Reviews Of Different Books

Catalog link: https://jainqq.org/explore/269335/1

Page #1
REVIEWS K. Hoffmann, Der Injunktiv im Veda. Eine synchronische Untersuchung. Heidelberg. 1967. Carl Winter. Universitatsverlag. (Indogermanische Bibliothek, Dritte Reihe, Untersuchungen). 298 pages. K. Hoffmann has chosen as an epigraph to his investigation the following words of Delbruck: "Wer die Injunctive des RV durchsieht, erhalt den Eindruck, als habe er eine Sammlung der Schwierigkeiten des Veda vor sich." These words are relevant to the present state of the studies of the Injunctive as well. Taking notice of the fact, that all his predecessors treated the Inj. mainly from the glottogonic or from the comparative viewpoint, the author sets himself as an object a purely synchronic aim - to describe the position of the Inj. as of a grammatical category functioning in the Vedic verbal system. Such an aim can be reached only by means of a philological investigation of the text. Realizing all the difficulties of such an approach, Hoffmann sees the only reliable way to find out, what the pisis had meant, when they used a certain grammatical category, the Inj. in particular, in an attempt to establish, whether the same contents could be expressed by means of other grammatical categories or not. In other words, the transformational analysis is taken here for the only safe criterion. These general methodological considerations are set forth by the author in a short foreword (Vorwort, 7-8). It is followed by a detailed table of contents of the book, which has rather acquired the shape of brief annotations of the chapters in German (Inhaltsubersicht, 9-17) and then in English (Summary of Contents, 18-26) and makes the book more accessible. Chapter I "Introductory" (Einleitung, 27-42) contains the definition of the object and method of investigation, and gives a short account of the history of the problem. It is a theoretical chapter, upon which the whole investigation is based. The author is quite right in stressing, that the Inj. should be identified only by means of a formal criterion (it is a category characterized by the absence of augment and by the secondary endings). It should be treated as a whole grammatical category, not to be divided according to the meaning of forms into the indicative and modal Inj., as it has been done by the representatives of the German tradition (Brugmann, Delbruck, Wackernagel). Stating, that the Inj. can be formed in the Vedas from three verbal stems - the Present (krnavam-cp. Impf. akrnavam, the Aorist (karam-cp. Aor. akaram), and the Perfect one (didet -- cp. Plusq. adidet), Hoffmann unfortunately does not mention the fact, that Inj. pr. and Inj. aor. are regular formations, while Inj. pf. are a few forms with an ambiguous morphological structure. As it has been shown by P. Thieme in his investigation of the Vedic Pluperfect, there exist no formal criteria able to identify each given form as belonging to the system of Perfect, to the reduplicative Aorist or to the Intensives. 1 See Das Plusquamperfektum im Veda (Gottingen, 1929).

Page #2
248 REVIEWS Two different meanings of the Inj. have been usually stated in the previous studies : a non-modal and a modal one. The sphere of the non-modal meanings has been described either as indefinite present and indefinite past (Avery), or as indiscernible in certain contexts from that of Impf. indic. and Aor. indic., and capable in later hymns of playing the role of a Pr. indic. (Delbruck). The sphere of the modal meanings of the Inj. has been described as that of Subjunctive-Optative (Avery) or of Subjunctive-OptativeImperative (Delbruck). The authors of modern studies of the Inj. stressed such a particular characteristic of this grammatical category as the absence of meaning of a certain tense or mood (Renou, Gonda), and marked its archaic character in the language of the Rgveda. The rapid disappearance of the Inj. in the post-Rgvedic language in all the functions, except in prohibitions with the particle ma, is explained by the process of development from the "primitive" mental structure (Gonda). Hoffmann quite soundly rejects the possibility of explaining the function of the Inj. in the complicated and ramified verbal system of the RV from the diachronic viewpoint. One can describe the volume of the grammatical category of the Inj. in the synchronic system only by means of its oppositions to other tenses and moods functioning in the same system. It is only the language of the RV, where the Inj. is represented in all its functions, that can be used for such a synchronic study. In his definitions of the function and of other general notions the author follows the views of E. Koschmieder. He considers it relevant for a linguistic description to apply the notion of noeme -- the smallest unit of what can be meant (p. 37). According to the author, it is necessary for a linguistic description to establish the degree of relevance of separate noemes