1. The terms “Northwestern Iranian” and “Southwestern Iranian” refer to a clear linguistic division of the languages of Iran proper and others located to the west and northwest of the borders of Iran. The following lists are a rough outline of the languages included in these two groups:

• A. The Northwestern group referred to throughout this article as Central Plateau Languages is usually called Central Plateau Dialects.

• B. For a complete listing of all villages whose dialects are mentioned in this paper, including some villages where other Central Plateau Languages are spoken, the reader is referred to Appendices I and II which consist of maps of north central Iran and the Tāti language area respectively.

• C. The use of the term Tāt, other than the Northwestern group discussed in this article, refers to a language of Southwestern Iranian origin spoken in the southern Caucasus in the Azerbaijan Soviet Socialist Republic. The language has two main branches, Muslim Tat and Jewish Tat.

• D. There is some dispute over the classifications of the different forms of Kurdish as a Northwest or Southwest Iranian language. For the present discussion I have assumed a Northwest origin for Kurdish and follow the classification of Kurmanji Kurdish as Northern Kurdish, Sōrāni/Mukri as Central, and Kermānshāhi as Southern Kurdish.

2. Various brief references to the interrelationship of Tāti languages have been culled from the following works of Prof. Yar-Shater: Yar-Shater, E., “The Tāti Dialects of Rāmand,” in A Locust's Leg: Studies in Honour of S. H. Taqizadeh, eds. Henning, W. B. and Yar-Shater, E. (London: Percy, Lund, Humphries and Co., Ltd., 1962), pp. 240–245Google Scholar; idem, “The Dialects of Alvir and Vidar,” in Indo-Iranica: Melanges Morgenstierne (Wiesbaden: Otto Harrassowitz, 1964), pp. 177–187Google Scholar; idem, A Grammar of Southern Tati Dialects (The Hague/Paris: Mouton, 1969)Google Scholar; idem, “The Dialect of Shāhrud (Khalkhāl),” Bulletin of the School of Oriental and African Studies, XXII (1) (1959), pp. 52–68.Google Scholar See Footnote 11 below for citations on other articles by Yar-Shater on Tāti languages.

3. Marāghei dialects of the Rudbār-e Alamut area refer to the dialects spoken by a group of people who have a secret religion and make a rigid distinction between themselves and the Shi'ite Muslim villagers whom they call “Pashei.” The Marāghei religion may be a form of Ismaili sect or, more likely, a mixture of elements from different religions of Iran and the Middle East. The Marāghei language is a form of Tāti and is totally unintelligible to the speakers of Pashei Tāti. Marāghei Tāti is spoken in some sixteen villages including Dikin (which I was able to visit), Mushqin, Garmābād-e Pāin, Aliābād, Avirak, Vashte, Dorchāk, Yavej, Aspemord, Sutkaš, Angerāzuj, Aleyn, Sapuhin, Vartovān, Zanāsuj, and Kashābād-e Pāin. Since I did not expect to find their form of Tāti in this territory, I would conjecture that the Marāghei speakers are immigrants to the area. Further study of the field notes collected on their language may reveal more information about its possible affinities. The Tāti of the Pashei speakers in the same villages is a form of Tāti that I would have expected for this area since it is not dissimilar to the Tāti spoken in Gozarkhān and Owrazān, discussed in the studies of Ivanow and Al-e Ahmad, respectively.

4. E. Yar-Shater, op. cit. (1959); Henning, W. B., “The Ancient Language of Azerbaijan,” Transactions of the Philological Society (London, 1954), pp. 155–177Google Scholar; Miller, B. V., Talyšskii Iazyk (Moscow, 1953), pp. 227–237, 254-266.Google Scholar

5. A distinction is made here between Southern Tatic as handled in this article, and Southern Tāti, which refers to the Tāti dialects spoken in the Qazvin area and investigated by Prof. Yar-Shater in his Grammar of Southern Tati Dialects.

6. While collecting material in the field on Tatic languages, I realized how common multilingual situations are in Iranian villages. For example, the people of Vafs with whom I had contact, all spoke Vafsi, Azerbaijani Turkish, and Persian in that order of usage. In Kabate, everyone seemed to speak Kabatei (Tāti proper), Rudbāri (preliminarily: a Caspianized Tatic language), Gilaki (Caspian), and Persian, in addition to which most of the men and some of the women spoke Azerbaijani Turkish. My observations in both villages included elderly women who had never left their native villages. In the cases of some of the villages, their Tāti language serves as their only language through childhood, and the non-Tāti languages, usually including Azerbaijani and Persian, are not learned until adolescence. In some villages, Tāti speakers are bilingual or multilingual from a very early age. The village of Dikin in Rudbār-e Alamut, for example, has three different languages (Persian; Tāti-“Pashei,” Owrazāni type; and Tāti-“Marāghei,” Tāti proper). These languages are spoken concurrently in Dikin owing to the presence of three different groups in the village. The Persian speakers (Shi'ite Muslim) are monolingual; the Tāti-“Pashei” speakers (Shi'ite Muslim) all speak Tāti-“Pashei” and Persian, whereas only the Tāti-“Marāghei” speakers (Marāghei sect, non-Muslim?) are completely trilingual.

