Are rules constructs of linguists or descriptions of actual associations latent in the speakers' mind? An interesting method of verifying the psychological reality of rules has been proposed by P. Kiparsky (1968). It can be summed up as follows. Given two chronological stages of a language (or two languages) where one can be said to be descended from the other, one first has to describe a phenomenon in the ancestor-language, then seek out and describe the phenomenon that can be considered a development thereof in the daughter-language. The two phenomena may be found somewhat different in certain respects, though still similar enough to be recognized as historically related. If the surface difference between the two phenomena, conceived as a change from the first to the second, can be explained, by comparing the formulation of the two descriptions, as a plausible modification of the first (set of) rule(s) naturally leading to the second, the chances are that the rules actually describe what appears in the speakers' mind, as part of collective knowledge. For Kiparsky, such a plausible change is primarily simplification.

The scholastic investigation of stress in Arabic starts with Lambert's pioneering article in 1897. Lambert demonstrated that the accepted stress of Classical Arabic is not based on genuine tradition, but was formulated at the beginning of the seventeenth century by Erpenius, who was influenced by the (Syrian–) Lebanese pronunciation of Classical Arabic. Recognizing the secondary origin of the traditional stress in Classical Arabic, Lambert established the relationship of stress in Classical Arabic to that of the various Arabic dialects on an entirely new basis. He himself emphasized the originality of stress as it is preserved in the Maghribī dialects (ibid., 409).3 In 1907 the first fascicle of Brockelmann's monumental Grundriss appeared, in which (pp. 82 ff.) Brockel-mann advocated the originality of the traditional accentuation in Classical Arabic, as generally preserved in the Syrian-Lebanese dialects. In 1908, Kampffmeyer reached the conclusion that the system of stress in the Maghribī dialects is more original than that of the Syrian-Lebanese group. However, since he published only the first part of his Untersuchungen, he dealt only with the stress in Spanish Arabic and the Moroccan (pp. 7–58), in addition to his results (pp. 1–7) contained in his brief introduction. A short synopsis of his views, with additional material, is included in Kampffmeyer, EI. An important review of the stress in Arabic (pp. 35–41) is contained in Sarauw. Despite the late appearance of his work (1939), he had almost completed it at the time of his sudden death in 1925. Moreover, the main body of his treatise had been composed as early as 1908 (cf. p. 3), the same year in which Kampffmeyer claimed the originality of the Western (Maghribī) stress.

Sībawaihi has occasionally been criticized for his obscurity and inconsistency. The former may be conceded, since it pertains only to his style and presentation, which undoubtedly make heavy demands on the reader, but the second criticism is an attack on Sībawaihi's linguistic reasoning and needs to be refuted. In the following pages we shall examine the background and use of the phrase ‘ishrūna dirhaman ‘20 dirhams’ as a locus probans in more than a score of grammatical arguments in the Kitᾱb, to which must be added a dozen or so other cases involving the use of similar expressions and grammatical principles. Two results should emerge: first we should gain the impression that the Kitᾱb is a work of the utmost coherence and consistency (with due allowance for the date of its composition), and second we shall uncover certain features of Slbawaihi's grammatical theory which appear to have been ignored by later grammarians.

In what follows I would like to discuss the structure of Amharic relative clauses. In the course of the discussion, I would like to make the following three claims which I will attempt to substantiate in turn. First, I believe that relativization is a kind of pronominalization and, consequently, the particle yä- that is attached to the main verb (or its auxiliary) of the relative clause is not a relative pronoun. Second, I maintain that the ‘yä- clause’ in subject position in Amharic cleft sentences is also a relative clause with an unspecified element as its head. My third claim is that Amharic genitive phrases originate from relative clauses and that the noun (phrase) in the genitive phrase to which the particle yä- is attached in surface structure is governed by a preposition in underlying structure, and the head of a genitive phrase is the head of the under-lying relative clause. In this connexion, I also argue that there is a rule in Amharic which moves the particle yä- (to the right) over, at least, one constituent.

