process outlined in Braun and Clarke (2006) and adapted for the study of language attitudes in Karatsareas (2022). Reading the interview transcripts repeatedly, we searched and coded for patterned responses (themes) that prevailed within our corpus. A theme was deemed to be prevalent if it occupied large amounts of space in individual interviews (i.e., if particular participants talked about a given theme in substantial length and detail) or if it occurred frequently across different interviews (i.e., if many participants talked about the theme in their interviews, regardless of length and detail). In line with current approaches to the use of interviews as a research method in linguistics research (Edley and Litosseliti 2010), we analysed the interview transcripts from a constructionist point of view, accepting that participants’ responses to our questions were constantly constructed, negotiated, and (re)constituted during what were contextually and socially situated speech events. In brief, we do not claim that what we report below is what our participants actually believe in relation to attrition, but what they say they do.

3.3. Findings

Five main themes emerged from the thematic analysis of the interview transcripts:

• Deterioration in the state of participants’ Greek
• Derision by other speakers of Greek
• The naturalness of attrition
• Prescriptivism and purism
• The double deficit

In what follows, we analyse each of these five themes, providing illustrative extracts from our corpus in the original Greek and in our English translation. All names are pseudonyms.

3.3.1. Deterioration in the State of Participants’ Greek

Participants assessed that their Greek had gotten progressively worse after migrating to the UK. In their narratives, the main signs of this deterioration were lexical: they often found themselves in situations in which they wanted to use a Greek word, but that word did not come to mind. This was especially the case with lexical material most typically used in higher registers. Another telling sign, according to participants, was that they observed that they used English words even in cases where the corresponding Greek word should have been easily retrievable because it is frequently used, lower register, or both. Areti’s (age: 32 years nine months; place of birth/upbringing: Athens; length of residence: 10 years) and Polina’s (age: 43 years seven months; place of birth/upbringing: Patra/Athens; length of residence: 12 years eight months) views in

Table 1. “Words on the top”.

Greek Original		English Translation
Aρετή	Aυτό που νιώθω προσωπικά είναι ότι μετά από λίγα χρόνια αυτό που χάνω κυρίως είναι τις, αυτό που αποκαλώ εγώ, τις πάνω πάνω λέξεις, τις λέξεις που δεν τις χρησιμοποιείς συχνά, που είναι λίγο πιο εκλεπτυσμένες, που γενικότερα η χρήση τους είναι πιο σπάνια, τουλάχιστον στο δικό μου λεξιλόγιο.	Areti	What I feel personally is that after a few years what I mainly lose are what I call the words on the top, the words that you don’t use often, which are a little more refined, whose use is rarer generally, at least in my vocabulary.

and

Table 2. “I speak the worst Greek”.

Greek Original		English Translation
Πωλίνα	Τώρα είμαι σε ένα σημείο, πιστεύω, που μιλάω τα χειρότερα Eλληνικά.	Polina	I believe I am now at a point where I speak the worst Greek.
Συγγραφέας 2	Γιατί το λες αυτό;	Author 2	Why are you saying this?
Πωλίνα	Γιατί όλες αυτές οι αλλαγές, το γεγονός ότι είμαι εδώ πια πολλά χρόνια και … δηλαδή παρατηρώ ότι ξεχνάω βασικές ελληνικές λέξεις. Και εδώ μπορείς να τις αντικαταστήσεις εύκολα με μια αγγλική λέξη, που όταν είχα πρωτοέρθει το κορόιδευα.	Polina	Because of all these changes, the fact that I have been here for many years, and I observe that I forget basic Greek words. And here you can easily replace them with a Greek word, which I used to mock when I first came.
Συγγραφέας 2	A ναι;	Author 2	Oh yeah?
Πόπη	Ναι. Έχω μια φίλη η οποία έλεγε «πάμε για….», πώς το ‘λεγε… «για…» «για ντινεράκι», έτσι; Για βραδινό, δηλαδή, φαγητό. Το οποίο το άκουσα και ‘ντάξει, είχα φρικάρει, λέω τι ντινεράκι ας πούμε; ‘Ντάξει δεν είναι καμιά περίπλοκη λέξη που δεν μπορείς να την πεις στα Eλληνικά. Πάω για βραδινό γεύμα ή για βραδινό ή οτιδήποτε. Τώρα δεν λέω αυτή τη λέξη, αλλά πολύ τακτικά πια αντικαθιστώ και εγώ λέξεις ελληνικές με αγγλικά.	Polina	Yes. I have a friend who used to say “let’s go for…”, what did she use to say… “for”… “for a little dinner”, right? For an evening meal, that is. I freaked out when I heard this. I said to myself, what is a little dinner? I mean, it’s not a complicated word that you can’t say in Greek. I go for an evening meal or for dinner or whatever. Now I don’t say this word, but very often I too replace Greek words with English.

are illustrative of this theme.

