Categorically misleading, dialectically misconceived: language textbooks and pedagogic participation in Central Asian nation-building projects

ABSTRACT

Persian language manuals uniformly adopt national categories such as Persian/Farsi (Iran), Dari (Afghanistan) and Tajik (Tajikistan). These categories at once impose an imagined contrast between the languages at the high register that is in fact marginal, while occluding profound linguistic variation within these nation-states at colloquial registers. Similar schemas apply to Central Asian Turkic languages such as Uyghur and Uzbek, which are closely related at the formal/literary register, but regionally diverse at lower registers. This dominant instructional approach ill prepares language learners for engaging the region on its own terms, rather than through the lens of nationalist aspirations. Students would be better served by an integrative method that teaches a transnational high language (in the case of Persian) while introducing a diverse range of dialects.

KEYWORDS:

• Language
• Persian
• Uzbek
• Uyghur
• Dari
• Tajik
• nationalism

Acknowledgments

The author is especially grateful to Daniel Sheffield and Joshua Freeman for valuable feedback, as well as Shelome Gooden for a number of terminological clarifications.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. For instance, The Modern Uzbeks: From the Fourteenth Century to the Present (Allworth 1990). Given that the term's usage as an ethnonym dates back barely a century, such a chronology now appears nonsensical (Khalid 2015, 259).

2. The Turkic languages considered here are of the Central Asian variety – i.e. Uzbek, Uyghur, and their predecessor, sometimes called Chaghatay. It does not deal with Turkish or other Turkic tongues, such as Tatar, Azeri, or those of the steppe region (Kazakh and Kyrgyz).

3. Shahsavari and Atwood's textbook enjoys the considerable advantage of being available in its entirety online (http://www.laits.utexas.edu/persian_teaching_resources/).

4. All of the Tajik language resources discussed here prefer to use the indigenized word (Tajiki) over the English equivalent (Tajik), likely taking inspiration from Farsi (as against Persian). Brian Spooner (2012, 92) directly ties this terminological shift to the partition of the larger Persian lingual world.

5. Careful readers will notice that Brookshaw and Jadidi at least acknowledge in the introduction that by ‘Persian colloquial’ they in fact mean ‘Tehrani’ (xiii).

6. Shahsavari and Atwood (80) note that ‘major cities and provinces are identifiable by their accents’, but do not point out that one of those accents – the Tehrani one taught in their textbook – is in fact one of them, thus leaving the impression that ‘spoken Persian’ is something separate from the country's ‘accents’.

7. This formulation is hegemonic to the extent that even students studying Persian in Dushanbe are offered ‘colloquial Persian’ (i.e. Tehrani); in contrast, Arabic students rarely study Egyptian colloquial in Damascus.

8. For instance, Sedighi (2015, 191) includes the lyrics to a song by Iranian singer Mohammad Nouri, with an oblique reference to its ‘regional accent’ (from Loristan). Note that Miller et al. (2014, 214). characterize Lori as an ‘advanced dialect’ broadly on par with Mazandarani or Kurdish.

9. A recent blog post offers a superb dissection of the national cultural categories deployed in Persian textbooks (Fani 2015).

10. The writing system used to transcribe the sounds that make up speech does not by itself define a language – at least by most definitions. Japanese is still Japanese, whether written in Hiragana, Katakana or Kanji. It is ironic that this most tangible difference between formal Persian in Tajikistan and its neighbours is the one most commonly cited as proof of Tajik's status as a separate language or dialect – given that it is precisely in written form that Tajik differs from Persian the least.

11. ‘Genetically’ in the sociolinguistic sense, meaning shared origins.

12. Although slightly less user-friendly, Baizoyev's book reaches a significantly more advanced level than any of the other Persian textbooks discussed here, offering exercises on complicated syntax in the later chapters (Baizoyev and Hayward 2004).

13. The participation of the Persianate elite in a ‘cosmopolis’ of high culture was broadly analogous to that of the Sanskrit realm in India during the premodern period (Pollock 2006).

14. The usage of ‘Tajik’ to refer to a language dates to the 1920s, conceived as an ‘afterthought’ to concurrent Soviet Uzbek language reform efforts (Perry 1996, 281). Curiously, the term does appear in that usage in at least one Persian manuscript prior to the Soviet period. Afżāl Pīrmastī al-Haravī's Afżāl al-Taẕkār fi Ẕikr al-Shuʿarāʾ va al-Ashʿār (ms. Insitut Vostokevedeniia Akademii Nauk Respubliki Uzbekistana [IVANUz], no. 2303, f. 105b), written in 1903, described a Central Asian poet as composing works in both ‘the Turkic and Tajik languages’, whereas most of his contemporaries would have referred to Turkī and Fārsī.