in a noematic category, i.e. in a function (p. 38). Leaving aside the philosophic grounds of this book, one should mention an important view of Hoffman, that only philological methods should be applied to establish the functions of a category in a synchronic study. One can prove, that he has established the main function of a category, only by the impossibility of its interchange with the other categories in the same context (p. 39). Nevertheless the author is aware of the fact, that there is no possibility of such a control in many contexts of the RV, that is in such cases, when several noematic categories can be taken as relevant for one and the same form. In chapter II the Inj. in prohibitive sentences is dealt with (Der Injunktiv im Prohibitivsatz, 43-106). Here and further on the meanings of the verbal forms are established with the help of a thorough philological analysis combined, where it is necessary, with the interpretation of a wide context, based on a deep knowledge of Vedic realities. Two types of prohibitions are distinguished by the author, i.e. the Inhibitive, which is a prohibition of an action, taking place at the moment of prohibition, and the Preventive, which is a prohibition of a future action. Further the prohibitive sentences are classified according to the type of stem, from which the Inj. used with the particle ma are formed. The Inj. with aoristic stems are the most frequent in prohibitive sentences (IIA. Der Aoristinjunktiv im Prohibitivsatz, 45-73). They are represented in the Vedas by the forms of all the three persons in the Sg., Du. and Pl., and are based on all the types of aoristic stems. The 1st person of the Inj. aor. with ma has a cupitive meaning, and in 2 The author is well acquainted with a wide range of scientific publications, dealing with the problem of the Inj. from different viewpoints, including the last publications on the subject by I. Thomas-Freiberger, Die Injunktiv-Konjunktiv-Formen im Rgveda, Diss. Gottingen, 1955 and J. Narten, Die sigmatischen Aoriste im Veda (1964). The author seems not to be acquainted with the book written in Russian by the author of these lines T. 8. EnugapeHKOBA, Aopucm e Puzbede (Mockba. 1960) 3 See Beitrage zur allgemeinen Syntax (Heidelberg, 1965).

Page #3
REVIEWS 249 this function can interchange with the 1st person Optative with the negative particle na. The author draws the conclusion, that the cupitive meaning of this form is characteristic only of the later Vedic hymns. The original meaning of ma in combination with the 1st person of the Inj. was that of appeal. The forms of the 2nd and 3rd persons, have a preventive meaning, e.g. RVX, 59, 4: mo su nah soma mrtyave para dah 'O Soma, do not give us away to death!' The inhibitive meaning is rare with the forms of the Inj. aor. This interpretation is convincing in general. The only thing that can give rise to objections is the desire of the author to qualify each form of the Inj. as belonging either to the Aorist or to the Present system. As a result of it a number of ambiguous forms, which should be attributed rather to the Present-Aorist system as a whole (or in other words to the non-Perfect system," are qualified as belonging to the Aorist. If one accepts the existence of such forms, which are functioning in the RV as a relic of a more ancient state together with the differentiated and mutually opposed forms of the Present-system and of the Aorist, there would be no need to prove, that thematic forms like karat etc. are at times Subjunctives of a root-Present, while at others they are Inj.-s of a thematic Aorist. The next section of this book deals with the functions of the present-stem Inj. in prohibitive sentences (II B. Der Prasensinjunctiv im Prohibitivsatz, 74-92). It is shown by Hoffmann, that the Inj. pr. combined with the particle ma has an inhibitive meaning, i.e. expresses the prohibition either of a durative or of an iterative action. This meaning of the Inj. is illustrated mostly by examples taken from the Vedic prose, where the situation behind the text is much clearer, than in the hymns. The inhibitive meaning of the Inj. pres. in the RV and AV in some cases might be evident as well, cf. for instance, AVX, 1, 26: parehi krtye ma tisthah 'Go away, o witchcraft! Stand not!' etc. Nevertheless in many cases the interpretation of a wide context is needed to prove the durative Aktionsart of the action, that is prohibited. The prohibition of an iterative action seems to be connected also with future, because it is implied, that the action will be repeated in future as well. Thus, for instance, it is difficult to draw a line between the inhibitive and the preventive character of the prohibition in such cases of the Inj. pres. as in RV X, 34, 13; aksair ma divyah krsim it krsasva 'Mit den Wurfeln spiele nicht (mehr), deinen Acker bestelle!' Is not the aim of this prohibition to prevent the future game of the gambler (no matter whether he is playing at the moment of prohibition or not)? It seems to be more precise from the viewpoint of method that a consistent delimitation of tense (Present-Future), aspect imperfect-perfect) and Aktionsart (durative, iterative etc.). should be carried out. It is not by chance that a separate section, dealing with the Inj. formed from the Perfect stem, functioning in prohibitive sentences (which is the very type of sentences, where the Inj.