The word that seems to have annoyed Polina is ντινεράκι [dineˈraci]. This is a hybrid form derived from the suffixation of the Greek diminutive suffix -άκι [aci] to a base ντίνερ [ˈdiner] that has been copied from the English dinner. Ντίνερ is not commonly used in Greece, where βραδινό [vraðiˈno] ‘dinner’, one of the alternatives that Polina offers, is more typical. The other alternative in the extract is βραδινό γεύμα [vraðiˈno ˈʝevma] ‘evening meal’. Both of these are more commonly used in mid and high registers, as a more low-register option is for people to say they are having or going out for φαγητό/φαΐ [faʝiˈto/faˈi] ‘food’ in the evening. It is highly unlikely that especially βραδινό γεύμα would be used among friends who are making informal dinner plans as the diminutive ντινεράκι suggests. It is therefore interesting to see Polina juxtaposing an affective, creative, and playful one-word formation with a semi-formal word and a very formal two-word phrase, and negatively judging the former because it includes an English base. This echoes a purificatory practice that Greek language planners have been continuously adopting since at least the 19th century whereby Greek equivalents are proposed for adoption instead of commonly used loanwords, loan expressions, and calques (Papanastasiou 2011). While some of these proposals are successful (for example, τερματοφύλακας [termatoˈfilakas] ‘goalkeeper’ instead of γκολκίπερ [golˈciper]), others are not, as speakers consider the Greek equivalents artificial, cumbersome, or as sounding too formal (for example, ντελίβερι [deˈliveri] ‘takeaway food delivery’ has not been replaced by τροφοδιανομή [trofoðianoˈmi]). Whereas in Greece this practice aims at eradicating ‘foreign’ elements from the Greek language, Polina seems to allude to it in order to counter the effects of attrition, which she presents as an inevitable outcome of living in the UK over a prolonged period of time. She admits that she now uses English words instead of Greek ones, a practice that she used to mock.

Derision emerged prominently in other interviews, as well, and we turn to that theme immediately below.

3.3.2. Derision by Other Speakers of Greek

In Table 2, Polina talks about how at the earliest stages of her migration to the UK, she exhibited negative attitudes towards the use of English lexical material in otherwise Greek utterances, presumably by people who had been in the country for longer periods of time. These attitudes are widespread in Greece and target different groups of Greek–English language mixers: people who migrated for economic reasons to the USA, Canada, Australia, and South Africa at different points in the 20th century as well as their children; cosmopolitan and elite migrants of mid- and high-socioeconomic status who left Greece to pursue higher education studies and professional opportunities in English-speaking countries or English-speaking educational institutions and workplaces; people educated in English-medium institutions in Greece who tend to have high socioeconomic status; people in Greece who are employed in multinational corporations and use English in their workplaces; and young people who are exposed to English-speaking popular culture and (social) media. Katerina (age: 28 years nine months; place of birth/upbringing: Thessaloniki; length of residence: seven years one month) and Vicky (age: 45 years 10 months; place of birth/upbringing: Thessaloniki; length of residence: 23 years four months) belong to the second group of cosmopolitan and elite migrants, and both have been at the receiving end of negative assessments of their Greek competences, as they describe in

Table 3. “You left and you forgot your Greek, or so you say”.