15. The textbook also briefly covers a colloquial future tense (e.g. ‘I will go’: raftanī), along with a handful of other canonized colloquialisms (e.g. ‘I can go’: rafta mī tawānam, as distinct from the literary mī tawānam [bi]rawam) (Khojayori, 2009a, 146, 160).

16. The otherwise very helpful companion grammar in fact asserts the opposite: ‘prepositions must always be immediately followed by their objects’ (Khojayori, 2009b, 44).

17. The companion grammar book includes discussion of -gī in its related use as a past participle (73), which is probably the origin of its colloquial usage as present perfect (Perry 2005, 269–270). A discussion of its use in the present perfect is included in Colloquial Tajiki Phrasebook (Aliev and Okawa 2009, 138).

18. The original provenance of boy is probably Indo-European, assuming it shares an origin with the Pahlavi bay (‘god, majesty’), though the word is long-attested in Turkic (Clauson 1972, 384). It was almost certainly adopted in Central Asian colloquial Persian from Turkic, given that the word means ‘rich’ in modern Uzbek but is not used in that sense in literary Persian. Khojayori (2009a, 87) also preserves the word for ‘already’ (allakay), which has no equivalent in literary Persian, though it is derived from the Uzbek allaqachon (kay and qachon both mean ‘when’, in Persian and Uzbek, respectively).

19. Note that Aliev and Okawa (2009, 168–172) very helpfully include a list of the Russian words most commonly used in colloquial Central Asian Persian along with the Iranian equivalent.

20. Like many nationalist programmes worldwide insisting on etymological authenticity (e.g. France), Soviet Tajik nationalist thought generally emphasized Indo-European-ness (e.g. the ancient Sogdians), but as a sizable body of scholarship has shown, the decisive split between Uzbeks and Tajiks dates only to the 1920s and 1930s (Hirsch 2000; Khalid 2015). Recently, the Tajik government has intensified nationalist language politics, mandating that all signage be in Tajik, including concepts usually understood through loan words, such as ‘dentist’ – stomatalog in Russian, now expressed with the neologism dandonpizishk (Eurasianet.org 2014).

21. Both of these words are included in some Soviet-era dictionaries and so were recognized as part of the Tajik lexicon on some level (Bertel’s 1954, 487, 489). Neither, however, was included in the even more voluminous Farhang-i Tafsīrī-yi Zabān-i Tājīkī (Dushanbe 2008). This shift maps perfectly with the shift from codification of colloquialisms in the 1950s and 1960s and the ‘re-Persianization’ of the 1980s and 1990s (Perry 1996, 283, 296).

22. To be clear, ‘Persian’ or ‘Turkic’ here refers to the language context in which the word is used; harakat qilmoq is Turkic in the sense that it is used in a Turkic language, even though ḥaraka is Arabic in origin (just like the Arabic saʿī in the Persian saʿī kardan). Whether the formulation was originally embedded in a grammatically Persian-dominant dialect or Turkic is unimportant next to the larger point that colloquial register speech patterns evolved in tandem.

23. The composition of the elite in Dushanbe (and even low-level government officials) shifted significantly following the civil war, which ended in 1997. Even though the more thoroughly Sovietized elite from the north (Khujand/Leninabad) were on the winning side of the civil war, they were junior partners alongside the set from Kulab, and this power shift has in turn affected the flavour of language spoken in the capital. While it has indeed dramatically reduced the prevalence of Russian, as the northern Tajik elite were displaced and ethnic Russians left the country en masse, it has not led to an increase in the prevalence of the ‘literary Tajik’ taught in these manuals. Quite the opposite: the ‘southern’ dialect (i.e. Kulabi) covered in Aliev and Okawa's Colloquial Tajiki Phrasebook is predominant.

24. Given that many of the textbooks discussed here focusing on Central Asian languages in particular (i.e. Khojayori, Azimova, and Nazarova and Niyaz) received Title VI funding (which also goes toward FLAS fellowships, for instance), one might assume that they ultimately have a policy audience in mind. However, as of 2016, the Foreign Service Institute uses Baizoyev's older textbook to teach Tajik, as does the American Councils language centre in Dushanbe, Tajikistan, which is now the de facto locus of Persian language immersion for American citizens.