-s are most frequent), is missing. The matter is, that in contradistinction to the Inj. pr. and Inj. aor. the Inj. pf. does not form an independent system. In a small section of this Chapter (IIC. Imperativ, Konjunktiv, Optativ und Indikativ im Prohibitivsatz, 92-98) all the other moods, that are used in prohibitive sentences, are described. It is shown by the author, that only a few cases of the Imperative and the Optative can be found in these sentences in Vedic proper; as to the Subjunctive and Indicative, they are practically missing. The description of the Inj. in prohibitive sentences is concluded by a section, where the results of the investigation are discussed (IID. Auswertung, 98-108). The fact, that in this type of sentences in Vedic only the Inj. can be used, and it is impossible to substitute it for any other mood, means, that the Inj. in prohibitions must have a function, which the other moods do not possess. One should keep in mind, that there are two elements used in prohibitions: the Inj. and the particle ma, a prohibitive negation; . The matt of this ch "nj. pf. does 4 Cf. Aopucm c Puzeede, pp. 29-35.

Page #4
250 REVIEWS so the function of the Inj. can be established only in contrast with the function of ma. The latter in its turn can be established from the opposition with the other particle of negation na, which is also used with the Inj. This second particle of negation, being combined with verbal forms (the Inj. among them) never expresses any modality. It is the negation of an action, expressing "resultative constatation". As for the negation ma, it can express prohibition not only in combination with the Inj., but also without any verbal form. That gives grounds for a conclusion, that the modality of the whole sentence depends on ma, which expresses the noeme averting, 'prohibition'. The Inj. in prohibitive sentences is allotted by the single remaining function, that of mentioning (Erwahnung) of the prohibited action, and of the person, executing this prohibition. That is why the modal meaning of the combinations with ma should not be attributed to the Inj. The Inj. is opposed to the Indicative as "mentioning" to "narrating" (Erwahnung: Bericht). The forms of the Indicative express beside the noeme "parration" that of "tense" (Present: Imperfect, Aorist). The Inj. does not express tense gradation, it is "tenseless" (zeitstufenlos). The difference of category between the Inj. with the Present-stem and the Inj. with the Aorist-stem lies in aspect, the Inj. pr. expressing the mentioning of an imperfect action, and the Inj. aor. that of a perfect one. The third chapter deals with the functions of the Inj. in non-prohibitive sentences (III. Der Injunktiv in nicht-prohibitiven Satzen, 107-204), where the Inj. is used without the prohibitive particle ma. The investigation is based upon the language of the RV, because already in the AV the Inj. without ma is comparatively rare, and it is almost completely extinct, as it was shown by Hoffmann, in the Vedic prose. It should be mentioned by the way, that although the conclusions, concerning the Vedic prose, are beyond any suspicion, the method of analysis, which is usually irreproachably strict, becomes sometimes less consistent, and in some cases the meaning of the "Inj. is established only with the help of a semantic criterion, disregarding the formal one. Cp., for instance, the arguments on pp. 108-109, where a number of augmentless verbal forms are qualified as Imperfects, that have lost their augments, and not as Inj.-s, only because of their meaning. The author is quite right in saying, that the rapid disappearance of the non-prohibitive Inj. after the RV was partly caused by morphologicai reasons the Inj. was morphologically a too ambiguous form. Some forms mentioned by the author might be interpreted in a different way. E.g., forms with the aorist-stem, having primary and not secondary endings, like krtha, bhuthah etc. might be interpreted not as a result of the tendency to differentiate the Inj. from the Iv., but as an archaism, going back to the period, when there was no clear differentiation inside the non-Perfect system into a separate Aorist and a separate Present. In the next section the Inj. expressing a general state of things is dealt with (IIIA. Der Injunktiv in generellen Sachverhalten). The aim of the description is to find out, whether there is any difference in functions of the Inj. and of the Pres. indic., expressing the noeme "general", and whether there exists a function, in which they could interchange. The