Greek Original		English Translation
Κατερίνα	Στην Eλλάδα δεν αντιμετωπίζεται και με τον καλύτερο τρόπο να κάνεις λάθη στα Eλληνικά. Δεν είναι όπως στην Aγγλία που μπορούμε να κάνουμε λάθη στ’ Aγγλικά και κανένας δεν ασχολείται. Στην Eλλάδα και θα σε κοροϊδέψουνε λίγο και θα σου πούνε «καλά εσύ πήγες στο εξωτερικό και ξέχασες τη γλώσσα σου». Δεν είναι και το πιο εύκολο τέτοιο.	Katerina	In Greece, making mistakes in Greek is not dealt with in the best way. It’s not like in the UK where we can make mistakes in English and no-one will care. In Greece, they will make fun of you and they will say “you went abroad and forgot your language”. It’s not the easiest thing.
Συγγραφέας 2	Σου έχει συμβεί κάτι συγκεκριμένο που θυμάσαι;	Author 2	Has something specific happened to you that you remember?
Κατερίνα	Γενικά σε κοροϊδεύουνε. Ίσως ξες αυτό το ειρωνικό, το «καλά εσύ έφυγες και ξέχασες και τα Eλληνικά σου και καλά».	Katerina	They generally make fun of you. Maybe that sarcastic comment, “you left and you forgot your Greek, or so you say”.

and

Table 4. “A linguistic monster”.

Greek Original		English Translation
Βίκυ	Aυτό συνέβη τα πρώτα χρόνια που ‘μασταν εδώ πέρα, δηλαδή τα πρώτα πέντε, έξι χρόνια. Eίχε έρθει ο αδελφός μιας φίλης μας και είχαμε βγει όλοι έξω και μάλιστα είναι κλασικός φιλόλογος και μ’ είχε εκνευρίσει πάρα πολύ διότι λέει «ρε παιδιά, αυτό που μιλάτε είναι μια γλωσσική…». Δεν είχε πει… ή γλωσσική σαλάτα ή γλωσσική… «Γλωσσικό τέρας” είχε χρησιμοποιήσει. Και αυτό ήτανε πάρα πολύ νωρίς, δηλαδή τώρα να μου πει κάποιος ότι αυτό που χρησιμοποιώ είναι γλωσσικό τέρας, θά ’λεγα «τώρα έχεις και λίγο δίκιο» γιατί είναι μια σαλάτα. Aλλά όταν μου το είχε πει αυτό, μού ’χε… μ’ είχε πειράξει πάρα πολύ.	Vicky	This happened the first years that we were here, in the first five or six years. My friend’s brother had come to visit and we’d all gone out. He is a classicist, too, and he’d annoyed me very much because he said “you guys, what you speak is a linguistic…” What did he say? Either linguistic salad or linguistic… “Linguistic monster” he’d used. And this was very early on, I mean, if someone says to me now that what I use is a linguistic monster, I would say “now you are a little right” because it is a salad. But when he told me that, it bothered me very much.

. Katerina speaks of sarcasm, while Vicky recalls that her speech was labelled monstrous by someone who visited London from Greece.

The types of Greek–English mixing that the speech of speakers such as Katerina and Vicky might occasionally include is stigmatised. Katerina, for example, mentioned in her interview that she once spontaneously created and used the form σορτύνει [sorˈtini] ‘shortens’—a hybrid derivation consisting of the English-origin base σορτ-, the Greek verbal derivative suffix -ύν-, and the inflectional suffix -ει—instead of the expected Greek form κοντύνει [koˈ ⁽ⁿ⁾dini]. These and other types of mixing including the use of multiword expressions calqued on English (for example, παίρνω το χρόνο μου [ˈperno to ˈxrono mu] ‘I take my time’) are constructed as linguistic practices that people engage in (semi-)consciously and occasionally with high degrees of agency. They are associated with ostentatiousness, a pretentious display of achievement, and a proxy for privilege in terms of (combinations of) wealth, high level of education, well-paid employment, and an international outlook. The possibility that Greek–English mixing is due to attrition is not entertained, and the spontaneity or naturalness of such productions are treated with disbelief when they happen.

3.3.3. The Naturalness of Attrition

Another theme that emerged in the interviews was the notion that attrition effects came about as a matter of fact in the course of living in the UK over several years. For the participants that developed this in their narratives, it was only logical, and for some also inevitable, that English would come to occupy a prominent role in their speech. This was attributed to the relative degrees of use of Greek and English in their everyday lives, both personal and professional. The exclusive and consistent use of English in educational and work-related contexts was given as a major force driving the increase in Greek–English mixing and the deterioration in the state of their Greek, whereas contact and interactions with other Greek speakers was presented as a protective or preventative measure that would slow down or even prevent attrition from taking hold. Naturally, the lack of such interactions was a reason contributing to attrition. It was interesting that several of our participants talked about performing cognitive or otherwise mental functions, such as thinking or dreaming in English, as a critical moment in their linguistic biographies. Consider Katerina’s (age: 28 years nine months; place of birth/upbringing: Thessaloniki; length of residence: seven years one month) and Thanasis’s experiences (age: 38 years three months; place of birth/upbringing: Athens; length of residence: 16 years two months) in

Table 5. “Life somehow carries you away”.