25. This is because the formal register of Tajik/Persian found in the textbook must be learned by Tajiks as well. Recently, the mayor of Kulob ordered his subordinates back to school because official correspondence was ‘riddled with grammar and spelling mistakes’ (Najibullah 2017).

26. It is hard to overemphasize the influence of Turkic on certain registers of spoken Persian in Central Asia. Indeed, Gerhard Doerfer (1995) described the dialects in northern Tajikistan as nascent Turkic dialects, i.e. Uzbek heavily influenced by Persian forms rather than the other way around. Stephen Bahry (2016, 12) observes that it is often difficult to tell ‘where does one language end and another begin’.

27. Sociolinguistics refer to this as a dialect continuum (Wardhaugh 2009, 44–45), such that Persian and Tajik might be understood as ‘antipodes on a scale of variability, with Dari as an intermediate form’ (Beeman 2010, 139).

28. ‘Tajik literature’ is generally understood to include classical Persian poetry composed long before the word Tajik was ever used to describe a language; but when ‘Tajik’ is used to refer to many dialects in Tajikistan that are genuinely distinct from literary Persian in grammar and vocabulary, the language is neither literary nor formal.

29. John Perry's (2005, 484–486) superb reference grammar also covers Turkic influence on Central Asian Persian.

30. The older Bukharan Tajik (Ido 2007) plays a similar role for a colloquial dialect in a historical centre of Persianate culture now located in Uzbekistan (whereas Aliev and Okawa's guide is limited to the nation-state of Tajikistan), though – like Mustafa Sayd's Spoken Dari – it consists exclusively of transcribed dialogues. The classic Soviet-era study of Bukhari (Kerimova 1959) remains unmatched.

31. William Beeman (2010, 139) states that the term ‘Dari’ (and ‘Tajik’) refers specifically to colloquial speech: ‘Persian is seen by all speech communities as a prestige standard, and Tajik and Dari as colloquial forms.’ Although there is truth to this in certain communities of native speakers, both Dari and Tajik are most certainly also used to refer to the high-register language nigh indistinguishable from Persian. This imprecision has led to exactly the sort of confusion this essay seeks to clarify.

32. In other cases, the words are drawn from Central Asian Persian (e.g. pēshak for ‘cat’, spelled pishak in Tajik).

33. Alongside intense regional (but not national) differentiation at the colloquial register, neologisms stand as a partial exception to the argument here. Neologisms active in high-register discourse do vary somewhat in accordance with national borders – sort of. Most dictionaries (by any branding – Tajik, Dari or Persian) furnish an etymologically Persian word for a modern concept absent in the historical lexicon, while in practice loan words are often used for the same purpose (especially Russian in the Tajik case, but also French and English in Iran and Afghanistan, respectively). For instance, for the concept of ‘satellite’, some Tajik dictionaries favour the Soviet-era neologism (hamsafar, literally ‘fellow-traveller’, a calque on the Russian sputnik), others the Iranian one (māhwāra, literally ‘resembling a moon’), while in practice most Tajiks would probably borrow the Russian term (sputnik). Confusing matters even further, specifically Iranian neologisms are increasingly being adopted into Tajik and used in parallel with the Russian and specifically Tajik neologisms (Perry 1996, 303–304). Most of the textbooks discussed here, however, do not progress to an advanced enough level for this to be a relevant issue.

34. The continuity drawn between modern Dari by Sayd in his introduction and ‘Dari’ as used to describe the court language of the very early Islamic period is entirely spurious: ‘Modern Dari began to develop in the 9th century’ (Sayd 2010, ii). The ninth-century court language of Dari Sayd is referring to has been investigated in several relatively recent publications (Lazard 1995; Perry 2012). In fact, the origin of associating the word Dari with the Persian language in Afghanistan was a political decision made in 1964 by King Amanullah of Afghanistan as a means of counteracting Iranian influence and placating Afghan nationalists (Spooner 2012, 99). This move is not so different from efforts to tie ancient usages of the word ‘Tajik’ (extant even in Middle Persian, but with a very different meaning) to the modern nation-state.

35. Cf. an interaction between speakers of the Kabuli ‘accent’ and the Kohdamani ‘dialect’ on p. 34, or the ‘Herati dialect of Dari, which is influenced by the Iranian Farsi’ on p. 232 (Sayd 2010).