word "general" describes the situation which cannot be modeled by the category of tense. When the Inj. is used in such contexts, one can't be sure, whether it should be taken for a tenseless, past or modal form. Cf., e.g., I, 168, 8: ava smayanta vidyutah prthivyam / yadi ghytam marutah prusnuvanti 'The lightnings are smiling (Inj. pr.) down, when the Maruts are sprinkling (Pr. indic.) with fat'. A clear and nonambiguous translation of such contexts is merely a matter of taste. The Inj. pr. in contexts, describing a usual state of things (III A1. Der Inj. Pras. in generellen Sachverhalten), is often used to express constant characteristics of gods (usual qualities and actions), for instance, in V, 11,2: sa barhisi sfdan ni hota 'He sits down as a hotar on the sacrificial straw'. It is not seldom that the Inj, pr. is used in the

Page #5
REVIEWS 251 same verses with the Pr. indic. They can interchange, but only in those contexts, where the Inj. pr. can be interpreted as an actual present - the meaning, which is the main function of the Pr. indic. Such general truths as phenomens of nature and characteristic actions of gods can also be expressed by the Inj. aor., and sometimes such Inj.-s can have the meaning of an iterative action, e.g. VI, 30, 2: dive-dive suryo darcato bhut 'Day after day the sun will be seen'. The general Inj. can function in rather large parts of hymns, connected with mythological events, that can not be related to a definite time. The author successfully illustrates it by some examples from the RV:1, 68; VI, 39, 3-4; 1, 173, 1-3.5 In a section, where the problem of the Inj. and the past tense is discussed (IIIB. Injunktiv und Vergangenheit, 145-235), first of all the question of the role of the augment in past tenses is put forward (IIIB 1. Augment und Augmenttempora, 145-160). As the use of the augment with past tenses is optional in the RV and AV, the augmentless forms are usually given the same status as the augmented ones. Therefore in a study of the preterital functions of the Inj. one should describe it in opposition to the corresponding augmented forms. The argumentation of the author is as follows. The Impf. (and the Plupf.) expresses in a historical narration a more remote past, while the Aor. expresses a proximate one, that is the augmented forms always express a past. As to the Inj., it may express a past tense, and may not, which means that the function of the obligatory expression of the past tense belongs to the augment. Further analysis of the contexts of the RV confirms this idea. At first the functions of Impf. indic. and of Inj. pr. are compared, these being the forms with the same type of stem different only in the presence or absence of the augment (III B2. Injunktiv und fernere Vergangenheit, 160-213). After a thorough analysis of several contexts of the RV the author comes to the conclusion, that the main function of the Impf. indic. is to narrate about the historical events of the bygone times, arranging those events in time in a logical order, i.e. correlating them with the moment of speech as a coordinate. Narration as a method of description of reality is opposed to the method of mentioning of events, their unrestricted enumeration without any logical arrangement. One of the strong points of Hoffmann's investigation is the stylistic stratification of hymns, which runs together with the linguistic analysis of the verbal forms. It is convincingly shown by the author, in what kind of hymns preference is given among the verbal forms to the Inj. Hymns of different style correspond to the opposition "narrative story" (berichtende Erzahlung): "description by mentioning" (erwahnende Beschreibung). The latter is mostly used in mythological scenes, where tense may be understood only from a wide context, or it may be irrelevant at all, for instance, in enumerations of constant characteristics of a deity (Beeigenschaftung) etc. This results in a conclusion, that if the category of tense is relevant for the Vedic augmented tenses, on the one hand, and irrelevant for the Inj., on the other, the former and the latter ones can not be functionally identical. It is shown by a comparative study of the functions of the Inj. aor. and of the Aor. indic. (III B3. Inj. Agr. und Ind. Aor., 214-235), that between the two forms a certain interchange is possible in the function of constatation. But the situation changes completely, when the Aor. indic. expresses an actual past tense: then it cannot be substituted for the Inj. The study of the functions of the Impf. indic. and the Aor. indic., on the one hand, and of those of the Inj., on the other, brings the author to the conclusion, that the 5 By the way, the principles of translation of such portions of hymns, where the actions are transmitted exclusively or mainly by the Inj., were discussed in an article in Russian, Т. Я. Елизаренкова, “о лингвистическом аспекте перевода Ригведы” История и культура древней Индии (Москва, 1963).