7. A very encouraging new step in this direction has been made by Masica, Colin P. in Defining a Linguistic Area: South Asia (Chicago/London: University of Chicago Press, 1976)Google Scholar. Dr. Masica examines in detail certain grammatical features that group languages across most of Eurasia into different subtypes, irrespective of their genetic affinity. A dynamic approach to Iranian dialectology is seen in a recent dissertation presented at the University of Michigan by Karl J. Krahnke, Linguistic Relationships in Central Iran (1976). His methodologies present a dynamic classification of Iranian dialects in the central plateau area and are not rigidly bound to the impossible task of genetic classification, as is the case with the work of his predecessors in this area. Through the use of isogloss maps for his area of study, Krahnke shows that it is impossible to delineate discrete genetic groups and subgroups in Central Plateau Languages. Unfortunately, Krahnke's dissertation came to my attention only recently. Because this article was already completed, I was unable to include here a discussion of his methodologies and results.

8. Important statements regarding convergence are made in Gumperz, J. J. and Wilson, R., “Convergence and Creolization: A Case from the Indo-Aryan/Dravidian Border in India,” in Pidginization and Creolization of Languages: Proceedings of a Conference Held at the University of the West Indies, Mona, Jamaica, April 1968, ed. Hymes, Dell (Cambridge, 1971), pp. 151–167.Google Scholar Speaking of the phenomenon of convergence in the village of Kupwar in Central India, where Marathi (Indo-Aryan), Urdu (Indo-Aryan), and Kannada (Dravidian) are spoken, the authors point out (p. 155) that the similarity among the grammatical structures of these three languages is so great, “that we were able to analyse an extensive corpus of bilingual texts involving all three local varieties without having to postulate syntactic categories or rules for one language which were not present in the other language. We may say, therefore, that the codes used in code-switching situations, in Kupwar have a single syntactic surface structure.” This statement is further supported by the claim that “For many Kupwar residents…a model of linguistic competence must comprise a single semological, a single syntactic, and a single phonetic component, and alternative set of rules for the relation of semantic categories to morphemic shapes” (p. 165). A similar case is made for Finnish and Lapp in Anttila, R., An Introduction to Historical and Comparative Linguistics York: The Macmillan Company, 1972), p. 169Google Scholar, in the claim that “Finnish syntax is largely Indo-European (Baltic and Germanic, apparently Swedish for the most part). Lapp syntax in Finland is strongly Finnish--in fact, almost like Finnish with Lapp words, but in Norway is markedly Norwegian.” The case for substratum influence on syntax is made for the Cushitic substratum in Ethiopic (Semitic) languages by W. Leslau in “The Influence of Cushitic on the Semitic Languages of Ethiopia--A Problem of Substratum,” Word, I (1) (New York, 1945). Aside from examples given in the realms of phonology and morphology, the author states that “It is the syntax that gives to Ethiopic, and especially to the modern languages, their particularly non-Semitic character. The Cushitic substratum is especially clear in matters of syntax” (p. 73).