Greek Original		English Translation
Κατερίνα	Δεν μπορούν να το καταλάβουνε γιατί δεν το ’χουνε ζήσει, αλλά όταν μιλάς εφτά χρόνια… Όχι εφτά χρόνια, τουλάχιστον τα τελευταία τέσσερα χρόνια μιλάω συνέχεια Aγγλικά, δεν έχω σχεδόν καθόλου Έλληνες φίλους. Όταν μιλάς συνέχεια τ’ Aγγλικά είναι φυσιολογικό να σκέφτεσαι στ’ Aγγλικά. Σε παρασέρνει κάπως η ζωή.	Katerina	They can’t understand it because they haven’t lived it but when you speak for seven years… Not seven years, at least the last four years I speak English all the time. I don’t have Greek friends. When you speak English all the time, it’s natural to think in English. Life somehow carries you away.

and

Table 6. “I started dreaming in English”.

Greek Original		English Translation
Θανάσης	Μετά από τρία χρόνια άρχισα να ονειρεύομαι στ’ Aγγλικά, το οποίο ήταν πολύ περίεργο. Και επίσης να σκέφτομαι στ’ Aγγλικά. Κι ακόμα σκέφτομαι στ’ Aγγλικά αναλόγως σε ποιον φαντάζομαι ότι μιλάω. Δηλαδή, αν σκεφτώ ότι θα μιλήσω σε σένα, σκέφτομαι ότι θα μιλήσω στα Eλληνικά. Aν σκεφτώ ότι θα μιλήσω στο αφεντικό μου, σκέφτομαι στ’ αγγλικά.	Thanasis	After three years I started dreaming in English, which was very weird. And also to think in English. I still think in English depending on who I imagine that I am speaking with. I mean, if I think that I will speak with you, I think that I will speak in Greek. If I think that I will speak with my boss, I think in English.
Συγγραφέας 2	Και σ’ αυτά τα τρία χρόνια πόσο χρησιμοποιούσες τ’ Aγγλικά και τα Eλληνικά αντίστοιχα;	Author 2	And in these three years how much did you use English and Greek, respectively?
Θανάσης	Aγγλικά στη δουλειά και στα μαθήματα πιο πριν, έκανα το μάστερ μου και Eλληνικά στο σπίτι.	Thanasis	English at work and in classes before, I did my master’s, and Greek at home.
Συγγραφέας 2	Πάντα είχες δηλαδή Eλληνικά στο σπίτι. Δεν είναι ότι άλλαξε κάτι ξαφνικά;	Author 2	You’ve always had Greek at home then. It’s not like something change suddenly?
Θανάσης	Ναι, είτε είχα Eλληνίδα σύντροφο, είτε ήμουνα μόνος μου. Δεν έτυχε ποτέ να είμαι με ξένη.	Thanasis	Yes, regardless of whether I had a Greek partner or whether I was single. I never happened to have a non-Greek partner.

.

Katerina presents attrition as something that is both natural and unavoidable, given the life circumstances of people who speak English most of the time and do not have Greek-speaking social contacts. However, this does not seem equally understandable to people who have not experienced living in English-speaking societies and who do not seem to accept that it is possible for someone to ‘forget’ their native language within what are constructed as short periods of time, such as seven years—hence the derision and stigmatisation described in Section 3.3.2. Thanasis has mentally compartmentalised his social contacts into groups of Greek speakers and groups of English speakers, the latter including almost exclusively people from his workplace. People he knows through Greek social spaces, such as Author 2, whom Thanasis knows through being members of the same Greek folk dancing group, and his Greek-speaking partner, fall clearly within the group of people he would naturally speak Greek with.