Page #6
252 REVIEWS usage of the augment in the language of the RV is not an optional, but a functional one. The functions of the "modal" Inj. are established in comparison with those of the other moods (III C. Der "modale" Injunktiv, 236-264). It is shown by the author in a section about the Injunctive and the Subjunctive (III C1. Injunktiv und Konjunktiv, 236-255), that many verbs do not mark clearly the difference between the forms of these two moods (for instance, bhuvat etc. can be taken both for the Inj. of the stem bhuva-, and Subj. of the stem bhu-.) In such cases one is forced to give up the morphological criterion, and to apply a semantic one, if this is possible at all. The analysis of the contexts of the RV shows that these two grammatical categories are interchangeable, only because the Subjunctive in its prospective function can acquire a tenseless meaning, and not because the Inj. possesses a special modal meaning. Such a neutralization of the modal meaning often takes place in interrogative sentences, and as to separate personal forms of the paradigm, it happens most frequently with the 1 sg., a form which is rather futural, than modal. Cf. e.g. X. 27.2: ama te tumram vrsabham pacani / tivram sutam pancadacam ni sincam.... dann will ich dir daheim einen kraftigen Stier kochen und funfzehn Tage lang scharfen Soma einschenken'. Comparing the Inj. with the Imperative (IIIC2. Injunktiv und Imperativ, 255-264) the author draws attention to the fact, that only the 2nd and 3rd persons of the Injunctive are characterized by modality proper -- they have a hortative meaning. This peculiarity of functioning corresponds to the morphological structure of the paradigms of the categories. A number of forms of both the paradigms coincide, and that explains the fact, why those forms of the Inj. and Imperative, which are morphologically differentiated, are functionally interdependent. It should be added, that the Imperatives with the aoristic stems are scarce. In the last Chapter of the book (IV. Zusammenfassung und Auswertung, 265-279) Hoffmann discusses the general results of his investigation. It is clear from the study of mutual relations between the Inj. and the system of tenses and moods in the RV, that there is no complete interchange between the Inj. and any other grammatical category. Interchange is possible only in quite definite syntactical functions. All the functions of the Inj. have one feature in common - they have nothing to do with the expression of tense, which is peculiar to the forms of the Indicative (Present, Imperfect, Aorist; it is not by chance that the Perfect is not mentioned here by the author). Thus the peculiarity of the Inj. consists in its tenselessness (Zeitstufenlosigkeit). One can see from the analysis of the non-prohibitive sentences, that the Inj. does not "narrate" (berichtet), a given action, but "mentions" (erwahnt) it. So, the characteristic functions of the Inj. are tenselessness and non-narration (Nicht-Bericht). One can see from the philological study of the "modal" Inj., that this meaning was fairly exaggerated, and only separate forms of the Inj. aor. (just those that should rather be treated from a synchronic viewpoint as the Imperative aor.) have an independent modal meaning. The difference between the Inj. with the present-stem and that with the aorist-stem is in opposition of aspects. To draw the conclusion, the Inj. expresses morphologically aspect (opposition of the type of stem), person, number and voice (opposition of the type of flexion). The function of the Inj. can be described as tenseless, non-modal and non-narrative. The Inj. is neutral towards the oppositions of tense and mood, but it is not neutral towards that of narration: non-narration. As a member of this opposition it is in contrast to the narrative tenses of the Present-Aorist system as a form, only mentioning an action. So the mentioning of action can be taken to be the main function of the Inj., and due to it the Inj. might be called Memorative. This is the most important conclusion of Hoffmann's investigation which is full of interesting ideas. A number of indices are added which are useful and exact. There is an index of quotations from the RV and other texts (Stellenindex), that of the analysed verbal

