1 Literature Review and Research Gaps

There is still limited scholarly engagement with language contact in written texts. One significant exception is Bilingualism in Ancient Society: Language Contact and the Written Text by Adams, Janse and Swain (2002). This book is an important introductory volume to the textual evidence of bilingualism in the ancient Mediterranean world. The authors provide insights into the challenges of bilingualism and show the interactions of multiple languages in written texts from antiquity and how they coexisted in the sources. In careful and well-crafted analyses, the authors expose not only the linguistic evidence of course of such interactions but also importantly, the broader sociocultural aspects of the phenomena. The book is useful for exploring the nature of how ancient societies dealt with multilingualism in written text and our historical linguistic environments. The book and other studies (Knauth 2007; Locher 2017) provide interesting nuances to the topic, but there are also a number of larger, even unexplored, avenues. Following are a few of those gaps:

2 Borrowing Typology

Analysing the language of the rare written forms of regional vernacular Iranian languages that traditionally lack a written form, such as Fahlavīyāt that traditionally have no written form reveals a pronounced borrowing of Persian lexis. Fahlavīyāt are a kind of poetry in regional languages of western and central Iran, in regions historically identified as Fahlah or Pahlav (Tafażżolī 2012). Similar to Fahlavīyāt, we see a strong influence from Persian in our letter. Borrowing was a necessary consequence of Persian’s dominance, due both to its multifunctionality and its prestige as the primary source of loanwords. Most spoken Iranian languages have little to no specificity to describe modernity, technologies, and cultural practices; whereas, related to modernity, Persian has abundant lexicon, which fills the lacunae of the other Iranian languages. As is evident, the letter we are analysing embodies two types of lexical borrowing from Persian: borrowings through direct borrowings or adaptations.

3 Borrowability Analysis

When examining examples of borrowing, one major question arises: which features are more frequently adopted from Persian into Dari, and which elements remain stable despite prolonged contact with Persian? To address this, it is essential to thoroughly review the existing literature on the subject.

Thomason and Kaufman (1988:14) assert that any language feature might be transferred from one language to another. However, the impact of contact-induced transfer varies among different linguistic features. For this reason, some borrowability scales, also known as hierarchies, have been developed to understand the varying impact of contact-induced transfer across different linguistic categories. These scales consider factors such as the intensity of contact, the degree of bilingualism (Thomason/Kaufman 1988; Thomason 2001), the structural compatibility of the languages (Myers-Scotton 2002; Field 2002), and the semantic-pragmatic features of the borrowed categories (Matras 2007; idem 2010).

Borrowability scales suggest that certain linguistic features are more likely to be borrowed between languages than others. For instance, unbound forms are more borrowable than bound ones, lexical items are more borrowable than grammatical items, and semantically transparent forms are more borrowable than semantically opaque ones.

In Whitney’s (1881) account of grammatical borrowing, function items such as prepositions are more frequently borrowed whenever possible than conjunctions, which are seen more frequently borrowed than pronouns. Derivational affixes are also more readily borrowed than inflectional affixes.

In his study of Norwegian and Swedish immigrant speech in the United States, Haugen (1950) observed the patterns of lexical borrowing among different parts of speech, establishing a hierarchy: nouns > verbs > adjectives > adverbs, prepositions, interjections.

Muysken (1981) also studied patterns of lexical borrowing and established a hierarchy: nouns > adjectives > verbs > prepositions > coordinating conjunctions > quantifiers > determiners > free pronouns > clitic pronouns > subordinating conjunctions.

Matras’ (2007) hierarchy is more detailed, including a broader range of grammatical categories: nouns, conjunctions > verbs > discourse markers > adjectives > interjections > adverbs > other particles, adpositions > numerals > pronouns > derivational affixes > inflectional affixes.

Melissaropoulou and Ralli (2020) focus on the borrowability of free grammatical elements, often argued to transition between the lexicon and the grammar or between content words and affixes.

This perspective on the transitional nature of free grammatical elements is particularly relevant to the present study, which seeks to assess the extent to which such elements, alongside lexical items, are susceptible to borrowing in a contact setting. Before proceeding, it is important to note a limitation of this study: the source text, specifically the letter, is short. The brevity of the text affects the statistical representativeness of the data and may skew overall interpretations of language contact, and structural coupling. The availability of a larger and/or more varied corpus may have permitted a greater empirical base for identifying borrowability hierarchies across language categories. Moreover, the opening section of the letter (lines 1–16), completely in Persian, contains only one instance of a Dari word xošta (line 6), limiting the available data for making conclusions about Persian-Dari interaction in that segment.

Nevertheless, the proposed methodology can still be applied to larger corpuses and provide a model for future research. This methodology, if applied to larger data sets, would allow for more rigorous statistical testing and would permit textual comparison for borrowings, phonological and grammatical convergence.

