There are two specific reasons for publishing these documents—quite apart from their intrinsic value—in an article in honour of H. J. Polotsky. My teacher H.J.P. and myself were, at different times, in statu pupillari, at any rate informally, in relation to Eugen Mittwoch (1876–1942). Polotsky took part in a Syriac class and in 1933–4 collaborated in the work of editing the Arŝam documents (private letter from H.J.P. to E.U. dated 25 December 1977), while I attended some informal lectures on Arabic given by Mittwoch in the mid-1930s. And, secondly, Polotsky has declared that any specimen of Aläqa Tayyä's handwriting is ‘always extremely pleasing’ (see below and BSOAS, XXXV, 2, 1972, 255 ff.).

Early in the sixth century A.D. the Greek author known generally as Cosmas Indicopleustes records in his ‘Christian Topography’ that he visited Adulis on the western Red Sea coast and there saw two Greek inscriptions, one on a marble throne and one on a stela standing behind it. The monuments have long ago disappeared, but we have the texts as copied by Cosmas. The stela belongs to Ptolemy III Euergetes (246–222 B.C.), but the throne text begins in mediis rebus without name of author, and is introduced by Cosmas with the slightly mysterious words ‘Then, as if sequentially, there is further written on the throne as follows’. Whatever he may have meant by this, it is certain, as modern scholars recognize, that the throne text is not part of the stela one: the former is drafted in the first person, the latter in the third; and the throne text was drafted ‘in the 27th year of my reign’, whereas Ptolemy HI died at the beginning of his twenty-sixth regnal year. This leaves the field open to speculations about the authorship and date of the throne text.

The name salūqī was that given by the Arabs, from, at the latest, the beginning of Islam, to the medium-sized gazehound with long pendulous ears which had hunted in the area of the Near and Middle East for perhaps thousands of years. As far as we are aware, the earliest occurrence of the word in Arabic literature is in a poem of Yazīd b. Dirār al-Muzarrid, who was born before Islam and who talks of banāt salūqiyyayn, ‘the offspring of two salūqīs’. Although the actual term salūqī does not, to my knowledge, occur in the classical hunting songs (ṭardiyyāt) of the Abbasid era, it is a word in general use in animal literature and that of the chase right through the medieval and late medieval periods. The word has survived to this day in the areas of the Arab World where the hound is still to be found, notably in the Hijaz and Najd areas of Saudi Arabia and the Gulf states.

In an earlier issue of BSOAS I published some notes on an interesting feature of the Socotri language, namely that certain imperfect forms lack the t-prefix which marks 2nd and certain 3rd person forms in Semitic languages.

The J(ibbāli), or Śherí, language of Dhofar also exhibits this negative feature, and in much the same series of forms; principally the indicative, subjunctive, and conditional forms of causative verbs, intensive-conative verbs, quadriliteral verbs and passive verbs.

The letters of the Byzantine emperor Manuel II Palaeologos, studied and exploited by Byzantinists since they were published for the first time by E. Legrand in 1893, are now available in a new edition by G. Dennis, in which they are translated, accompanied by the appropriate comments and, in most cases, accurately dated. These letters constitute a most valuable contemporary source for the years 1382–1417 because they help to fill one part of the much deplored ‘gap’ in Byzantine historiography. Moreover, they include evidence which sheds light on some details of early Ottoman history, as will appear from the following analysis of Manuel's material dealing with the conflict between the sultan Bāyezīd I and the ruler of Sivas, ḲāḲi Burhān al-Dīn Aḥmad in 1391.

The Indo-Iranian linguistic frontier constitutes one of the most complex and interesting language-areas of the sub-continent. Given the nature of the area, it is perhaps inevitable that scholarly attention should have been directed particularly to its remoter corners, where so much that is of historical importance has been preserved, and we certainly have every reason to be grateful for the fascination which such out of the way survivals have held for the minds of several outstanding linguists. It is, on the other hand, a matter for regret that so little has been done by comparison on the languages which flourish in less inaccessible parts of the frontier, particularly on the Indo-Aryan side. The wide distribution of such languages alone, quite apart from their intrinsic interest, demands that they too be accorded adequate coverage if the peculiarly complex language-patterns of the area are ever to be properly understood as a whole. The present article, based largely on material collected during a recent field-trip to Pakistan,1 represents an attempt to fill one such gap in contemporary coverage, by providing descriptions of the extreme north-western extensions of the main body of Indo-Aryan.