Page #7
REVIEWS 253 forms (Verbalformenindex), that of the discussed words (Wortindex) and at last an index of notions (Sachindex). Appreciating Hoffmann's conception in general for its being based on convincing philological analyses and on consistent application of strict methods of description, which brings the author to a new interpretation of the main function of the Inj., one should also make some critical remarks. The interpretation of the Inj. as a Memorative does not give rise to objections from the synchronic viewpoint. Nevertheless this conception is modeling the Inj. in a somewhat incomplete manner because the volume of the functions of this grammeme is established not by all the oppositions, in which it takes part in the Vedic verbal system. The matter is that the author does not pay attention to the relation of the Inj., which is based on the Present-Aorist system, to the Perfect. It has been mentioned several times in the description of the morphological structure of the Inj., that its forms can be built on three types of verbal stems: Present, Aorist and Perfect, though it is well known to the author, that the whole body of the Inj-s. is formed from the present and aorist stems (and it is just these two structural types of the Inj. that are actually analysed in the book), while there are only a few morphologically ambiguous Inj.-s, based upon the perfect stem. This important peculiarity of the Vedic verbal system is mentioned only in a reference on p. 276, where it is said, that one can practically disregard the Inj. pf., because the scarce forms of it are built by analogy with the Inj. pr., and have no independent function. If so, the Inj. should have been described in opposition to the Perfect as well, which should have inevitably brought the author to the necessity of a diachronic reconstruction of a state previous to that of the RV. The Inj. of this reconstructed system might be treated as a predecessor of the system of differentiated tenses and moods, and it would be opposed in this system to a pre-Perfect. This binary opposition might be interpreted as "action connected with an agent": "impersonal action, expressing state". Though the author of the book Der Injunktiv im Veda pursues a purely synchronic aim, and is not supposed to deal with reconstructions, nevertheless it would be highly desirable that in a synchronic description of the functions of the Inj. all the oppositions in which this grammeme takes part should be taken into consideration (the Perfect among them). T. Elizarenkova Satyavrat Sastri (sic), Essays on Indology. Delhi, Meharchand Lacchmandass, [1963). ii, iii, 236 pp. This work consists of thirteen essays, some unpublished, by an able Sanskritist of Delhi University. They range in length from four pages to fifty-four, and in subject from articles of general interest to the educated Indian reader to more specialized philosophical and literary studies. The opening article *Sanskrit Language and Indian Culture" is of the first type, though it shows evidence of original thinking, as when the author suggests from linguistic evidence that wheat was very little eaten in classical India, even in those areas where now it is the staple crop (p. 7). Less valuable is the next essay "Sanskrit Semantics", which relies too much on traditional etymologies -- for instance vadanya meaning "generous" is explained vadati (diyatam iti vadanyah), without basis or reference (p. 15). More specialized are the brief but interesting studies "On the words Kahala and Kirata" (pp. 32-35), "Sanskrit Originals of a few Hindi and Panjabi Words" (pp. 36-41), See Aopucm e Puzeede, Chapter 3.

