Peoples and Languages of Nuristân

The following ethnolinguistic overview is a revision of a paper given at the 3rd Himalayan Languages Symposium, University of California at Santa Barbara, on 18 July 1997. The symposium was supported with funds from the Wenner-Gren Foundation.



An Overview of The Nuristâni Languages

by

Richard F. Strand


The five languages of Nuristân are spoken by about 70,000 people in the Hindu-Kush mountains of northeastern Afghanistan. Nuristânis inhabit the drainages of three main rivers: the Alingar in the west, the Pech-Pârun in the center, and the Lanḍai Sin in the east. The latter two flow into the Kunar-Chitral River. I list the languages here in a north-to-south order.

Transcriptional notes:

Vowels: a represents the unmarked vowel in each language. In groups 1, 2, and 3 it is a high central vowel, while in groups 4 and 5 it is a low central vowel. a contrasts with â, which is articulated with the jaw more open. Nasalized vowels are indicated by a following ˜. Stress is indicated by ' before a vowel in groups 1 and 2, where it is phonemic. In the remaining groups (including Tregâmi?) stress falls automatically on the last syllable of a stem.

Consonants: Retroflex (apico-postalveolar) consonants are indicated by letters with underposed dots or overposed hacheks: ṭ, ḍ, c̣, J̣, ṣ, ẓ, ṇ, and ř ň. ř represents a retroflex approximant, as opposed to a tapped r or flapped . is a retroflex nasal stop, with a flapped intervocalic allophone in group 4. ň is a retroflex nasal flap intervocalically in group 3, but a nasalized ř elsewhere. Palatal (lamino-alveolar) consonants include the affricates č and ǰ and the spirant š. ć is a voiceless dental affricate; ź is a voiced dental affricate in groups 1 and 2; z is a voiced lamino-dental spirant in the other groups.

  1. vâs'i vari ('Vâsi Language') is spoken in upper, middle, and lower dialects by the vâs'i people of the Pârun Valley.
  2. kâmk'ata mumkṣt'a viri ('Kom, Kâta, Mumo, and Kshto Language') is spoken by the kât'a, mum'o, kṣto, and kom peoples who range across the breadth of Nuristân. The Kâta speak two main dialects, Western and Eastern kât'aviri. Western Kâtaviri is spoken in řâmg'al, as they call the upper Alingar Basin, in the kt'ivi Valley off the upper Pech, and in the p'ařuk Valley off the upper Lanḍai Sin. Each of these regions has its own subdialect. Eastern Kâtaviri is spoken in the upper Lanḍai Sin and in a few villages across the border in Pakistan's Chitral District. The Kom and the Kshto speak another dialect, kâmv'iri or kṣt'aviri, depending on which ethnic group one addresses; the Kom inhabit the lower Lanḍai Sin and upper Kunar basins, with some Kom spilling over into southern Chitral. Three client peoples, the ǰâš'i, the bini'o, and the ǰâmč'o, live within Kom territory and speak Kâmviri. The Kshto inhabit one village surrounded by Kom. The Mumo speak a transitional dialect, mumv'iri, between Kâmviri and Eastern Kâtaviri in the middle Lanḍai Sin Valley.
  3. âṣkuňu vi:ri is spoken by the âṣku˜ people in separate dialects in the bâźâigal, kolâtâ˜, and titin Valleys off the middle Alingar, and by the saňu and gřâmsaňâ peoples of the middle Pech Valley, each with their own dialect.
  4. kalaṣa alâ is spoken by the kalaṣa people of the Vaygal Valley off the lower Pech and to the east in the veligal and d'u˜gul Valleys off the middle Kunar. A major dialect division separates the varǰan people of the upper Vaygal Valley from the čima-nišey people of the lower valley. Minor dialect divisions separate each village, especially among the Varǰan, where the chief subdialect is that of the vay people of the village of vaygal.
  5. tregâmi is spoken in three villages of the Tregâm Valley off the lower Pech.

On the basis of shared features, Groups 1 and 2 form a northern cluster, opposed to a southern cluster consisting of Groups 3, 4 and 5. However, there is much overlapping of dialect traits.