The main purpose of this article is to show that the conjugations (binyanim) in Colloquial Israeli Hebrew, looked at from the point of view of their syntactic and semantic functions, form together a coherent, tightly organized system, in which each conjugation is characterized by a fixed configuration of semantic and syntactic properties. These properties are represented in terms of three concepts, defined and explained in §4: semantic value, semantic tie-up, and syntactic tie-up. There are many deviations from the regular pattern of the system, but relatively few of these are true exceptions. In the majority of cases either there is a rule which explains the apparent deviation, or the deviation should be considered unimportant in some definite sense.

It is now more than 90 years since Fortunatov first enunciated that rule of phonetic development in Sanskrit which has since borne his name. The article in which this theory was proposed appeared in A. Bezzenberger's Beiträge zur Kunde der indogermanischen Sprachen, VI, 1881, 215–20, and the rule formulated by him was that in the group l + dental in Sanskrit the l disappears and the dental is changed to lingual (cerebral, retroflex). This process of phonetic change is illustrated by the comparison of the Sanskrit word paṭa- ‘cloth’ with OSlav. platǐno, Russ. polotnó ‘linen cloth’, etc., where Sanskrit paṭa- has replaced a pre-Sanskrit *palta-. Fortunatov cited a large number of examples in support of this rule, and although many of these are unacceptable, there still remain enough good etymologies in his list to provide significant support for his theory. It is a noteworthy feature of this change that the cerebral replacing the combination of l with dental is a single consonant, in marked contrast to the later assimilation of r followed by dental in Middle Indo-Aryan where the result is always a double consonant (e.g. -ṭṭ- in vaṭṭai 〈 vartate, etc.). Another significant difference is that whereas in MIA the combination of r with dental results in dental as well as cerebral (vattai beside vaṭṭai, etc.), the product of l + dental in Sanskrit is invariably cerebral.

In an earlier article (Sprigg, 1963b) I drew on Burmese, with only a few examples from certain Tibetan dialects (Lhasa, Sherpa), to advocate applying prosodic analysis, the name commonly given to the polysystemic type of phonological analysis devised by J. R. Firth, to the languages of the TibetoBurman group for purposes of comparison and reconstruction; since then I have had an opportunity of studying two typologically different Tibetan dialects, the non-tonal Golok and the slightly tonal Balti, both of them remarkable for syllable-initial consonant clusters; and the relations of tone in the tonal Tibetan dialects to syllable-initial consonant clusters in the non-tonal dialects much strengthens the case, to my mind, for basing comparison of Tibetan dialects, and. through them, Proto-Tibetan reconstruction, on polysystemic analysis. The two main characteristics of such an analysis would be: (i) separate phonological systems for different types of syllable feature and syllable-initial feature; (ii) the emphasis on the syntagmatic association of successive phonetic features of the utterance rather than on purely paradigmatic contrast.

When revolution broke out in China in 1911 and rapidly led to the overthrow of the dynasty and the founding of a republic, opinion throughout the world was startled and deeply impressed. The European press expressed amazement that a quaint Oriental people such as the Chinese should take it into their heads to carry out an anti-monarchical revolution, and that hard-headed political realist Vladimir Ilich Lenin waxed positively lyrical about the Chinese revolution and its leader. ‘The East’, he wrote, ‘has committed itself to the Western path,…further hundreds and hundreds of millions of people will from now on join in the struggle for the ideals for which the West has striven’. In Sun Yat-sen, he saw ‘a revolutionary democrat, imbued with that nobility and heroism peculiar to a rising rather than a declining class’. Even when Sun had been supplanted as President by Yüan Shih-k'ai, Lenin still saw a ‘mighty democratic movement’ in progress throughout Asia, whereby hundreds of millions of people were ‘awakening into life, light, and freedom’. Succeeding generations of scholars and political observers, instructed by the spectacle of corruption, division, and military domination which unfolded itself in various guises throughout the era from 1911 to 1949, have gone almost completely to the opposite extreme. What happened in 1911 was not a revolution at all, or only in the most superficial sense; the monarchical system was indeed overthrown, but there was no real change in the locus of power in society.