To investigate the distinctive characteristics (borrowability) of various language characteristics borrowed from Persian into Zoroastrian Dari (ZD), a quantitative study based specifically on the distinctive forms occurring across this letter is employed. This analysis avoids any forms that both languages share, as they can offer no insight into the sociolinguistic behavior of borrowing. This is particularly problematic when comparing two genetically related languages which share traits and vocabulary. For example, the preposition az (meaning “from”) and the conjunction ke are common in the whole letter. Nonetheless, as both of these forms are identical in both Persian and ZD, they were omitted from the calculation of borrowability.

Fundamentally, this study is concerned with the frequency of genuine Dari forms of each linguistic type according to its linguistic category. When engaging with the full occurrence as the denominator and with the native Dari forms as the numerator of each category, I created a rate of percentage to quantify the degree of borrowability for each category. That is the simplest illustration of the strength of the probability of borrowing Persian forms into ZD particularly in these linguistic categories.

For example, in working with prepositions, there are eight prepositions (none of which are shared between Dari and Persian); in total, there were only three were genuine Dari forms and all other forms (five) were from Persian. Thus, the borrowing rate or borrowability for prepositions was 62.5 % showing that ZD draws on Persian prepositions at a greater rate than utilising its own forms. This methodical process allows me to assess the influence of Persian on ZD in this letter in a systematic manner and to quantify the extent of borrowing linguistically throughout the various grammatical categories. For this reason, I excluded certain categories with low occurrence. For example, the category of reflexive pronouns was excluded because there was only a single instance of its use in the letter, in which the genuine Dari word xa was employed. This lone occurrence is insufficient to support a reliable analysis.

In the case of independent personal pronouns, I focused only on the first- and second-person singular forms (again as set apart from the object marker form, rā). The pronoun ūn, which relates to the third-person singular pronoun, is common to both Persian and ZD. Therefore, ūn is neutral to this analysis.

Based on the frequency of native Dari forms used in each category and the calculation of the proportion of these forms to derive a borrowability rate, the following key points can be deduced from the analysis.

First, the analysis indicates that the borrowability rate of the ezafe particle is 94.12 %. This means that all uses except one, are Persian forms. The categories of substantives (82.61 %) and demonstratives (83.33 %) also have high borrowability rates indicating a heavy dependence on Persian forms.

The category of prepositions shows a borrowability rate of 62.5 %, which, while lower than some other categories, still indicates a considerable adoption of Persian forms over native Dari equivalents. This indicates that prepositions are also susceptible to borrowability, but perhaps not as much as other grammatical categories.

The borrowability rates of the verbs and independent personal pronouns are 0 %. Therefore, all forms of verbs and independent personal pronouns in the letter are genuine Dari forms with no borrowing from Persian. This could indicate a strong retention of native forms in these essential grammatical functions, possibly reflecting a core area of linguistic resistance to borrowing.

In putting together, a borrowability scale based on this information, the ranks of the parts of speech from highest to lowest borrowability are as follows:

Our letter illustrates that essential grammatical structures, such as verb morphology and basic morphosyntactic elements including verbs and Independent Personal Pronouns, always remain stable and resistant. This stability primarily stems from these elements forming the backbone of a language’s grammar and being deeply embedded in the linguistic competence of native speakers. Their high frequency and central role in sentence construction make them less prone to borrowing and alteration.

These results diverge from known borrowability scales in some important ways. In our study, verbs and independent person pronouns are resistant to borrowing with a 0 % rate or level of borrowability from Persian. This differs to the borrowability scales developed by Haugen (1950), Muysken (1981), and Matras (2007), which indicate verbs and pronouns as some of the most borrowable of grammatical constructs but tend to be less borrowable than nouns. The stability of these categories in ZD, as reflected in our data, suggests that they form a core part of the language’s grammar and are resistant to borrowing, likely due to their essential role in sentence structure and communication.

Our results show that the Ezafe particle has a 94.12 % borrowability rate, indicating it is highly borrowed from Persian. This high rate differs from earlier borrowability scales, which typically place more emphasis on the borrowing of content words like nouns and less on functional elements like particles. This finding underscores the unique dynamics of Persian-ZD contact, where even functional grammatical elements are highly susceptible to borrowing.

In our data, prepositions have a borrowability rate of 62.5 %, indicating significant, though not overwhelming, borrowing from Persian. This finding is somewhat consistent with previous research, where prepositions are often cited as more borrowable than other grammatical elements like pronouns or conjunctions. However, the consistently observed borrowability percentage of ZD suggests a more moderate tendency of borrowing relevant to Persian and ZD language specific contact and actual situation.