The phylogenetic position of these languages within Indo-European was disputed by earlier scholars, until Georg Morgenstierne of Oslo University established that they formed a third branch of Indo-Iranian, in between Indo-Aryan and Iranian. I can add little to his view, except to flesh out a best-guess scenario for the history of these languages, based on native traditions, recent archeological syntheses of scholars like Marija Gimbutas, and my own observations of phonetics in the field.

Around the middle of the fourth millennium B.C., some 800 years after the first Indo-European peoples expanded out of their Volga Basin homeland into Europe, new waves of horse-mounted tribesmen who called themselves Aryas expanded south and east around the Caspian Sea from the Volga Basin, driving other Indo-European speaking peoples before them. Those Aryas who spread south into the region between the Caspian and Black Seas bumped up against the Caucasus Range, which for some fifteen hundred years served as their southern border. These southern Aryas were the precursors of the Indo-Aryan speakers of the Indian Subcontinent. By the beginning of the third millennium the remaining Aryas to the north had expanded to the west around the Black Sea, into Europe's Danube basin, and to the east around the Caspian Sea, into the basin of the Amu Darya. These were the precursors of today's Iranian speakers. Ahead of the eastern Iranian expansion were driven the precursors of the Nuristânis.

The earliest Aryas distinguished themselves expressively through the processes of harshening, apical suppression, and prognathizing. Harshening lowers the spectral frequency of a fricative by apicalizing and retroflexing it. Apical suppression holds the tongue's apex firmly fixed to the back of the lower teeth as the default articulation; only during retroflex, dental, and perhaps labiovelar sounds, does the apex leave its suppressed position. This process produces the palatalization of velars next to front vowels. Prognathizing is the process of jutting out the jaw to present a belligerent demeanor. Combined with apical suppression, it moves the blade of the tongue into contact with the upper teeth, so that palatal affricates become dental and apicodental sounds become laminodental. This process is also responsible for the change of postconsonantal v to a stop in may languages of the region. These three processes have acted periodically in the history of the Aryan languages up to today, as any phonetician who has observed the harshening and belligerent prognathizing of Kunar Afghans can attest.

One way to account for the early differentiation of proto-Nuristâni from the rest of the Aryan group is to assume that the Aryas first initiated a harshening of s after u. The proto-Nuristânis did not participate in that process, because they were beyond the influence of Aryan at that time. Thus, the word for 'mouse' is mus'a in Kâtaviri but muṣ'a in a neighboring Indo-Aryan language like Gâhwar bâti.

Summary of Nuristani Developments:
1st Palatalization
'dog'
*kuon-'e-
*čuon'e-
*čuan'e-
*ćuan'e-
*ćuan'a-
*ćun'a-
*ć'auna-
*ć'aun-
*ć'un-
 ć'u˜
'ten'
*d'ekm
*d'eka
*d'eča
*d'eća
*d'aća
*d'ać
 d'uć
'knee'
*g'enu-
*ǰ'enu-
*ǰ'ênu-
*ź'ênu-
*ź'ânu-
*ź'ân-
*ź'â˜-
 ź'o˜
'snow'
*ghim'o-
*ǰhim'o-
*ǰhim'a-
*ǰim'a-
*źim'a-
*ź'ima-
 źim

Later, when the Proto-Nuristânis were within the Aryan sphere, both groups underwent apical suppression, which caused the plain velar stops to become palatal affricates, so that, for example, PIE *kuon- 'dog' became *čuon-, *d'ekm 'ten' became *d'eča, *genu- 'knee' became *ǰ'ênu-, and *ghi-m'o- 'snow' became *ǰhim'o-. This process also caused the laminalization of s after i, as in *nišanna- 'seated', and it may also have been involved in the change of Indo-European schwa (vocalic H) to i.

Other assimilations and further harshening changed s to after r and k throughout the region. Later all harshened s's, that is, those after u, i, r, and k, merged to become š in Iranian and in Indo-Aryan. In proto-Nuristâni š after i (as in viš 'poison'; cf. Sanskrit viṣ'a-) and after r and k remain apart, while s after u remains unharshened.

Throughout the Aryan sphere there was, after the first palatalization, a loss of lip-rounding accompanying the labiovelar stops and the vowel o; the labiovelars , , and ghʷ became k, g, and gh, and o merged with a.

