In a recent article, I argued that the study of early Arabic grammar of the second and third centuries A.H. should be based on the sources of that era, and not on the work of the later grammarians, as these often do not report the real stand of a particular grammarian on a particular issue, and more often misunderstand certain early grammatical concepts, or use their own terminology rather than the older one to elucidate them. The problem becomes more acute when we find that many modern scholars can still only see early Arabic grammar through the eyes of the later grammarians, beginning with the fourth-century figures and probably ending with Suyṭ. Learning about early grammar from later sources is inescapable, given the scarcity of existing second- and third-century sources, but this is nevertheless permissible only if we do not impose the method, and especially the terminology, of the later authors on early grammar.

It was remarked by the Czech linguist Pavle Ivi that it would be possible for particular linguistic landscapes to fit easily into particular theories about linguistic diffusion. He was referring especially to the difference between the Stammbaum or family tree theory and the Wellentheorie or wave theory as expounded by the Neo-grammarians. The family tree theory sees dialect features as genetically transmitted and traceable to the homeland of the group concerned. The wave theory on the other and concentrates more on the present geographical location of the speakers and traces the geographical diffusion of dialect features across the landscape independently of the movement of groups of people. If a distinction between family tree landscapes and wave theory landscapes is possible, as Ivi suggests, then the dialectology of the bedouin tribes of north-eastern Arabia falls quite definitely into the first group. In a previous article the writer described the dialect of the Muṭair tribe of Eastern Arabia which a dialect of the Central Najd type akin to the dialects of Central Najd and the 'Aniza tribes of the Syrian Desert. The Ḍhafr, although direct neighbours of the Muṭair to the north and west, speak a dialect of a quite different kind, basically North Najd, akin to the speech of the Jabel Shammar and the Shammar tribes, but with some admixture of Central Najd features. It seems that in this area dialect is a marker of tribal identity and since there is no gradation between membership of the Ḍhafr tribe and membership of the Muṭair, the dialects are similarly clearly demarcated geographically. The boundary between the Ḍhafr and the Muṭair also Muṭair also constitutes a more far-reaching boundary between Ahl al-Shiml and Ahl al-Jinb, northern and southern bedouin. This division is marked by dialect features and also certain other cultural externals such as dress and tent structure.

Yska, in his Nirukta, divides words into four classes: nman (noun), khyta (verb), upasarga (preposition) and nipta (particle). He is the only one of all the ancient writers on grammar who deals at some length with the last mentioned category of particles. The introduction to this class runs as follows:

atha nipt uccvaceṣv artheṣu nipatanti. apy upamrthe. api karmopasaṃgrahrthe. api padapraṇḥ

The Ming government's grave concern with the Mongols appears on every written page of the history of that era. Moreover, this obsession with the danger threatening from the north is reflected in a great number of place names. Towers, fortresses, garrison towns of every kind and description built and repeatedly repaired along the border, carried names referring directly or indirectly to the Mongols. Whenever a new construction was begun, it had to be given a name, and every often a name indicating a role in the defence of the country: in fact, these names call to mind such Russian place names as Vladivostok Domination of the East, and Vladikavkaz Domination of the Caucasus.

Malay is one of the best-studied of all Austronesian languages, both synchronically and diachronically. We would not therefore expect a study of Malay historical phonology to uncover much that is new. In this paper, however, I will suggest that a basic change has been overlooked in past work on the language.

Since the publication of the catalogue of manuscripts in Indonesian languages in British public collections (Ricklefs and Voorhoeve, 1977), further MSS have of course come to light, as have some of those errors which inevitably attend such a work. It seems appropriate to publish here a list of additions and substantive corrections known to the authors by the beginning of 1981. The authors of the catalogue are grateful to those librarians and scholars who have brought such things to their notice, and particularly to Dr. A. C. Milner for offering his descriptions of Malay MSS found in the Wellcome Institute for publication here.

There are several problems intimately connected with pronominal affixes in Chadic. These problems may be connected with each other and an explanation of one may possibly shed some light on other problems. The following problems will be discussed in the present paper:

(a) In several languages there exist pre-pronominal markers whose function is poorly understood. I will illustrate this problem in the example of Hausa, because it is the most familiar Chadic language and also because it is fairly typical for the situation under consideration.

Writing in 1941, in the foreword to his Khotanese texts I (Cambridge, 1945), H. W. Bailey related concerning the Jvaka-pustaka: The late Dr A. F. Rudolf Hoernle had begun a study of part of this text. Some six years ago I saw the MS on one afternoon, but when about two years later I wished to examine it, it was not to be found. I have not seen it again and do not know what Dr Hoernle had done.