Our results reinforce the idea that borrowability is highly dependent on context. These findings align more closely with the perspective of Melissaropoulou and Ralli (2020), who argue against the existence of a universal borrowability scale with predictive power across all language contact situations. Our analysis suggests that the borrowability of linguistic features in ZD cannot be fully predicted by existing scales, but rather depends on the particular communicative context of Persian-ZD interaction.

4 Code-Switching

Studies of Code-Switching often rely on analytical frameworks built for spoken language, and these taxing frameworks do not typically transfer easily to the analysis of written texts due to the intrinsic differences between spoken and written language. The written texts are typically more permanent and lack the immediate, dynamic interaction found in spoken conversations. Written language does not occur in a series of exchanges between parties like spoken language, and once it is produced written language will likely remain the same, rather than changing or shaping new patterns of language. However, to some extent, written forms of language can more closely approximate interaction and spoken discourse when produced as letters. Although written texts vary in type, typical categories connecting to spoken texts may often not hold. A key, essential way to approach the study of historic texts is to see them as social practices, which need to be studied in their social, historical and linguistic contexts. Also, when it comes to the analysis of mixed language written texts, I would argue it is appropriate to think about who is writing this type of text, for whom they are writing, how they are used, and ultimately, why they are written.

There is a rare but important example of a book focused on the phenomenon of code-switching in written language: Language Mixing and Code Switching in Writing: Approaches to Mixed-Language Written Discourse, edited by Sebba et al. (2012). The book provides research on this alternation of languages in written forms, ranging from historical to contemporary periods and covering a wide array of media, from traditional as well as modern forms such as emails, SMS and blogs. The book offers new perspectives on written discourse and it emphasises the study of mixed-language literacy as a social practice and the textual and visual dimensions of written communication.

Two further works that are particularly relevant to this chapter are those of Schendl (2002) and Nurmi and Pahta (2004). Nurmi and Pahta’s study examined code-switching between English and other languages in letters written in fifteenth- and early sixteenth-century England. Their research explores code-switching in the context of a variety of social variables but with a primary focus on social stratification. Their analysis showed a clear association between writers’ social rank and their code-switching behaviour, both in terms of the frequency with which they used code-switching and the types of switches they employed. They also mention that other factors, such as gender and education, play a significant role in this linguistic phenomenon.

Schendl (2002) argues that the choices of language made by Richard Kingeston in his two mixed-language letters had a heavy correlation to the social positions and relationships of the writer (Kingeston) and the audience (the King). The study finds that Kingeston was likely trying to create a more personal and urgent message by using English in the postscript of his first letter. This interpretation contrasts with earlier views by Hingeston and by Suggett that the switches to English were a result of Kingeston’s poor command of French or offered as an impulsive reaction to some form of stress. Schendl thinks the switches are intentional, and that it serves a purpose, especially in the case of wanting the message to be more personal and direct.

From these two works, which focus on the social motivation of code-switching, we will also take the opportunity to examine our letter with an expectation of explaining the writer’s decisions to use distinct languages for different purposes.

We must therefore consider not only how the writer of our case study switches between Persian and Zoroastrian Dari, but to contemplate the pragmatic functions behind the writer’s language choices, especially when examining the motivation for switching to the Zoroastrian Dari in the middle section of the letter.

The writer uses Persian most of the time but also includes a considerable amount of Zoroastrian Dari. Persian is used in the beginning (lines 1–16), briefly in the middle (lines 31–35), and again at the end (line 35). Zoroastrian Dari appears primarily in the central section (lines 16–31), with brief additional usage at line 35, overlapping the transition between middle and end. In Table 3, the details of the code-switching are presented.

We can easily recognise the five parts² which Haskins (1955) offered on composing letters. This indicates that whatever the time and location of the letter, they seem to have some commonality of structure and content. The letter begins with the “Salutation” section, comprising greetings and expressions of respect toward the recipient: “May I be sacrificed for you. After expressing my devotion and sincerity, which day by day (2) has increased due to the sentiments of familiarity-worship and kindness to friends of that existence”.

It proceeds with attempts at obtaining the goodwill of the recipient, while establishing a positive tone (Exordium portion) of common ground, i.e., previous friendship, unity, and moral distinction. The writer creates a positive, optimistic tone despite acknowledging challenges, and prepares the reader to be receptive to the writer’s values and emotional appeals: “Truly (3) at the time of testing friendship and unity, it has become clear and good has been distinguished from bad. Truly (4) I have been and remain your friend, and I hope that at the time of testing and demonstrating service and servitude, (5) I hope that favor will be shown. It has been a while since I have received any information about your well-being. I hope (6) you are well and happy”.

The letter continues, line 6, with the facts or background information (exposition), which constitutes one of the reasons for writing the letter. It seems as though the author is now answering the one question the recipient likely asked previously and provided an explanation in answer to that question.

