Landscape of Code-Switching in Tunisia

Julién Godman
10 min readJun 1, 2021


Photo credit: Julien Godman, taken in Tunis in 2018

Herein is a review and brief analysis of the state of code-switching and/or code-mixing in Tunisia. I examine some of the Tunisian contexts, settings, and uses of code-switching as it relates to identity, and effectiveness or purpose of communication. I conclude with thoughts and light speculation of code-switching and code choice in Tunisia in the near future.

Keywords: Tunisia, codeswitching, codemixing, French, Standard Arabic, identity, culture, language, dialects

Code-Switching Situation in Tunisia

Code-switching (CS) in Tunisia, and to a greater extent that of the Maghreb, has a rich and dynamic code-switching and diglossic landscape, often a combination of French and local Arabic dialects, and Modern Standard Arabic (MSA), excluding herein Amazigh languages (Bassiouney, 2020). The main purpose of this paper is to examine the contexts and uses of CS in Tunisia as they relate to identity and purpose. This is especially of interest today, now years after decades of French colonialism, as well as the more recent resolve of the 2011 Jasmine Revolution (Tunisian Revolution), first of the Arab Spring (Schraeder, 2012).

Historical Points: Colonialism and Revolution

Two historical points to account for, when trying to understand identity, culture and the language situation in Tunisia are: French Colonialism and the Jasmine Revolution. The period of French Colonialism lasted for 75 years, during which extreme efforts were made by the French to colonialize the very cultural spirits of the people of Tunisia, especially language (Aitsiselmi & Marley, 2008). During this time, French was promoted, or status planned, as the official language of the country, used for most official capacities, in schools, and for economic purposes, and was the second official language for some time after independence in 1956 (Aitsiselmi & Marley, 2008). This is important to note, both for the extent of French immersion into society and for the duration of French Colonialism and influence thereafter. Because of this, whole generations in North Africa (including Tunisia) have grown up within a cultural and linguistic duality, whereby French was used for important purposes, but diglossic variants of Arabic were used in the home and in everyday functions (Bassiouney, 2020).

The second historical point to note is the relatively recent Jasmine Revolution (Schraeder, 2012). Like any revolution, there was unrest, and varied perspectives and sociopolitical camps, to an extent differing ideological camps.

On a personal note, when I was in Tunisia in 2018, I noticed there was still noticeable levels of political instability or struggle within the country, particularly in Tunis. I witnessed demonstrations, speeches, small marches, a strong police presence, suicide bomber threats, and even on an interpersonal level the young urban friends I was with reiterated a perspective of an ongoing social struggle within Tunisia. This personal note is important on a linguistic and identity level because though the Jasmine Revolution was nearly a decade prior, the country seemingly is still working through complicated sociopolitical, perhaps even, cultural polarities.

Photo credit: Julien Godman, taken in Tunis in 2018

Code-Switching Function and Consciousness

Bouzemmi (2005) recognizes the importance of French in Tunisia, though also outlines a certain steady demise of its popularity and use. She highlights three primary functions in the Tunisian context that CS serves. First, referential function is when speakers CS between French and Tunisian Arabic (TA) because they ‘lack knowledge’ or ‘facility’ of the word in either one of the languages. This kind of CS highlights the bilingual state of Tunisia, whereby speakers may converse back-and-forth between French and TA because of perceived semantic appropriateness within a conversation. Referential function CS is on a conscious-level and the distinction or choice between which language often is related to specific topics within a conversation.

Bouzemmi’s (2005) second highlighted CS function is metalinguistic function, which is essentially when CS occurs because the speaker wants to impress, include, or exclude others (among other things). Speakers may try to impress listeners by CS from TA to MSA, French, or English to express concepts or ideas of an important nature, or under specific circumstances where CS to a ‘higher’ linguistic choice presents that speaker amicably. The third CS function, expressive function, is usually unconscious and is recognized by the utilization of both French and TA in conversation solely based on dualities of identity. This last function of CS is not about inclusion or exclusion, nor semantics or lexicon but more so on an individual-level, expressive of speakers themselves. Some, like Lawson & Sachdev (2000), view this demographically and geographically diverse distribution of CS in Tunisia, on an identity-level, as a mode of speech or full-on dialect of its own, known as Tunisian-Arabic French code-switching variety (TA-F).