In an earlier study, ‘Assimilation, and the definite nominal particle in Balti Tibetan’ (Sprigg, 1972), I dealt with vowel harmony in words in which the noun is colligated with the definite nominal particle, a particle that has the constant phonetic form po when in junction with the final syllable of a noun ending in a consonant (-Cpo), e.g. smenpo ‘the medicine’ sman, but variable phonetic features where the final syllable of the noun ends in a vowel: a share in the features of the resulting word-final long vowel, either -o: or -u: according to vowel harmony.

The colophons of the Mongol Ganjur contain a large number of terms and expressions having to do with ‘writing’, ‘translation’, ‘revising’, ‘correcting’, ‘copying’, etc.; but, with these expressions appearing in almost numberless varieties of combinations with each other and with other words, it is extremely difficult, if not impossible, to assign a precise meaning to each particular one as distinct from the others. Some of these terms are synonyms and are used either together, or as substitutes one for the other, duplicating and repeating each other. The taste of the writers may very well have dictated the choice of one term in one place, and a different one, with roughly the same meaning, in another passage. The various ways in which these terms are combined and strung together, themselves strongly suggest that the Mongol translators, or the authors of the colophons, maintained no sharp distinction between the various terms: where is the precise line between ‘correcting’ and ‘revising’, ‘editing’, and ‘compiling’? Nevertheless, it should be possible to classify those expressions or terms into a few categories, even if there exists no clear line of demarcation between the various notions shading off into each other. I do not intend to trace such neat lines of separation, but I do believe that it is possible to define, if only in broad terms, the general meaning of the expressions concerned. And if carefully analysed and compared, these very words may tell us something about the manner in which the translation of the Buddhist sūtras was achieved.

It is a commonplace in the study of pre-Han texts to acknowledge that the received version of a given text cannot be assumed to reflect with any significant degree of fidelity its original form. A text is, as the literal sense of the English word implies, something woven, something stitched together (cf. Skt. sūtra, Ch. ching), and once woven, it may ravel. It would then be subject to re-weaving, either in its entirety alone, or together with other, originally distinct texts. In the former case individual phrases or lines might be introduced throughout the piece for reasons for euphony, stylistic balance, or perhaps because of a misunderstanding of the original sense on the part of a later scribe. In the latter case wholly independent accounts might become interwoven, begetting a new, hybrid document. Such re-weaving with its gratuitous additions of new material might occur several times, further distorting the primary content on each occurrence. The end result of such a process of textual alterations would be a composite and thoroughly heterogeneous work of diverse provenances, and of uncertain internal uniformity.

Most collections of Japanese wood-block printed books contain some items that were once in the possession of the circulating libraries known as kashihonya and the library of the School is no exception. Kashihonya were certainly in existence by the Genroku period (1688–1704), for the collection Haikai sugatanazo, which was published in 1703, includes the verse, Karihon no kakidashi ga kuru toshizakai ‘The year's end—when the bills come in for books borrowed’. Thenceforward they thrived on the growing literate population and in the Tokugawa period they reached their peak during the first half of the nineteenth century. It is from this time that almost all the former kashihonya books in the School date. The books contain the usual range of seals and admonitory labels attesting to their use by kashihonya, but I list only those with sufficient information to be of interest.

Studies of political and administrative institutions in pre-colonial Burma have almost invariably adopted a synchronic perspective. In those inquiries limited to two or three generations, the effort to present a composite picture entails perhaps few dangers. But in so far as these investigations embrace very long periods, or in so far as studies of a particular era have attempted to project their findings on to the pre-modern period as a whole, they present an artificially static image of indigenous civilization.

I.o. The phenomenon of reciprocity has, within the framework of generative grammar, been dealt with either as an aspect of sentence coordination as in the following:

1(a) the boy saw the girl

(b) the girl saw the boy

(c) the boy and the girl saw each other

A recent visit to Iraq (made possible by a fellowship grant from the British School on the Gertrude Bell fund and by the kind assistance of the Directorate General of Antiquities in Iraq) enabled the writer to look closely at a number of manuscript collections. What follows is an indication of these resources and a review of a new catalogue.

A few uears ago, Roger Cowley edited an interesting text of the Tamhartä haymanot in older Amharic with an English translation under the title given above. The text is of special value to te study of the history of Amharic: it is relatively lengthy (14 fols., 2 cols., 12 lines to a column and 6–8 characters to a line), in prose, and shows most of the features of archaic Amharic known so far. In spite of the fact that, to use the author's own words.