Page #8
254 REVIEWS and "The Text of the Dik and Kala-Samuddesas of the Vakyapadiya and its Commentary" (pp. 42-45). "Induduta of Vinaya Vijayagani -- a Textual Study" (pp. 46-56) contains suggestions for the emendation of this little known work. "The Story of Udayana and Vasavadatta through the Ages" (pp. 57-69) compares various versions of the tale. "Poetry in the Vamana Purana" (pp. 70-81) draws our attention to the literary merits of some passages of this text. Lengthier and more important is "A Critical Survey of Sanskrit Dutakavyas" (pp. 82-138). This gives an account of the contents of over fifty duta poems in Sanskrit from Kalidasa's prototype to modern productions of a satirical character. "The Anyoktis of the Vasistha Ramayana" (pp. 139-148) draws attention to some pleasant examples of this figure of speech in the poem, giving text and translation. The longest of the essays "Conception of Time in Post-Vedic Sanskrit Literature" consists of a rapid survey of the treatment of Kala in a wide range of texts from the Smstis to the works of the Sampradayas of later Vedanta. Students intending to embark on a deeper study of the subject may find this paper useful. The last two essays of the book, "Conception of Space (Dik) in the Vakyapadiya" (pp. 205-215) and "Conception of Daiva and Purusakara in the Valmiki and the Vasistha Ramayanas" (pp. 216-236), are interesting surveys on the subjects. While the book contains little evidence of startling originality or very profound research, the essays are the work of a sensitive scholar with a wide range of interests and deep feeling for Sanskrit literature. Dr. Satyavrat's English prose, though occasionally unidiomatic, is generally readable and clear. Despite the unfortunate omission of a diacritic in the name of the author on the title page, transliteration is fairly accurate, It is unfortunate that a book which ranges so widely in the fields of Sanskrit literature, religion and philosophy should contain no index. A. L. Basham Franz Kielhorn, Kleine Schriften mit einer Auswahl der epigraphischen Aufsatze. Herausgegeben von Wilhelm Rau (=Glasenapp Stiftung, Band 3,1 + Band 3, 2). Wiesbaden, Franz Steiner Verlag GmbH, 1969. 2 Teile, XXIX + 1105 pp. DM 92,-. In 1965 the Franz Steiner Verlag reprinted Kielhorn's Grammatik der Sanskrit-Sprache (see IIJ, XI, 1968, p. 35). The Kleine Schriften contain all his articles with the exception of his editions of inscriptions, of which only those which are important for the history of Indian literature have been selected. Moreover, this collection reprints his Katyayana and Patanjali (Bombay, 1876), the introduction to the Report on the Search for Sanskrit manuscripts in the Bombay Presidency during the year 1880-81 (Bombay, 1881), the Tafeln zur Berechnung der Jupiterjahre nach den Regeln des Surya-Siddhanta und des Jyotistattva (Gottingen, 1889) and the Bruchstucke indischer Schauspiele in Inschriften zu Ajmere (Berlin, 1901). The first volume contains a systematic bibliography which simultaneously serves as table of contents (pp. VII-XXIII). Kielhorn was mainly interested in the study of Indian grammarians, epigraphy, chronology and manuscripts. All these subjects are of a highly technical nature. His works, written in a severe style, do not make easy reading and they must be studied carefully. His publications fulfil the most exacting standards of scholarship. As many of Kielhorn's articles were published in the Epigraphica Indica and the 1 A reprint of the original English version has been announced by the Chowkhamba Sanskrit Series Office.

Page #9
REVIEWS 255 Indian Antiquary it has been necessary to reproduce them in a reduced size. It is a pity that one of the two volumes could not have been published in a larger size. It would have made it easier to read the articles reprinted from these two periodicals. However, the reproduction is very clear. Professor Rau, the editor of these two volumes, has added four indices (Indische Autoren, Indische Werke, Indische Worter, Sachregister) and a concordance of the articles in the alphabetical order of the periodicals in which they were first published. In view of the fact that these two volumes together with the indices will be extremely welcome to everyone interested in Kielhorn's work, it may seem ungrateful to express a desideratum. Kielhorn's articles mostly deal with specific topics and problems. Only the specialist will know the relevant literature in each case. One wonders whether it would not have been possible to ask specialists in Indian grammar, epigraphy, etc. to list briefly the publications which relate to the problems dealt with by Kielhorn. This could probably have been done in a limited number of pages and would have made these volumes of even greater use to the reader. Australian National University J. W. de Jong Helmuth von Glasenapp: Bibliographie. Bearbeitet von Zoltan Karolyi. Wiesbaden, Franz Steiner Verlag GmbH, 1968. xiv + 100 pp. Ln. DM. 24, Helmuth von Glasenapp (1891-1963) was one of the most prolific scholars in the field of Indian studies. We learn from his biography that his motto was "nulla dies sine linea". 