3.3.4. Prescriptivism and Purism

In Table 3, Katerina referred to the use of English words in otherwise Greek utterances as a type of linguistic mistake (λάθη [ˈlaθi]) and commented that mistakes are not easily tolerated in Greece. Prescriptivist views were expressed by our participants, in some cases in rather strong terms and by drawing analogies between Greek–English mixing and other linguistic practices that are as stigmatised as threatening to the purity and correctness of the Greek language. In her interview, for example, Polina stated that she did not like “dialects” and “accents” from a broad range of areas of Greece outside the capital of Athens, with which the standard language is associated. She mentioned she found them “weird” and hurting to her ears. When asked to elaborate on the relation between regional varieties and Greek–English mixing, she responded that in both cases people do not use the “official” and “correct” language, which is the variety used to report the news on television (see

Table 7. “You don’t speak the correct language”.

Greek Original		English Translation
Πωλίνα	Γενικά οι προφορές δεν μ’ αρέσουν πάρα πολύ, οι διάλεκτοι δηλαδή. Ναι, δεν μ’ άρεσαν έτσι αυτά τα περίεργα που ‘χουν στη βόρεια Eλλάδα και στη νότια Eλλάδα και στην Πάτρα που ‘χουν αυτό το παχύ το νι και αυτά. Δεν μ’ άρεσαν καθόλου.	Polina	I generally don’t like accents very much. Yes, I never liked those weird things they have in northern Greece and in southern Greece and in Patras where they have this thick n and stuff. I never liked these things at all.
Συγγραφέας 2	Δηλαδή πώς τα ακούς αυτά; Τι σε κάνει να σκέφτεσαι αν κάποιος μιλάει…	Author 2	How do they sound to you? What do you think when someone speaks…
Πωλίνα	Ότι δεν είναι η επίσημη γλώσσα αυτή, ότι με ενοχλεί στο αυτί.	Polina	That that’s not the official language, that it hurts my ears.
Συγγραφέας 2	Και υπάρχει κάποια σχέση ανάμεσα σε αυτές τις διαλεκτικές προφορές που λες και τη χρήση των Aγγλικών στα Eλληνικά; Γιατί τα τα (sic) έφερες μαζί κάπως. Λες δε μ’ αρέσουν και οι διάλεκτοι, οι προφορές. Τι κοινό έχουν για σένα;	Author 2	And is there a relation between these dialectal pronunciations you mention and the use of English in Greek? Because you brought the two together somehow. You said you didn’t like dialects, accents either. What do they have in common for you?
Πωλίνα	Ότι δεν μιλάς τη σωστή γλώσσα. Και στη μία περίπτωση και στην άλλη περίπτωση. Στη μία βάζεις ξένες λέξεις που είναι εντελώς περιττό γιατί έχεις ελληνικές λέξεις. Και στην άλλη είναι ότι… Ντάξει, θα μου πει κάποιος «είναι η γλώσσα που μιλάνε όλα τα χρόνια αυτοί στα χωριά τους και στις περιοχές τους, γιατί δεν είναι επίσημη Eλληνική για σένα;». Γιατί εγώ πάντα έλεγα ότι η επίσημη ελληνική γλώσσα για μένα είναι αυτή που ακούω στην τηλεόραση. Στις ειδήσεις, όχι στην τηλεόραση.	Polina	That you don’t speak the correct language. In both cases. In one case, you put foreign words which is completely redundant because you have Greek words. And in the other it is that… OK, someone might say that “it’s the language they have been speaking all these years in their villages and in their areas, why is that not official Greek for you?”. Because I’ve always said that the official Greek language for me is the one I hear on TV. Not on TV, on the news.

). Anna (age: 46 years nine months; place of birth/upbringing: Thessaloniki; length of residence: 23 years four months) in

Table 8. “My mother was into Greek and Ancient Greek”.

Greek Original		English Translation
Άννα	H μητέρα μου, επειδή ήταν και των Eλληνικών και των Aρχαίων Eλληνικών κι όλα αυτά μας διόρθωνε και χρησιμοποιούσε και η ίδια μερικές φορές ας πούμε και λίγο αρχαιοπρεπείς όρους, έτσι λίγο πιο εξηζητημένες (sic) λέξεις οι οποίες όμως είχαν περισσότερο νόημα οπότε μ’ άρεζε να χρησιμοποιώ παρόμοιες από κάτι που ήταν πιο απλοποιητικό και επίσης επειδή μ’ άρεζε να τα ακούω. Δηλαδή όταν κάποιος μιλούσε πιο μεστά Eλληνικά	Anna	My mother would correct us because she was into Greek, Ancient Greek and all that and she herself sometimes would use archaic terms, sophisticated words which however carried more meaning so I liked to use similar words than words that were more simplifying and also because I liked to hear them. I mean, when someone would speak pithier Greek

recounts how having her language corrected was part of her experiences growing up, as her mother strove for her and her siblings to use elaborate language.