9. The term “isorule” was coined and communicated to me orally by Dr. Gernot L. Windfuhr of the University of Michigan.

10. The following sources were used for the languages and dialects cited in the examples in the rest of this article and on the isomaps: Tatic--Alviri: Hāshemi, J., “Guyeš-e Alviri” (Ph.D. Dissertation, University of Tehran, Department of Linguistics, 1353/1974)Google Scholar; Vidari: E. Yar-Shater, op. cit. (1964); Southern Tāti (including Eshtehārdi, Tākestāni, Esfarvarini, and Chāli): E. Yar-Shater, op. cit. (1969); Upper Tāromi: Yar-Shater, E., “The Tāti Dialects of Tārom,” in The Henning Memorial Volume, eds. Boyce, M., Gershevitch, I. (London, 1970), pp. 451–467Google Scholar; Khoini: “Khoini--yeki az lahjehā-ye āzari,” Farhang-e Irānzamin, 6 (Tehran, 1958), pp. 324–327Google Scholar; Gozarkhāni: Ivanow, W., “The Dialect of Gozarkhon in Alamut,” Acta Orientalia, 9 (1931), pp. 352–369Google Scholar; Owrazāni: J. Al-e Ahmad, Owrazān (Vaz'- e Mahall, Ādāb va Rosum, Folklur, Lahje) (Tehran, 1954); Shāhrudi: E. Yar-Shater, op. cit. (1959); Kajali: idem, “The Tāti Dialect of Kajal,” Bulletin of the School of Oriental and African Studies, XXIII (2) (1960), pp. 275–286Google Scholar; Lerdi: idem, “The Distinction of Feminine Gender in Southern Tāti,” in studia classica et Orientalia Antonino Pagliaro Oblata, III, ed. Belardi, W. (Rome, 1969), p. 300Google Scholar; Harzani: Kārang, A. A., Tāti va Harzani (Tehran: Isma'il Vā'ezpur, 1954)Google Scholar; Keringani: idem, and Zokā, Y., Guyeš-e Keringān (Tehran: Ketabxāne-ye Dāneš, 1954)Google Scholar; Northern Talyshi: B. V. Miller, op. cit. (1953); Central Plateau Languages--Āmorei: Sā'eb, R., “Guyeš-e Āmore” (Ph.D. Dissertation, University of Tehran, Department of Linguistics, 1976)Google Scholar; Meimei: Lambton, A. K. S., Three Persian Dialects (London, 1938)Google Scholar; other Iranian Languages--Baluchi: Rastorgueva, V. S., “Beludžskii Iazyk,” in Iazyki Narodov SSSR, I, Indoevropeiskie Iazyki (Moscow: Nauka, 1966), pp. 323–341Google Scholar; Central Kurdish (Sōrāni): Abdulla, J. J. and McCarus, E. N., Kurdish Basic Course, Dialect of Sulaimania, Iraq (Ann Arbor, 1967)Google Scholar; Northern Kurdish (Kurmandji): Kurdoev, K. K., Grammatika Kurdskogo Iazyka (Moscow/Leningrad: Akademiia Nauk SSSR, 1957)Google Scholar; Gurāni (Auromani): The Dialect of Awroman (Hawrāmān-ī Luhon): Grammatical Sketch, Texts, Vocabulary, Det Kgl. Danske-Videnskabernes Selskab, Historisk-Filologiske Meddelelser, 4 (3) (Copenhagen, 1966); Caucasian Tat: Griunberg, A. L., Iazyk Severoazerbaidžanskix Tatov (Leningrad: Akademiia Nauk SSSR, 1963).Google Scholar My own field notes were used for examples in the following dialects and languages: Tatic--Vafsi, Razajerdi, Dikini/Marāghei, Dikini/Pashei, Kalāsi, Kabatei, Southern Talyshi (Māsulei and Māsāli), and Central Talyshi (Asālemi); Caspian--Gilaki (Rashti and Lāhijāni dialects), Māzanderāni (Bābolsari dialect), and Shahmirzādi; Transition Tatic/Caspian--Rudbāri; Central Plateau Languages--Se-Dehi, Esfahani Jewish, Yazdi (Zoroastrian), and Abuzeydābādi; Other Iranian Languages--Bakhtiari; Non-Iranian Languages--Azerbaijani Turkish, Armenian, Romany (of Ābyek area), Neo-Aramaic (Assyrian), and Arabic. The transcription system of my own examples is the same as that used throughout this article with the exception of the vowels of Māzanderāni; the Mazanderani vowel equivalent to the /e/ of Persian and English (mid vowel) is represented as /E/, whereas /e/ is used to represent a high mid-vowel phoneme in Mazanderani, similar to the “é” of French.

11. The nominal system (including pronouns) of the Tatic languages and many other Northwestern Iranian languages differs drastically from that of Modern Persian (I use Persian for contrast as a point of reference for most readers of Iranian studies). Nouns in these languages have a two case system: Direct Case and Oblique Case. The functions of these two cases are listed below:

Direct

1. 1.

   1. subject of intransitive verbs in all tenses

2. 2.

   2. subject of transitive verbs in the present tenses only

3. 3.

   3. object of transitive verbs in the past tenses only (optional, depending on the language)

4. 4.

   4. destination without preposition (“I'm going home” “man miram xune”).

Oblique

1. 1.

   1. subject of transitive verbs in the past tenses only

2. 2.

   2. object of transitive verbs in the present tense (and in some languages in past tenses also)

3. 3.

   3. possessive

4. 4.

   4. indirect object

5. 5.

   5. object of prepositions (optional)

6. 6.