At this point the pronunciation of the palatal affricates diverged. The Indo-Aryans dropped the affrication of the voiceless palatal, producing a palatal spirant, so that *čuna- became Sanskrit šuna- 'dog' and *d'eča became d'aša 'ten', while the voiced palatals remained (*ǰ'ênu- 'knee', later ǰ'ânu-; *ǰhim'o- 'snow', later him'a-). The Iranians prognathized to produce dental affricates, and the proto-Nuristânis, being under Iranian influence, followed suit, so that *čuna- became Kalaṣa alâ ću˜ 'dog' (through strengthened *č'âuna-), *d'eča became Kâmviri duć 'ten', *ǰ'ênu- 'knee' became Kâmviri źo˜, and *ǰhim'o- 'snow' became Kâmviri źim.

2nd Palatalization
'migrate'
*kʷel'e-
*kel'e-
*čel'e-
*čal'a-
 čâl'a-
'five'
*p'enkʷe
*p'enke
*p'enče
*p'anča
*p'anč
*p'a˜č
*p'u˜č
 p'uč
'alive'
*gʷiHu'o-
*gʷi:v'o-
*gi:v'o-
*ǰi:v'a-
*ǰiv'a-
*ǰü'a-
 ǰ'üa-
'smite'
*ghʷ'ene-
*gh'ene-
*ǰh'ene-
*ǰh'ana-
*ǰan'a-
*ǰaṇ'a-
*ǰâṇ'a-
*ǰâň'a-
 ǰ'âňa-

Throughout the Aryan-speaking region the default apical suppression continued, producing a second palatalization of the remaining velars (from labiovelars) before palatal sounds (i and e). In Indo-Aryan the resulting voiced affricates ǰ and ǰh, as in the development of PIE *gʷiH-u'o- 'alive' > *giHu'o- > Skt. ǰi:v'a- and PIE *ghʷene- 'smite' > *ghene- > *ǰhene- > Skt. h'ana-, were identical to those of the previous palatalization, while the new voiceless affricate č, as in *kʷele- 'migrate away' > *čele > Skt. č'ala-, contrasted with the spirant š from the earlier palatalization. In both Iranian and proto-Nuristâni, the older dental affricates ć and ź remained fully distinct from the new palatal affricates č and ǰ, as in Kâmviri čâl'a- 'migrate', ǰüa- 'be alive', and ǰâň'a- 'kill'.

A final innovation occurred throughout the Aryan-speaking region: the loss of palatality of the vowel e, merging it with a.

In the Iranian region, a further innovation, the loss of voiced aspiration, also spread to the proto-Nuristânis; but they escaped subsequent Iranian laxing innovations, such as the changes of s to h, of stops to spirants before other stops, of voiceless aspirates to spirants, and the deaffrication of the dental affricates.

By the beginning of the second millennium B.C. we find the Indo-Aryans spreading south over the Caucasus, perhaps under pressure from their Iranian cousins to the north. By the middle of that millennium they had mostly left their formative region. They mingled with Caucasian Hurrians to form the kingdom of Mitanni in the west, and they spread east across the northern Iranian Plateau, through Khorâsân and the Helmand basin and up into the Kabul River basin. At that time they began to collide with Iranians coming through Khorâsân from the north, and by the beginning of the first millennium B.C. the Iranians had prevailed and obliterated the Indo-Aryans from Khorâsân westward. The proto-Nuristânis escaped further Iranization by moving east from Khorâsân into Indo-Aryan territory.

At this point we begin to get traditional accounts of their origins. According to Kom tradition, their route took them through the Helmand Valley to Kandahâr, and then up to Kâbul, Kâpisâ, and finally down the Kâbul River to Kâmâ, at the confluence of the Kâbul and Kunar Rivers. They probably followed that route first to escape advancing Iranians, and later to escape advancing Muslim armies that began their conquest of the region at the start of the 8th century A.D. They remained in Indo-Aryan-speaking territory at least from early Middle Indo-Aryan times, and participated in many of the major linguistic developments of the northwest Middle-Indo-Aryan region. A large portion of the Nuristâni vocabulary may be traced back to Indo-Aryan borrowings.