However, it is important to note the difference between TA and that of TA-F. TA is an Arabic dialect, though many linguistic elements of TA, including some lexicon, phonetics, and more, have been influenced by French. TA’s basic structure and form is recognizably Arabic. However, TA-F is a CS mode of communication (perhaps dialect) of two entirely different grammar systems, French and TA together. In this, the choice to CS to/from TA, French, TA-F, English, MSA, or other, is greatly influenced by both context and identity.

Boussofara-Omar (2003) in a study and cross-examination of Myers-Scotton CS models, of 17 political speeches of former Tunisian President Habib Bourguiba, offers many examples of diglossic CS on syntactic and morphological levels. The aim of these speeches was to best connect with a wide and diverse Tunisian audience or to strengthen a statement. One transcribed example, below, combines forms of MSA with TA: ‘I do not think they were’ uttered as “NEG-1SG IMP think-NEG were.” For comparison the MSA variety of this statement is “NEG 1SG IMP-think that-they were” and the TA variety is “NEG-1SG IMP-think were.”

Regarding identity, Labben (2017) highlights the works of other linguists that cover some basic components of identity-making and identity classification. First, she touches on the work of Jan Stets and Peter Burke, whom outline that identity can be derived from social groups and social categories, i.e. political and religious affiliations, nationality, ethnicity, even sports interests. Second, she reviews the work of Marilynn Brewer, who breaks up identity into two categories — personal identity (individual level) and social identities (social or group level).

These divisions of identity (and the making of identity), either within a group or on an individual level, help understand what and when CS is appropriate. For example, on a personal note, in Tunisia in 2018, I noticed a difference in language choice of my urban friend group (20–35 years old), between TA-F amongst each other at cafes and clubs, and TA when they communicated with serving staff and cashiers (or other Arabs in polite functions) at markets and restaurants. And when speaking directly to me, either French or English was prescribed, with intermingled specific TA word, of which were particularly hard to translate. Their code choice within each context (formal, polite, familiar, friendly, etc.) speaks directly to identity. And though subtle, this level of CS signifies a distinct social and linguistic structure of members of a larger group (Tunisians, Arabs, etc.), dynamics between group members and non-members (Arabs and foreigners), and familiar/unfamiliar sub-groups, such as friends or strangers.

Code-Switching Implications of Gender and Education

Sayahi (2011) conducted a series of twelve interviews with varying levels of educated (either secondary school or higher education) males and females between 20–42 years old. He then compiled the recorded interviews into 3 hours of 15-min segments (cut from 60-min segments each), and then marked every instance and syntactic category of CS and linguistic borrowing of each speaker. First, he noticed a total of 1,721 instances of CS, with 36.73% CS on an inter-sentential level (CS between sentences), and the remaining 63.27% CS on an intra-sentential level (CS within a sentence, and often but not exclusive to nouns and noun phrases). He also noted that the majority of CS was from TA or MSA to French or TA-F.

Sayahi (2011) presents a prime example of the frequently seen intra-sentential CS of TA to French (or perhaps a form of code-switching TA-F), from one of the interviewees as: “lɛ: lɛ:. Jamais istaʕmiltu l-Français. D’ailleurs ʕumru: ma kɛ:n fi-l-luɣa imtɛ:ʕi: maʕnitha ma nitkɛlmu:ʃ systématiquement.” Translated as ‘No, no. I never used French. In fact, it was never part of my speech. I mean I don’t use it systematically.’

Sayahi’s (2011) last two conclusions were that females had a slightly higher likelihood to CS from TA to French (particularly Standard French) compared to males, which he corroborated with other linguist’s Tunisian studies, such as Dhaouadi’s (1996) hypothesis of the existence of ‘franco-arabe masculine et franco-arabe féminin’ or male and female variants of TA-F. And last he found that the higher the education level and depending on what field of study, the higher the frequency of CS occurred from TA to French, TA to TA-F, or TA to MSA.

Photo credit: Julien Godman, taken in Hammam-Lif in 2018

Code-Switching Identity and Context

Lawson & Sachdev (2000) conducted a series of studies (two reviewed herein), examining social and psychological aspects of CS, such as attitudes and individual perspectives. The first study, questionnaire-style, was comprised of 169 Tunisian college students using a ‘matched-guise’ technique, whereby participants listened to different dialects and languages spoken by the same speaker (without seeing or knowing the speaker) — including TA, MSA, French, English, and TA-F. The participant reactions and viewpoints on each vernacular were then recorded. The findings indicated that both men and women perceived TA-F as less favorable in general, but especially in status and prestige. However, though TA-F was negatively viewed a moderate percentage of the participants also more closely identified with TA-F, as well as attributed modest levels of ‘modernity’ to TA-F. The highest perceived levels of prestige, status, and social solidarity were recorded as TA (by men) and MSA (by women), with French and English both floating around the middle between perceived High varieties (MSA and TA) and the perceived Low variety (TA-F).