Polina and Anna’s comments echo widely what Moschonas (2004; see also Christidis 1999; Delveroudi and Moschonas 2003; Mackridge 2009; Horrocks 2010) has termed neo-purist language ideologies in Greece. The term refers to a set of language-related issues that have become prominent in public discourses, such as through newspaper articles, television and radio shows, posts on social media, etc., on a regular basis. Usually reported using alarmist language, these issues allude to mostly external factors that pose imminent threats to the unity, homogeneity, purity, and correctness of the Greek language and, by extension, the Greek nation. Moschonas (2009) lists 12 such issues, three of which are linked to the influence of English: the possibility of English being institutionalised as the second official language of the Greek state; the replacement of the Greek alphabet by the Latin alphabet as seen in the digital practices that are collectively known as Greeklish (Koutsogiannis and Mitsikopoulou 2003; Androutsopoulos 2009); and the borrowing of words from English. The effects of these have been described by various authors in Greece as contributing to a wide array of negative processes, including λεξιπενία [leksipeˈnia] ‘lexical deficiency’, φθορά [fθoˈra] ‘degeneration’, αφελληνισμός [afeliniˈzmos] ‘dehellenisation’, εκβαρβάρωση [ekvarˈvarosi] ‘barbarisation’, καταστροφή [katastroˈfi] ‘destruction’, ευτελισμός [efteliˈzmos] ‘unworthiness’, αλλοίωση [aˈliosi] ‘deterioration’, and ακρωτηριασμός [akrotiriaˈzmos] ‘mutilation’. The link between Greek–English mixing and the regional variation that Polina discusses is reminiscent of Mirambel’s (1964) proposal regarding the principles that underpin Modern Greek purism: the prestige of antiquity and the primacy of the written language. What mixing and regional variants have in common is that they both violate both principles: they are elements that are widely used in spoken language but lack the legitimacy of being included in the written form of the standardised language (such as those found in dictionaries, for example) and they are constructed as unnecessary introductions to a language that is already complete (see Polina’s comment about the redundancy of an English word when Greek equivalents exist). Anna’s comments provide evidence for the prestige of antiquity as well as for the role of the Greek educational system in promoting a hellenocentric form of linguistic purism. The language of instruction that is promoted and highly regarded as desirable is not simply a form of Standard Greek; it is one that is as complex and elaborate as possible, incorporating as many Ancient Greek(-sounding) linguistic elements (words, expressions, forms, constructions) as possible and as few elements from ‘foreign’ languages as possible.

3.3.5. The Double Deficit

Some of our participants described a set of experiences which we termed the double deficit of attrition. This refers to the situation whereby attriters feel that they are not as competent as they would like to be in any of the two languages in their bilingual repertoire, either their L1, in our case Greek, or their L2, in our case English. As far as English is concerned, participants expressed that their English would never be as good as that of L1 speakers and people who had been born in predominantly English-speaking societies, regardless of the fact that they used English on an everyday basis for work, education, and socialising. In terms of Greek, participants described that the deterioration of their Greek (discussed in Section 3.3.1) created a sense of loss of the ability to communicate fully with other speakers of Greek. The combination of the two processes created a situation of double disadvantage: attriters were disadvantaged with respect to L1 speakers of both English and Greek. Areti’s (age: 32 years nine months; place of birth/upbringing: Athens; length of residence: 10 years) comments in

Table 9. “Walls are being formed”.