Nevertheless, it appears the main purpose of the letter lies in the confidential Dari part, where the writer states the purpose of the letter (Petitio), and submits a request to the recipient (lines 16–31). The detailed instructions contained in these lines suggest he is advising the recipient on how he should respond if questioned on their finances, explaining they dealt with a higher-up of a close friend, and discussing whatever attempts he made to inquire about the well-being and finances of the recipient. The writer is suggesting the discomfort of prolonged waiting—both being kept waiting and having to make others wait for hours. The writer cautions against the potential investigations or information fishing endeavors, emphasising the need to keep business affairs hidden.

In closing, the writer also expresses thanks for support, as well as mentions expenses incurred during the process. Thus, the writer summarised the main points, restated his request, and offered a polite closing.

The letter involves inter-sentential code-switching. In this particular case, the language switch occurs at the paragraph level. It means that entire paragraphs are written in one language before switching to another language in the following paragraph. In addition to inter-sentential code-switching we can see a further variety of intra-sentential code-switching, meaning that the code-switch happens within the same sentence as we saw in line 35.

Each discrete part of the letter is written in a distinct language, either Persian or Zoroastrian Dari. The intention of the writer in separating the Persian language from the Zoroastrian Dari language shows that the writer has intended to divide various segments of an audience. The writer writes in Persian when the letter enters the non-confidential parts of the letter, e.g. greeting and general information that can be shared with a broader audience. In contrast, Zoroastrian Dari is used for confidential content intended specifically for the main recipient.

In strategic language switching in letter writing such as this, beyond managing the privacy and dissemination of information, it also reflects the writer’s intention that only the intended recipient fully understand the sensitive content. The code-switching in this case is driven by a combination of linguistic and social factors. It reflects the writer’s intent to balance openness with confidentiality, engaging different readers appropriately based on the content being conveyed. Furthermore, the switch to Zoroastrian Dari in the middle section of the letter serves specific pragmatic functions. The switch may have indicated solidarity to the recipient associated with perceived shared background and trust in the underlying messaging. It also allows the writer to express nuances such as seriousness or confidentiality that might not be as effectively communicated in Persian. A sophisticated use of switching also shows the attention the writer paid to the audiences and content, and shades of meaning, thereby managing the message that they were trying to convey in a letter.

Schendl’s (2002) argument that code-switching can be an intentional and strategic choice rather than simply a result of language competence issues provides a useful framework for interpreting the sophisticated use of inter-sentential code-switching in this letter.

5 Conclusion

This binational letter adds to the body of linguistics literature on Iranian languages and serves as a case assignment to examine issues of language contact and multilingualism in written texts. It provides evidence of a low source language and illustrates complex bilingualism between two genetically related languages.

The letter highlights two types of borrowings from Persian: direct borrowings and adapted borrowings. Direct borrowings include terms for administration, everyday life, and formal communication, highlighting the extensiveness of influence from Persian. Adapted borrowings generally involve complex verbs that have substantives in Persian, combined with morphologically adjusted Dari verbs. Additionally, the use of the Persian object marker rā and the presence of both genuine and adapted vocabulary highlight the trend towards Persianisation. Genuine Dari alternatives either omit the object marker or use e instead. Moreover, the letter reveals instances of confusion and inconsistency in using rā, as well as the Persian subjunctive marker be‑ versus the traditional Zoroastrian Dari ve‑. A phonological adaptation is also noted where maʿlum (clear) becomes maʿlim (obvious), displaying typical Zoroastrian Dari vowel shifts.

In the analysis, I examined the stability of grammatical items in Zoroastrian Dari (ZD) and formulated a borrowability scale based on the ratio of genuine Dari forms to borrowed Persian forms. I found that core grammatical items, such as verbs and independent personal pronouns, show high stability and no borrowing (0 % borrowability rate). Items such as the Ezafe particle, demonstratives, and substantives have fairly high borrowability.

Analysis of code-switching in the letter revealed a sophisticated and strategic use of both languages to manage a number of communication requirements and privacy issues. The writer used Persian for the general content of the letter, while Zoroastrian Dari was used in places where highly confidential information is conveyed. The inter-sentential switching allows the writer to communicate across different discourse domains in ways appropriate to the general and intended segments of the audience.

View Full Size

Figure 1

Letter dated April 11, 1931, Pouladi Collection, Yazd, Page 1

Citation: Iran and the Caucasus 29, 4-5 (2025) ; 10.1163/1573384X-02904008

© Mehraban Pouladi

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Figure 2

Letter dated April 11, 1931, Pouladi Collection, Yazd, Page 2

Citation: Iran and the Caucasus 29, 4-5 (2025) ; 10.1163/1573384X-02904008

© Mehraban Pouladi

• Download Figure
• Download figure as PowerPoint slide