The chronology of Darius' Behistun inscription (= DB) has been the subject of detailed examination especially since 1938 when A. Poebel established the order of the Old Persian months. One problem which has received particular attention is the apparent lack of agreement between the statement of Darius that he gained 19 victories in one year after he became king, and the fact that the time between the first and the last dated events mentioned by him exceeds one year. Furthermore, the recorded battles number 18 instead of 19. The purpose of this note is to study Darius' claim once more, to investigate the previous attempts at tackling the question, and to offer a new explanation for the discrepancies just noted. We start with the order in which the victories of Darius are recorded, for which the following list may be drawn up:

1. Pseudo-Smerdis put to death by Darius in VII, 10 (DB, I, 56).

2. Assina, king of Elam, captured and slain, date not given (DB, I, 81–3).

3. Nidintu-Bel defeated on the Tigris in IX, 26 (DB, I, 88–9).

4. Nidintu-Bel's second defeat in X, 2 (DB, I, 96).

5. Martiya rebelled in Elam, but the Elamites seized and slew him while Darius was in Babylonia (DB, II, 8–13).

6. Fravartiš's army suffered defeat in X, 27 (DB. II, 26).

7. First victory of Dādaršiš over the Armenians in II, 8 (DB, II, 36).

8. Second victory of Dādaršiš over the Armenians in II, 18 (DB, II, 41).

9. Third victory of Dādaršiš over the Armenians in III, 9 (DB, II, 47).

10. Vahumissa overcame the Armenians in X, 15 (DB, II, 56).

11. Second defeat of the Armenians by Vahumissa in II, 30 (DB, II, 61–2).

12. Darius won a decisive victory over Fravartis in I, 25 (DB, II, 69).

13. The Sagartian Čissantaxma captured and slain, date not recorded (DB, II, 78–91).

14. Hystaspes won a victory over the Parthians in XII, 18 (DB, II, 98).

15. Hystaspes routed the Parthians in Patigrabana in IV. 1 (DB, III, 8).

16. Dādaršiš, satrap of Bactria, quelled the revolt of the Margian Frada in IX, 23 (DB, III, 18).

17. Vahyazdata's first defeat in II, 12 (DB, III, 39).

18. Vahyazdata's second defeat in IV, 5 (DB, III, 46–7).

19. Vivana's first victory over the army of Vahyazdata in Arachosia in X, 13 (DB, III, 63).

20. Vivana routed the supporters of Vahyazdata in XII, 7 (DB, III, 68).

21. Araxa rebelled in Babylonia, but was captured and slain in VIII, 22 (DB, III, 88).

The preposition rm ‘with’ is common in Buddhist Sogdian texts, but has not been found in texts in Manichaean or Christian writing; nor is such a preposition known in any other Iranian language. Its peculiar distribution was noticed by Benveniste, who also remarked: ‘rm…se confond dans l'usage avcc δ'wn, δnn, et entraîne le cas oblique.… Il admet, comme δnn, la postposition pr'yw‘. These facts in themselves strongly suggest that rm is not merely a synonym, but a purely graphic variant, of δnn; and this is confirmed by the ease with which rm can be explained as an ideogram.

Source material for study of the evolution in practice of the forms of music in Indian classical tradition are meagre, although prescriptions for the various types of prabandha abound in the treatises on music. The reason is not far to seek. An essential feature of Indian art music has always been that it is acquired orally from a teacher and transmitted in turn in like manner. It is true that a non-diastematic notation, varying somewhat from one area of the country to another, has been observed, but its purpose traditionally was to present exercises and to facilitate the committing to memory of certain passages. To the present day, no musician would perform from a ‘score’ although, with the advent of radio and the concomitant interest in ‘orchestras’, it should be noted that the music performed is frequently written down beforehand and distributed to the performers.