36. Oddly, this report’s nuanced appreciation of Iran's dialect continuum is belied by occasional invocation of the nationalized understanding of Persian: ‘The Persian language as spoken today can be roughly divided into three varieties depending on the country in which it is spoken: Farsi, spoken in Iran, Dari, spoken in Afghanistan, and Tajiki, spoken in Tajikistan’ (Miller 2014, 7). However, the authors are clearly aware of the limitations of this tripartite schema: ‘The more colloquial language a speaker uses, the more distinction can be seen between Farsi and Dari’ (10).

37. However, the report is part of an effort to develop an ‘intermediate and an advanced course in the accents and regional dialects and languages of Iran’ (Miller 2014, 5).

38. Students and instructors in need of more emphasis on grammar might consult An Introduction to Persian (Thackston 1993) and Persian Grammar and Verbs (Hillman 2012). Thackston's textbook remains in many ways unsurpassed.

39. ‘Turkic’ refers broadly to a language group, including Turkish, Uzbek, Azeri, Kazakh, Kyrgyz, and others.

40. While I am focusing on the Turkic languages of the sedentary Central Asia that interacted with Persian the most extensively (and with which I am most familiar), the same could be said about the relationship between Kyrgyz and Kazakh, which are mutually intelligible and probably even closer to one another than Uyghur and Uzbek.

41. Nigora Azimova's Uzbek: An Elementary Textbook and Gulnisa Nazarova and Kurban Niyaz's Uyghur: An Elementary Textbook are timely additions to a rather uncrowded field of Central Asian Turkic language textbooks. Greetings from the Teklimakan (Engesæth, Yakup, and Dwyer 2009) stands as one of the chief alternatives to these newer works. One notes that these authors do not follow Khojayori's lead in deploying the emic term for the language. In other words, we have ‘Tajiki’ but not ‘O’zbekcha’.

42. Unlike the Soviet Uzbek Latin alphabet of the 1930s, and unlike Turkish, the modern incarnation does not use any diacritically modified characters.

43. I am using ‘eastern Turki’ and ‘Central Asian Turkic’ as a neutral means of broadly referring to Uzbek/Uyghur and their literary predecessor (sometimes referred to as Chaghatay), analogous to the usage of ‘Central Asian Persian’ previously.

44. The BBC Uzbek service offers coverage in Cyrillic, Latin, and even Arabic script for the Uzbek population in northern Afghanistan (as distinct from the substantially modified Arabic script used for modern Uyghur).

45. Happily, Azimova's textbook does include a ‘Cyrillic reader’ in an appendix, which consists of a reference chart, examples, and 50 pages of exercises and sample texts. Beyond this, language learners can consult Uzbek Textbook (Ismatulla 2001).

46. This is because so much of the vocabulary in literary eastern Turki / Chaghatay is Persian and Arabic loan words, which are effectively transliterated in the new script (since many Arabic letters are redundant in both Persian and Turkic). For instance, tem (‘taste’) would have been written as ṭamʿ in literary eastern Turki – i.e. with two out of three completely different letters. Students coming from Uyghur and Ubzek (ta’m) are essentially in the same boat for most loan words.

47. One notes an interesting parallel between the modifications made to the Uyghur modified Arabic script and the officially promulgated Standard Chinese Characters, both of which have the end result of divorcing the modern language from the deeper heritage (Sahlins 2013).

48. This is almost directly analogous to buzūrg and kalān (‘big’) in Iranian vs. Central Asian Persian, both of which are active in both regions, but have inverse nuances in meaning.

49. Significantly, there is no single schema for describing the numerous local languages considered dialects of Uyghur (Abdurehim 2014, 26–27).

50. Geographically, this makes perfect sense, as Khorezmian Uzbek is influenced by Turkmen, which is from the Oghuz branch of Turkic languages, sharing features with Azeri and Turkish.

51. Although Chaghatay is the name used to describe literary eastern Turki in the few manuals that do exist (Bodrogligeti 2001; Eckman 1966), it was very rarely used to describe a language in pre-twentieth-century sources – and even in the few such uses, it seems to have been referring to a literary style rather than a distinct tongue (Péri 2003).

52. This is not to say, of course, that New Persian remained completely static over time (Paul 2002).

53. The two manuals that do exist are indispensable, but take a densely grammatical approach and draw examples from classical literature mostly from the sixteenth through eighteenth centuries, as opposed to the chronicle or document traditions from the nineteenth-century khanates (Eckman 1996; Bodrogligeti 2001). As far as dictionaries, aside from Clauson's etymological work, scholars tend to use Budagov's (1869) reference, which covers all Turkic dialects of the Russian empire. A Soviet dictionary (Fazylov 1966) details Khorezmian Turki, and a Turkic–Persian dictionary written by one of Nadir Shah's chroniclers remains useful to the present day (Mahdī Khān 1995).

