Baku presents a compelling case for studying linguistic landscape dynamics, given its history of language shifts and script changes – from traditional Arabic to Latin, then to Cyrillic, and back to Latin in the 20th century – and the recent elevation of Azerbaijani as Azerbaijan’s only official language.
These dynamics have left their imprint on Baku’s urban space, which is visually and linguistically rich and reflects its vibrant and competitive social and political realities. Analysing this linguistic landscape (LL) reveals a spectrum of influences: from Soviet ‘Russification’ policies and post-Soviet language policies to promote the status of the Azerbaijani language in the public sphere, to the increasing presence of European and Asian languages in commercial areas.
Our blog post examines multilingualism in modern Azerbaijan through a qualitative and quantitative analysis of Baku’s linguistic landscape. Situated at the intersection of sociolinguistics, anthropology and urban studies, our aim is to understand how language ideology and language usage are displayed in signs and reflect societal dynamics. In our project ‘Ethno-religious diversity between regulation and practice’, we aim to provide insights into lived and contested ethno-cultural diversity.
Methods
Our study was conducted in central Baku during fall and winter 2024-2025, documenting signage (design and use of signs) across ten selected streets to evaluate the interplay of languages and scripts. For that, we examined the presence, variety, distribution and frequency of written languages using LL mixed methods and addressing three main questions:
- Which languages and scripts are visible in the linguistic landscape and street communications in Baku, Azerbaijan?
- How does the language on street signs distinguish one area of the city from another?
- How are language ideology and recent regulations reflected in language usage on public signs and billboards? What is the strategic use of specific languages on street signs?
The research also raised a methodological question about how to collect and analyse data that captures Baku’s unique urban language landscape. What follows is our discussion on how to interpret the existing linguistic situation in the street signs of Baku and what to count as ‘street signs’.
What is a ‘sign’? Defining the unit of analysis
The considerable advantage of the LL method, especially in the context of this post-Soviet country, lies in its ‘unobtrusive nature’. First, the ‘sign’ as the primary object of the research is present in the public space regardless of whether the researcher intends to include it in the analysis. Second, collection of data does not involve the researcher’s direct interaction with the author/owner of the sign or require any informed consent. (The perspective of the owners and exploration of their motivations behind the use of certain languages are included in the follow-up research phase.) Third, the research on the street signs does not require any ‘clearance’ from the authorities or permission for data collection as ‘signs’ are part of the visible public space and accessible to everybody.
Nevertheless, one of the first questions encountered by all the LL researchers was this: What counts as a linguistic sign? Do we include only shop signs and public signs, which are printed on durable surfaces, professionally designed and produced? Or should any written text, advertisement, personal announcement or writing on a wall be included? How do we consider temporary signs like chalk drawings or graffiti? Should we focus solely on signs that clearly display certain functions (for example, advertising or announcements) or should we include signs of all kinds, for example, drawings of certain logos or symbols (such as swastikas)?
Our research team agreed to adopt a moderate ‘inclusive’ approach and to collect all visible formal and informal signs which contain textual information and can thus be used in analysing the structure of the sociolinguistic landscape.
The next question is related to research techniques and to the question ‘How do we ensure consistency and comparability when addressing the signs?’ When working in a team, different researchers may have different interpretations of what constitutes a ‘sign’ or how to categorise it. For this reason, joint documentation guidelines were developed, and coding was introduced to help the researchers make the distinctions. Moreover, the researchers were instructed to take photographs of all signs that might be interpreted as ‘ambiguous’ or stand out in terms of their creative approach or original/imaginative use of languages.
How to handle multilingualism
The next set of questions concerns the documentation of the ‘signs’ and the decisions on which attributes they need in order to be classified. Since the key focus of the current study is ‘language usage’, the questions to be addressed were as follows: Which language(s) is/are displayed? How should signs with mixed languages or specific code-switching be categorised? What should be done with ambiguous signs, where the language used is not apparent or the sign intentionally or unintentionally introduces misspelling from the perspective of formal grammar rules? The word ‘salafan’ is an example: used in the Azerbaijani language for a plastic bag or plastic cover, adopted from the Russian целофан (English cellophane), which is ‘misspelled’ in the Latin script, but understandable for every Azerbaijani citizen.
In order to facilitate the processes of interpreting the signs and language attribution, we developed a coding scheme that includes variables focusing on language used. It captures the script, the number of languages on the sign, the language(s) on the sign, the placing of the sign, and, in the case of bilingual or multilingual signs, the combination of languages used.
Although the coding procedure appeared quite straightforward, in reality the team faced numerous challenges during the interpretation and language attribution processes. For instance, which language should be attributed to such signs as ‘Ponçik-monçik’, an echo-word formation deriving from the Russian for ‘donut’ with the popular addition ‘monçik’ in the Azerbaijani language? Since there are no direct alternatives in Azerbaijani culinary language for ‘donut’ (пончик in Russian, meaning a round piece of sweet dough fried in oil), we interpret this creative ‘borrowing’ in the Azerbaijani language as an example of vernacular mixed-language signage. Such echo-word formations are typical of Turkic languages and serve a pragmatic function, namely to create a certain appeal, aesthetic or humorous effect. A similar pattern can be observed in the case of ‘Coffee Moffie’, the name of a popular coffee shop in central Baku, or the pizza chain ‘Pizza-Mizza’. Both cases led to a discussion of whether we should treat the sign as ‘monolingual’ (Russian or English) or ‘bilingual’ (English plus Azerbaijani) or something else, such as a hybrid ‘third space’ construct generated through regional language contact.
A further example of vernacular language mixing in everyday life is the street food location ‘DANA BURGER-çi’. The word ‘dana’ in Azerbaijani means beef, ‘burger’ is English, and the ending ‘-çi’ is Azerbaijani and denotes profession/occupation, so the sign can be translated as ‘Beef Burger Maker’, providing an example of code-switching, defined as the practice of alternating between two or more languages in areas with linguistic diversity.
Relying solely on quantitative approaches was not always feasible and often provided a limited perspective. For this study, it was crucial not only to mechanically ‘assign’ the codes but also to place the sign within a broader sociolinguistic context, observing contested interplays and offering qualitative interpretation. In these ‘complex’ cases, the sign was included in the dataset as ‘ambiguous’, accompanied by a qualitative explanation and photographic evidence.
Conclusion
Studying the linguistic landscape is a dynamic exercise. At the same time, it poses significant methodological challenges for the researchers, which need to be addressed in more detail based on the specific context. This essay discussed briefly the challenges faced by the researchers during the fieldwork in Azerbaijan and at the later stage of data analysis, when new findings and solutions were discussed by the team. The above reflections are therefore not conclusive but are intended to invite further exploration from both a methodological and a socio-cultural perspective.
Acknowledgements
The authors are grateful to our field assistants Gabil Hajiev, Alina Vasiutsina and Aysu Usubova for their reliable support in the documentation of linguistic landscapes and for their participation in discussions ‘on the ground’.