At the beginning of the 11th century Afghan expansion forced the Nuristânis into their present region. Islamic Turkish and Afghan holy warriors from Ghazni swept through the Kâbul Valley, including Kâmâ, on their way to India. They tried to convert the Hindu "infidels" of the region to Islam; those who refused fled south to the Safed Koh and north to the Hindu Kush for refuge. The Nuristânis were driven part way up the Kunar and then north over the Kunar-Pech watershed, leapfrogging the Indo-Aryan speaking Degans who were settled in the lowlands of the Kunar basin. The Nuristâni refugees probably bypassed the Indo-Aryans of the lower Pech and Vaygal Valleys, to settle in more sparsely inhabited areas. The Kâta, Kṣto, and Mumo ended up in the region where the Ktivi and Pârun Rivers converge to form the Pech. One Kâta account claims that they displaced an indigenous people, the ǰâš'i, about whom more later. The Vâsi, being less powerful, ended up in the high and less desirable Pârun Valley, while the Âṣkuňu speakers established themselves above the Pech's bend and over into the Alingar basin. The Vay expelled the inhabitants of what is now the village of Vaygal at the head of the Vaygal Valley. Those new refugees went east to the upper Kunar, where they live today, speaking Gahwâr bâti. The Kom apparently split up; one group went as far as lower Řâmgal, while another group went to highlands further up the Kunar.

The invasion of the Ghaznavids was the first of a series of violent encroachments by Afghans into the lowlands of the Kunar, Pech, and Alingar valleys. This period of violent encroachment lasted into the 16th century, and displaced both Indo-Aryan and Nuristâni speakers. Encroachment by Afghans into the region continues in diminished intensity even until today; but with increasing intensity, contact with Afghans is transforming the lexicons of the Nuristâni languages through the introduction of the Perso-Arabic vocabulary of Islam.

During the millennium that the Nuristânis have lived in their remote valleys of the Hindu Kush, they have had chaotic relations with one another, and there has been much shifting and mixing of populations. One area of refuge was the middle and upper Lanḍai Sin Valley: an early split among the Kâta sent the speakers of Eastern Kâtaviri there, and the Kṣto, Binio, Mumo, and Jâmčo also sought refuge there. The Kom left Řâmgal and settled between the people of Saňu and Ktivi. There they became troublemakers and were driven out. They too fled to the middle Lanḍai Sin, where they encountered the Jâši and the Binio. They expropriated the Jâši lands in the lower Lanḍai Sin, down to the Kunar, where they apparently met up with their distant kinsmen living in the highlands across the Kunar. They also took most of the land that the Kṣto had settled.

Another refuge area was Tregâm. A group of Gřâmsaňâ people invaded the lower Vaygal Valley to become the Čima-Nišey; they expelled the Preǰvře˜ inhabitants into Tregâm. Another portion of the population of Tregâm were exiles from Ktivi.

By the time I started my fieldwork in Nuristan in the late sixties, the internecine fighting of earlier centuries had mostly disappeared, and any hostilities were aimed at Afghan and Guǰar interlopers. When that largest of interlopers, the Soviet Union, set up a communist regime in Afghanistan, the Nuristânis were the first to revolt. For almost twenty years now, the societies of Nuristân have been scarred by war, and I fear that the old internecine chaos may be on the rise. The languages, for the time being, remain vital; and their vocabularies have acquired a new Persian layer that allows them to interface well with the modern, external world of war and technology.

Let me turn now to a brief typology of these languages, which will be based on their cognitive geometry. I note at the outset that Vâsi vari is highly aberrant vis-à-vis the other languages, both in its phonological and morphological development.

Like most languages of ancient Eurasia, the Nuristâni languages first depict the geometric relationships of the objects of discourse, and then they depict a change in the objects. In "traditional" terms, they are SOV languages.

All objects in the cognitive field have a location, which is specified by components of distance and direction. Distance is measured by dichotomous divisions of space into internal vs. external spheres. Such divisions generate a three-way contrast of distance between archetypical, perceptual, and conceptual.

Nouns at archetypical distance represent internal archetypical objects, traditionally labelled as indefinite, and thus do not appear with deictics. An example is Kâmviri

         g'o gać. 'Give me a cow' (any cow),

with just the basic form of the noun.

Nouns at the remaining distances represent external objects, traditionally labelled definite, and thus require deictics to locate them in space, unless they are nouns with implied location, like nurist'on.