54. Wheeler Thackston designed a class at Harvard, and wrote an accompanying textbook, in which Levantine Arabic was taught through transliteration with no reference to MSA – a rare exception to the overwhelmingly predominant trend.

55. Beyond sections in the text itself, the Al-Kitaab language programme includes companion websites for students wishing to specialize in regional dialects other than Egyptian (see e.g. Haki bil-Libnani at http://www.alkitaabtextbook.com/bookstore/52023.html).

56. The differences between MSA and Classical Arabic have been described as ‘stylistic and lexical rather than grammatical’ (Bassiouney 2009, 12).

57. Before the Napoleonic period, French peasants spoke Latinate dialects of varying degrees of mutual intelligibility (i.e. a dialect continuum).

58. This point is distorted somewhat in the secondary literature by a disagreement over the matter of diglossia. Taking the lead of Charles Ferguson, sociolinguists are generally in agreement that Arabic is a classic example of diglossia in that there is a relatively clear distinction between the high and low registers of the language (Bassiouney 2009, 13). Yet controversy over this issue remains (Miller 2014, 9), and some linguists argue for Persian's status as a diglossic language, similar to Arabic (Jeremiás 1984). Although the degree of diglossia varies with population education, it does appear that literary Persian and Tehrani are less diglossic than Egyptian and MSA (Perry 2003, 11). However, even if Arabic is on average more diglossic than Persian (there has been little concrete research), the difference is one of degree rather than kind, and the existence of a complex array of regional dialects connected to a common high language through social registers is clear in all cases. Moreover, even the aspirational categories forwarded in Persian textbooks of all stripes undermine the notion of Persian as a ‘standard with dialects’. If the Persian spoken in Tehran is the ‘standard’, then where does the ‘spoken Persian’ of these textbooks fit, given that it is spoken primarily in Tehran?

59. Despite being a rather uncommonly studied language, Pahlavi boasts an excellent textbook: Prods O. Skjærvø, Pahlavi Primer (unpublished).

60. Hindustani (along with several other terms) refers to a single language widespread in India before the appearance of modern nationalist movements, which invented and codified separate Urdu and Hindi variations over the late nineteenth and early twentieth centuries.

61. For instance, the Hindi Urdu Flagship Program at the University of Texas at Austin: Two Languages or One (http://hindiurduflagship.org/about/two-languages-or-one/). Despite this trend, I am not aware of any textbooks that fully integrate Urdu and Hindi into a single course, with the partial exception of Conversational Hindi-Urdu (Gumperz 1967), which comes in a Devanagari-script edition, but not an Arabic-script version.

62. Urdu language reformers tapped into Arabic even more than Persian to invent vocabulary for modern concepts, tending rather to excise both Indic and Persian elements from the language (Rahman 2006, 114).

63. Intriguingly, interest in a particularly Central Asian dialect of Persian through Judeo-Persian may have predated its Arabic-script counterpart. Having immigrated from Bukhara to Jerusalem in 1890, Shim’on Ḥakham (1986, xxii–xxiii) highlighted the Bukharan features of his Persian writing and endeavoured to communicate with Persian-speaking Jews from Central Asia and Afghanistan, just as Arabic-, Spanish- and Yiddish-speaking Jews could read texts in their native tongues.

64. Bukhārā-yi Sharīf, no. 26 (10 April 1912), f. 3. The pan-Aryan idea, however, did not ultimately gain much traction (Khalid 1998, 209).

65. The letter to the editor complained of pervasive Arabic vocabulary used in Persian as ghayr al-fahm (incomprehensible). If this Arabic syntax and vocabulary was intended ironically, the sarcasm was lost on the editor, who proceeded to systematically parse out the Arabisms in the letter. Bukhārā-yi Sharīf, no. 13 (26 March 1912), f. 2.

66. Bukhārā-yi Sharīf, no. 11 (23 March 1912), f. 2; Bukhārā-yi Sharīf, no. 8 (20 March 1912), f. 2.

67. The Bukhari language transcribed in the newspaper seems to be very much a conceptual precursor to literary Tajik in that it contains patterns which are uniquely Central Asian but eschews, for the most part, Turkic vocabulary and syntax. Bukhārā-yi Sharīf, no. 8 (20 March 1912).