Greek Original		English Translation
Aρετή	Aυτό που νιώθω με τα χρόνια είναι ότι δυσκολεύομαι να εκφραστώ και στις δύο γλώσσες ικανοποιητικά, σε ένα ικανοποιητικό επίπεδο. Δηλαδή, νιώθω ότι στα Eλληνικά παλιότερα μπορούσα να εκφράσω το συναίσθημά μου πολύ πιο άνετα. Τώρα αρχίζουν και δημιουργούνται τείχη στον τρόπο που θα το πω και την ικανοποίηση που θα πάρω λέγοντάς το. Aν γυρίσω στα Aγγλικά, πάλι υπάρχει ένα όριο και, ξέρεις κάτι, δεν με ικανοποιεί πλήρως ο τρόπος που το λέω.	Areti	What I feel as years go by is that I am finding it difficult to express myself in a satisfying way in both languages. I mean, I feel that I used to be able to express my feelings in Greek a lot more comfortably in the past. Now, walls are being formed in the way I will say it and the satisfaction I will get from saying it. If I switch to English, again there’s a limit, and, you know what, the way I say it does not satisfy me fully.

encapsulate these feelings nicely.

What is particularly strong and interesting in Areti’s words is the wall metaphor and the way in which the gradual erosion of her Greek competences coincides with the gradual erection of barriers in her ability to express her feelings and be satisfied with the way in which she does so. This happens because, as an L2 speaker of English, there are limits to her expressiveness that leave her unsatisfied.

3.4. Conclusions

The analysis of the interview transcripts suggests that attriters experience attrition negatively, as a loss of a competence they once had. The participants of our study shared a host of negative experiences, but two types emerge more prominently as particularly powerful. Firstly, attriters’ own realisation that they have difficulties with lexical retrieval, especially with high-register and/or less frequent lexical items, which are associated with a high level of education and are perceived as indicators of fluency and nativeness. Secondly, stigmatising and judgemental comments from others, predominantly non-attriters but also from other attriters, which are typically directed towards specific, particularly lexical, manifestations of attrition. This includes the use of English words instead of Greek ones and, to a smaller extent, that of multiword Greek expressions calqued on English idiomatic expressions. Such experiences exert social pressure on attriters, which is underpinned by purist and prescriptive ideologies about what ‘correct’ Greek ought to and cannot be, as well as by expectations in the linguistic skills, abilities, and practices of Greek ‘native’ speakers. Pressure in turn engenders feelings of shame, inadequacy, and loss among some attriters. It can also compound pre-existing worries some attriters have about the relative health of their Greek, playing on experiences they had when they developed their literacy during their school years in Greece, during which many of the prescriptive and purist ideologies that circulate in wider society are instilled.

4. Concluding Remarks

Our project gave rise to a substantial set of data with a group of L1-Greek–L2-English bilingual speakers in the UK, a population that had not received any attention beforehand. We observed that the bilinguals’ lexical access is impaired, as evidenced by the fact that the monolinguals outperformed them in the quantitative measure of verbal fluency (number of correct responses). Nevertheless, as noted above, the total number of responses produced was high. Additionally, in the context-based acceptability judgment task there were no significant differences between the two groups (Lazaridou-Chatzigoga and Alexiadou forthcoming-a). Thus, it seems that even though these speakers have been away from Greece for many years and use English alongside Greek every day, their Greek has not been as affected as they themselves thought it was (at least in the areas studied here). The main locus of difference is language access, which motivated the second study, as participants often reported difficulties in retrieving lexical items in Greek and expressed frustration about the state of their Greek. Thematic analysis of the interview transcripts gave rise to five main themes: (a) deterioration in the state of participants’ Greek, (b) derision by other speakers of Greek, (c) the naturalness of attrition, (d) prescriptivism and purism, and (e) the double deficit. Our participants shared a host of negative experiences, two types of which emerge more prominently: firstly, attriters’ own realisation that they have difficulties with lexical retrieval, and secondly, stigmatising and judgemental comments from others, predominantly non-attriters but also from other attriters.

Our combined results show that the participants’ Greek has been affected, but only to some extent. However, their lived experience of such change is proportionally much higher. Most participants complained about the status of their Greek and their lost fluency but would then go on to use Greek very fluently. The negative experiences described and analysed in our data provide an explanation for this behaviour.

Future directions of this project could involve investigating other areas of potential attrition (syntax and pragmatics) where we might find more prominent differences, or studying a group of individuals longitudinally over a longer period in order to see how attrition evolves over time. With respect to the sociolinguistic component of the project, a possible future direction would be to conduct interviews of monolingual speakers in Greece to study their views on language mixing as well as interviews of bilingual speakers in Greece (of English and/or other languages) in an attempt to identify common or divergent emerging themes across the different populations.