Nouns at perceptual distance represent objects internal to the speaker's current perceptual sphere. Such objects are again specified as internal or proximal vs. external or distal, as in Kâmviri

         ina g'oa gać. 'Give me this cow.'

vs.

         i·a g'oa gać. 'Give me that cow.'


Nouns at conceptual distance represent non-perceived objects, and are therefore external to the speaker's perceptual sphere; thus Kâmviri

          âska g'oa gać. 'Give me that (unseen) cow.'

With plural nouns conceptual distance is foreshortened, so that plurals appear only with perceptual deictics, whether unseen or not.

In addition to distance specification, Âshkuňu Vi:ri and Vâsi Vari allow directional pointers to be prefixed to deictics.

At least one object in the cognitive field stands in the foreground as the subject. Other non-archetypical objects may be backgrounded, as indicated by an oblique case suffix on the noun. Several postpositions and one or two prepositions may stand with an oblique-case noun to indicate regions or direction relative to the noun's object referent.

The overview of Kâmviri verbs that I presented earlier today [see abstract] is representative of the systems for the other languages, with some exceptions in Vâsi Vari. Finite verb forms are built on the ancient present active participle, the past passive participle, or, less ubiquitously, on the present stem, followed by a subject pointer. A participial base may be followed by a closely compounded form of *âsa- 'is' with subject pointer, or by an enclitic form of *âsa- with subject pointer. The forms of *âsa- serve to place the change further away on a distance scale analogous to the one for nouns. Here such distance is temporal, allowing discriminations between present vs. imperfect, or vivid vs. resultant past, or imaginative vs. resultative.

For transitive verbs built on the present stem (including the present participle), the agent stands in the foreground as subject, and an external patient is backgrounded by an oblique case marker. For transitive verbs built on the past passive participle, the situation is reversed; the focus is on the end-point of the verbal change, so that the patient stands in the foreground and the agent is backgrounded. This latter configuration implies that the speaker is looking back into time, with the final state of the verbal change being closer than the initial state. This typical "split-ergative" system is not found, however, in Vâsi Vari. There past forms are built on a participle that has an Iranian look, ending in -k, and definite patients are always backgrounded.

At least three modes are represented in verb stems, as were noted earlier today for Kâmviri; these include a real vs. an imperative vs. an imaginative mode.

Numerous non-finite forms exist besides the participles on which finite forms are built. There are conjunctive participles, which depict an agent-focus change prior to the change of the finite verb. There are infinitives, which depict a change as a temporal goal. Of interest are the forms that occur with verbs of motion. One of these depicts a change as a goal of motion, as in Kâmviri

          vâll'oaň gu·sa. 'He went to call on someone'.

The other depicts a change concurrent with motion; an example is Kâmviri

          kân'am 'oaso. 'He came laughing'.


A salient feature of these languages is the depiction of imbedded images, which are represented by subordinate finite clauses. They depict the internal states or compulsions of an oblique-case experiencer, as in Kâmviri

          'i˜a 'oata bo. 'I'm hungry'

or

          'i˜a i'e˜ sta_âsa. 'I have to go'.

With postverbal particles, they depict past or hypothetical scenes, as in

         'o˜ć g'um_to, t'ü â·k'i n'â_âsaš.
          'When I went, you weren't there'

or

        'o˜ć g'um bo, t'ü di ieloš. 'If I go, then you should go, too'.

With the postposed conjunctive participle (e.g., Kâmviri kti) of the verb 'make', in the sense of 'said', they depict quotations and causes, as in

          'i˜a i'e˜ sta_âsa kti giǰa_kâřo. 'He said that he has to go'

or

        gâtr'a bi·sam_kti n'â go. 'He didn't go because he was tired'.


Finally, I should mention these languages' unique system of directional pointers, which I will discuss in detail tomorrow morning [see abstract].

There is still much to be discovered about the languages of Nuristân. The pioneering research of Morgenstierne on these languages has been followed up with more concentrated field studies by only a handful of other scholars, including Georg Buddruss of Mainz University on Vâsi Vari, Aleksandr Griunberg of Leningrad and Jan Mohammad, a native speaker at the University of Arizona, on Western Kâtaviri, and myself and my mentor Qâzi Ghulâmullâh on Kâmviri. Unfortunately, the prospects for further research in the homelands of these languages are eclipsed by the hostilities still current in Afghanistan.