   6. preceding postpositions

The languages of this system are known as “ergative” languages. The ergative construction has also been referred to in the Iranist literature as the “passive” or “agentive” construction, although sentences of this type are by no means passive in meaning. The ergative construction contrasts with the nominative construction of Persian and other languages. In the nominative type sentence, the subjects of intransitive and transitive verbs are in a direct or nominative case (often the unmarked case, i.e., with no case ending) and the direct object of transitive verbs is marked by an accusative case or an accusative marker such as /-ra/ of Persian. In the ergative type sentence, as seen above in the list of functions of the two cases, the subjects of intransitive verbs and the object of transitive verbs are represented by the same case, the Direct case (or unmarked case) in Iranian languages. The subject of transitive verbs is then marked by a special case used only for this syntactic function; this case has been named variously the ergative case or the agentive case. The contrast of the ergative and nominative systems in Iranian languages is demonstrated in the following schema:

Iranian languages have an ergativity which is tensebased; this is not necessarily true of ergative systems in other languages. That is, in their pure forms, Northwest Iranian languages are nominative in the present system and ergative in the past system. This tensebased ergativity, therefore, yields a flip-flop of the case functions according to tense of the transitive verb:

Examples from Vafsi:

In the instance of pronominal cases, some Northwest Iranian languages have the same system of cases in pronouns as in nouns. In other languages, however, some of the functions listed above under “oblique” case are taken over by specialized pronoun case forms which may number from three to five, depending on the language. Hezārrudi (Tāromi), for example, has five cases in the pronoun:

Case One: Subject of intransitive verbs, subject of (Direct) transitive verbs in the present tense, destination--i.e., most “direct” functions

Case Two: Possessive; object of the postposition (Genitive)

Case Three: Subject of transitive verbs in the past (Ergative) tense

Case Four: Object of transitive verbs in the present tense

Case Five: Object of transitive verbs in the past tense.

In the Iranist literature concerning the use of cases in Iranian languages and dialects, Case Three is usually referred to as the “agentive” case, Case Four is the “objective” case, and Case Five is the “logical direct object” which refers only to objects of past tense verbs in the ergative system.

12. Individual isomaps appearing in the text of this article do not indicate the names of individual dialects mentioned. Isomaps correspond exactly to the maps of Appendices I and II, where the names of the villages are given in full.

13. There seems to be a relationship between the č/j morpheme of Tatic and the oblique marker /ž/ in the third person pronouns of Sorkhei and Semnāni, as already mentioned in W. B. Henning, “The Ancient Language of Azerbaijan,” op. cit., p. 162:

It is too early to say, however, whether these are parallel developments or whether they can be considered the same isogloss as Tāti.

14. The independent possessive/oblique pronouns in Alviri/Vidari seem to follow their head nouns, connected to them by the ezāfe (which, as in Persian, connects noun with noun in a genitival function and noun with adjective in adjectival modification). More field work is necessary to test whether these pronouns may also precede their head nouns.

15. Yar-Shater, Grammar of Southern Tati Dialects, p. 21.

16. Historical retention of lexical roots (as opposed to morphemes) in the languages of northern Iran as a tool to use in the establishment of genetic affinity has not been discussed in this paper simply because no systematic comparative work has been done on the lexical composition of the members within each group of Northwestern Iranian languages nor have comparative studies among the different groups been conducted. We do not even understand clearly the criteria for establishing Tatic, Central Plateau and Caspian languages, Kurdish, Gurāni, Zaza, etc. as separate groups. Even less do we understand the interrelationships and hierarchical ordering of the relationships of groups of Northwestern Iranian languages, and certainly the transitional areas between these groups have never been established on a systematic basis.

17. It cannot be said that grammatical morphemes are never borrowed. There are many clear examples of morphemeborrowing from one language into another. There are also examples of extreme cases of borrowing in which whole morphemic systems of one language are taken over in toto by another language, and the new set of morphemes, calqued upon the roots of the original language, form a type of unusual hybrid language. Such may have been the case to some extent in the Rudbāri dialects. They seem to have been originally members of the Tatic languages which now show a great percentage of Tatic lexical roots but share most, or possibly all, of the morphemic system of Gilaki, a language of the Caspian group. Such extreme cases of morpheme borrowing in languages must be studied more carefully to understand the reasons for, and mechanics of, this process.

18. The dialects of the Semnān area and of Āmore, which seem to be transitions between Tatic and Central Plateau, are possible exceptions in that they may prove to share these prefixes in their morphology.

19. W. B. Henning, op. cit., p. 175.

20. Elwell-Sutton, L. P., “The Vafsi Dialect (North-western Persia),” in Trudy XXV Mezhdunarodnogo Kongressa Vostokovedov (9-16 Avgusta, 1960) (Moscow: Vostochnaia Literatura, 1960), p. 315.Google Scholar