¹ The Rev. Bailey, T. Grahame, Grammar of the Shina (Ṣiṇā) language (London, 1924), 280Google Scholar. I have retained Bailey's transcription.

² Shaw, R. B., ‘Stray Arians in Tibet’, JASB, 47, 1878, 56Google Scholar. The transcription is Shaw's.

³ The data presented in this paper were collected over a four-day period spent in Skardu in the summer of 1987, during the first phase of a survey of the languages of the Indus Valley undertaken by me and my student, Dr. Elena, Bashir. That survey comprises the linguistic component of a multidisciplinary study of Pakistani culture, organized by DrWilma, Heston under the joint direction of Professor William Hanaway of the University of Pennsylvania and DrUxi, Mufti, Director of Lok Virsa, Islamabad. I gratefully acknowledge the financial and administrative support of the Smithsonian Institution, the guidance of DrAdam, Nayyar of Lok Virsa, and the assistance of Mr. Baig, G. M. (Gilgit) and Mr. Qazmi, S. M. Abbas (Skardu) in making local arrangements and obtaining introductionsGoogle Scholar.

⁴ Skardu is about 80 kilometres to the north of Dras. Our informants there were Mr. Ghulam, Mahdi, about 30 years of age, an employee of the Forest Department stationed in Skardu and a native of Satpara (known locally as sāspar), about 16 kilometres south by south-west of Skardu, and Gholam Sahab, 55 years of age, retired after 19 years military service to his native village Tandal, about 14 kilometres west of Skardu. Like many settlements of Shina speakers in Baltistan, Satpara (S) and Tandal (T) are located on the flanks and high valleys of the Deosai Mountains on the left bank of the Indus. Skardu itself is Balti-speaking and our informants, like nearly all the Shins (in Balti known as Brokpas) living in Baltistan, were bilingual in Balti and Shina (Brokskat). For elicitation we used Urdu, the chief language of literacy in the areaGoogle Scholar.

⁵ The system used here for the transcription of Shina data from areas other than Gilgit is that used in Turner's, R. L.Comparative dictionary of the Indo-Aryan languages. The sound represented by ‘j’ seems to vary in Skardu dialect between [j] and [ž]Google Scholar.

⁶ Abbreviations used in this paper include:

1, 2, 3 first, second, third person

CP affix of conjunctive participle

DT dative case or postposition

ER ergatiye case or postposition

f. m. feminine, masculine gender

G Shina of Gilgit

GN genitive case or postposition

IN instrumental case or postposition

NM nominative case

P past tense affix

PR present tense affix

QM question marker

s. p. singular, plural number

S from village Satpara (Skardu)

T from village Tandal (Skardu)

⁷ The Shina of Skardu agrees with Drasi in having the dative-of-agreement with lěj-. Although we cannot conclude anything from their silence, neither Shaw nor Bailey mentions the construction as characteristics of other predicates in Drasi.

⁸ ‘Dative (of) subject’ is the term commonly found in studies such as Colin Masica's, P.Defining a linguistic area: South Asia (Chicago, 1976)Google Scholar. My ‘Coexistent analyses and participant roles in Indo- Aryan’, Indian Linguistics 46,3, 1985, provides a detailed discussion of the syntactic and conceptual bases for modifying the terminology.

⁹ The standard construction, of course, is:

(a) mìĩ tumhārī bāt bhūl gayā

I-NM your word forget went

‘I forgot what you said.’

with nominative case for (and verbal agreement with) the noun phrase that expresses the person who forgets.

¹⁰ The figures are 21 datives to 10 ergatives for speaker (T) versus 19 ergatives to 3 datives for speaker (S) (in their use of paš). This difference in the preferences of the two speakers accords with a local tradition of a Chilasi origin for the Shins of Satpara. It is possible that their speech (still a form of Chilasi? Emerson, R. M., ‘Charismatic kingship: a study of state-formation and authority in Baltistan’, Journal of Central Asia, 7, 1984, p. 130, n. 16)Google Scholar has been gradually but not completely influenced by a dialect such as Tandal's spoken by Shins indigenous to the area.

¹¹ I was able to inform myself on the relevant syntactic points from speakers of Gilgit and Gurez Shina.

¹² See Bailey, , Grammar of the Shina, op. cit.;Google Scholar Col. Lorimer, David L. R., ‘The forms and nature of the transitive verb in Shina (Gilgit dialect)’, BSOS, 3, 1924, 467–93;CrossRefGoogle ScholarRuth, L. Schmidt, ‘Morphological criteria for distinguishing categories of transitivity in Shina’, in Arlene Zide, R. K., David, Magier and Eric, Schiller (ed.), Proceedings of the Conference on Participant Roles: South Asia and adjacent areas (Bloomington, Indiana, 1985), 33–47Google Scholar.

¹³ Gilgit Shina data are from Mr. Muhammad, Amin Zia in a transcription that he developed with George Buddruss. It recognizes the distinction between vowels accented on the first as opposed to the second mora, thus requiring the use of a digraph to indicate length. See Zia's šinā qāidā aur grāmar, Gilgit, 1986, for discussionGoogle Scholar.

¹⁴ If the š in skomušemus is a misheard ṣ then the form may be analysable as skom(u) ṣ-emus where ṣ-emus < ṣono ‘to attach’ (cf. Urdu lagānā)Google Scholar.

¹⁵ Shina, like other languages of Pakistan's far north (Burushaski, Balti, Dumaki) is ergative in all tenses. Verb agreement, however, is with the subject, not the object. Compare Kashmiri, Pashto, Gujuri, Hindko, Urdu, and other more southerly languages in which the verb agrees with the object but only in certain past tenses.

¹⁶ Consequences of this for the supposed unidirectionality of certain types of syntactic change are discussed in my and Usha, Lakshmanan's ‘Not every diachronic path is a one-way street: the emergence of inversion in South Asian languages’ in Proceedings of the Conference on the Theory and Practice of Historical Linguistics, Need, B. and Schiller, E. (ed.), (Chicago, 1989)Google Scholar.