Lawson & Sachdev (2000) second study, a linguistic diary, was conducted on a similar small pool of 28 college students. Participants were instructed to write down any instance a language variety was used in/per conversation (in any setting or function), over a six-day period. The results showed that TA-F was the dominant language of choice in social settings and was the second variety of choice in study atmospheres (English being first) and utilized second in the home between differing familial age groups (TA being first).

Lawson & Sachdev’s (2000) findings suggest ‘that CS should be treated as a distinct linguistic variety in the Tunisian context’ and that a speaker’s setting can act as a heavy indicator of code choice. Last, the studies also indicated that the use of TA-F or multiple CS within communication was something observed within an intragroup (group based on close familiarity).

Interestingly, between the two studies TA-F was negatively perceived and yet was more widely used. And conversely, the perceived more prestigious and of higher status varieties of MSA and TA were utilized less than TA-F in most functions, except for the home where TA was the highest. Last, one may extrapolate a certain covert level of social or national solidarity at play (element of identity of a speaker), when TA-F is used or when diglossic CS between Arabic varieties is utilized on an intragroup level.

Conclusive Thoughts and Speculation

Looking at the varied uses of CS outlined above, there are intentional and clear-cut levels of CS — to include, exclude, or impress others. The use of these types of CS, from MSA or TA to French or TA-F, demonstrates a strong consciousness of speakers in Tunisia of both the bilingual landscape in Tunisia but also on some level consciousness of a bi/multi-culturalism. Choices within a group, or on an individual level, to CS between friends, at the grocery store, in school or the home, etc., also show an inherent structure of socially appropriate linguistic expectedness between speakers. However, the ambiguous nature of TA-F, which is some variant of CS and/or dialect of its own, shows both conscious and unconscious use, signifying that on some level two grammatical systems and their lexicons may have or may be merging into something more tangible and systematic than merely CS for functional or identity-marking purposes.

Given that English is overtaking French as the dominant foreign language, and the higher surveyed uses of TA-F or TA over that of French or MSA, I predict that TA-F will formalize or strengthen itself, within Tunisian society. Along with this, I see TA-F becoming less perceived as an inferior linguistic variant. Additionally, I think we may see more English lexical and syntactic elements infused into the TA-F form, perhaps soon formulating into an even more common, creole-like, language combining French, English, and TA all in one theoretical mouthful acronym as TAFE. In any case, given the political and uniting dominance of MSA in the Arab world (Bassiouney, 2020), and the current TA-F landscape, I do not see French returning to the level it once was.


Aitsiselmi, F. & Marley, D. (2008). The role and status of the French Language in North Africa. Studies in French Applied Linguistics. Volume 21(6). 185–222. doi:10.1075/lllt.21

Bassiouney, R. (2020). Arabic Sociolinguistics, Topics in Diglossia, Gender, Identity, and Politics. Washington, D.C: Georgetown University Press

Boussofara-Omar, N. (2003). Revisiting Arabic diglossic switching in light of the MLF model and its sub0models: the 4-M model and the Abstract Level model. Bilingualism: Language and Cognition. Volume 6(1). 33–46. doi:10.1017/S1366728903001032

Bouzemmi, A. (2005). Linguistic Situation in Tunisia: French and Arabic Code Switching. Interlinguistica. Volume 16(1). 217–223. Retrieved from

Labben, A. (2017). Revisiting face and identity: Insights from Tunisian culture. Journal of Pragmatics. Volume 108. 98–115. doi:10.1016/j.pragma.2016.11.011

Lawson, S. & Sachdev, I. (2000). Codeswitching in Tunisia: Attitudinal and behavioural dimensions. Journal of Pragmatics. Volume 32. 1343–1361. doi:10.1016/S0378–2166(99)00103–4

Sayahi, L. (2011). Code-switching and language change in Tunisia. International Journal of the Sociology of Language. Volume 211. 113–133. doi:10.1515/ijsl.2011.040

Schraeder, P. (2012). Tunisia’s Jasmine Revolution & the Arab Spring: Implications for International Intervention. Orbis. Volume 56(4). 662–675. doi:10.1016/j.orbis.2012.08.009