Qur'ān 2:114 describes those who prevent God's name being uttered in His mosques (man mana'a masājida ‘llāhi an yukhdara fīā’smuhu) as ‘most unjust’ (aẓlamu). If further states that ‘they shall not be allowed to enter them except in fear; they shall endure the curse of chastisement in this world and great torture in the aftermath ’.

The distinguished orientalist Reinhart Dozy, writing in 1849, and referring to the possible influence of Arabic poetry on Spanish romances and on the early troubadour lyrics, pronounced the whole question as ‘tout à fait oiseuse’. He went on to say about this highly controversial question: ‘nous voudrions ne plus la voir débattue, quoique nous soyons convaincu qu'elle le sera pendant longtemps encore. A chacun son cheval de bataille’.

Despite the similar titles of the two books reviewed here, they are very different in aims and scope. While Samadi's book is primarily an etymological dictionary of verb forms, Kellens's is in effect the first volume of a comprehensive Avestan grammar, as a companion to which the author promises ‘ in the near future' (p. 435) a complete lexicon of the Avestan verb. What they do have in common, in addition to the relationship in subject-matter indicated by their titles, is a comprehensiveness which ensures that they will both become standard works of reference. It is this characteristic above all which seems to justify reviewing them in more than usual detail.

The following study contains a translation of al-Bīrūnī’s rendering into Arabic of the fourth and last chapter of Patañjali’s Yogasūtra cum commentary. Our translation of the three preceding chapters was published in BSOAS, xxix, 2, 1966, 302–25 (henceforth abbreviated as BSOAS ch. I); BSOAS, XL, 3, 1977, 522–49 (henceforth abbreviated as BSOAS, ch. II); BSOAS, XLVI, 2, 1983, 258–304 (henceforth abbreviated as BSOAS, ch. III). This translation is based on Ritter’s edition of the Arabic text. Comparison has been made with the unique MS of Ritters Text: Kopriilii, 1589, fols. 412a–419a (written on the margins). We have also compared the text with parallel passages and expressions in al-Birunl's India.

When in 1857 Ferrier reported a rock relief, apparently Sasanian, in the region of Sar-i Pul in North Afghanistan, he left us with an enigma which is still unsolved and which has for over a century remained a subject of intense discussion for specialists. The intrepid scholars who searched around the remote and often inaccessible rocky mountains of the area found no trace of the rock relief, but A. D. H. Bivar, while searching for the relief in 1964, was the first to notice some important and yet previously little known Islamic shrines near the village of Sar-i Pul. The two most interesting monuments were the shrines (ziyārat) of Imām-i Khurd and Imām-i Kalān. The first, a small squaredomed chamber, with several inscriptions, one commemorating the martyrdom ofYahya b. Zayā (d. 125/742–3), was described in detail by Professor Bivar who suggested a mid eleventh century date for the building, and it is not discussed again here. The other shrine was briefly mentioned, and two photographs showing details of the cut stucco work of the building were given.

My purpose in this paper is to outline a preliminary reconstruction scheme for the agreement morphology of the Proto-Tibeto-Burman (PTB) verb. I will deal with two separable issues: the provenence of the widespread suffixal agreement paradigm, and that of a less widely attested set of prefixes. The suffixal paradigm has, since the work of Bauman (1974, 1975, 1979, see also DeLancey 1980, 1981b) been recognized by some, but by no means all, Tibeto- Burmanists as being reconstructible for PTB. The prefixes, on the other hand, are as far as I know currently universally viewed as of secondary origin (although Bauman (1975) attributes a PTB provenence to one of the morphemes involved, he does not consider it to have been part of an original agreement system). I will show, however, that there is good evidence on which to base an argument that these also are of PTB provenence, and in fact very likely older than the suffixes.