'Mr. Zoltan Karolyi's bibliography lists fifty-two publications in book-form and one hundred and eighty-seven articles. To this must be added newspaper articles, contributions to encyclopedias, etc, and reviews. Yet H. von Glasenapp was not a scholar who remained sitting in his study. He undertook many voyages and travelled in all five continents. In his numerous publications he has dealt with many subjects, but his main interest has always been in Indian religion and philosophy. He had a gift for explaining clearly and concisely the most difficult problems. His more popular books, which have been translated into many languages, reached a very wide public. Few scholars have contributed so much to the spread of a better knowledge of the Indian world. The bibliography, which has been carefully compiled by Mr. Zoltan Karolyi, is a very fitting tribute to the memory of Helmuth von Glasenapp. It is divided into six sections: 1. Selbstandige Veroffentlichungen; 2. Zeitschriften- und Festschriftenbeitrage; 3. Zeitungsaufsatze; 4. Beitrage in Nachschlagewerken; Tatigkeit als Mitarbeiter und Herausgeber; 5. Besprechungen; 6. Literatur uber Helmuth von Glasenapp. The index consists of two parts. The first lists von Glasenapp's publications in alphabetical order. The second is an index of proper names. Each item of the bibliography contains a full description and, if necessary, explanatory notes and cross-references. The first section lists also the reviews of von Glasenapp's books. All items have been numbered consecutively. This will make it possible to refer henceforward to von Glasenapp's publications by quoting the relevant number of the bibliography. In nr. 500 "L'absolut" must be corrected into "L'absolu" and in nr. 664 "17th and 18th centuries" into "seventh and eighth centuries". Under nr. 164 we find a cross-reference to nr. 152. One must add a reference to nr. 191. Likewise references to nrs. 164 and 181 and to nrs. 152 and 164 should be added under nrs. 152 and 191. The curators of the Helmuth von Glasenapp-Stiftung have rendered a very useful service to Indian studies

Page #10
256 REVIEWS by publishing this excellent bibliography as volume two of the publications of the Stiftung. Australian National University J. W. de Jong L. Boulnois et H., Millot, Bibliographie du Nepal, Volume 1: Sciences humaines. References en langues europeennes. Paris, Editions du Centre National de la Recherche Scientifique, 1969. 289 pp. 55 Frs. Comme l'explique dans l'introduction le professeur J. Millot, le responsable de la R.C.P. 65 (Recherche cooperative sur Programme ayant pour object le Nepal) instrument de documentation fut considere comme une tache urgente. Ce premier volume, relatif aux sciences humaines, contient une bibliographie, classee par sujets, de publications en langues europeennes. Un deuxieme volume est prevu pour les publications en langues asiatiques. L'introduction et les notes preliminaires (sources; bibliographies; a propos du classement par matieres; notes pour l'utilisation de la bibliographie) renseignent sur les sources et le classement. Chaque notice porte un numero. Les notices contiennent les details bibliographiques, et, a l'occasion, de breves annotations et des references aux comptes rendus. La bibliographie s'arrete au debut de l'annee 1968. L'ouvrage est divise en dix-huit sections; I. Guides et ouvrages generaux; II. Voyages; III. Alpinisme himalayen; IV. Cartes; V. Climatologie et meteorologie; VI. Geographie generale; VII. Histoire ; VIII. Economie; IX. Instruction publique; X. Les hommes; XI. Metiers et techniques traditionnelles; XII. Religion et magie; XIII. Art; XIV. Jeux, musique et danse; XV. Mythes; XVI. Langues et litterature; XVII. Biographies; XVIII. De quelques organismes de recherches. L'ouvrage se termine par un index des anonymes, un index des auteurs, une liste des titres de periodiques cites dans la bibliographie et une table des matieres. Dans l'introduction M. Millot ecrit: "Peu de taches sont plus malaisees et plus ingrates que l'etablissement d'une bibliographie satisfaisante: il y faut des connaissances etendues, un rare sens des langues, une ardeur au travail et une attention a toute epreuve. Aucune de ces qualites cependant ne saurait empecher une bibliographie d'etre vouee a l'imperfection. Mais il lui suffit, pour recevoir sa justification, de constituer un progres sensible sur celles qui l'ont precedee: les services rendus font oublier toutes les peines de l'execution." Le present ouvrage constitue certainement un grand progres sur ceux parus jusqu'a ce jour. Les imperfections inevitables sont mineures. Elles se limitent presqu'entierement a des points de detail. Tout au plus, on peut attirer l'attention sur le fait que les titres sanskrits ne sont pas toujours corrects (cf. 4401, 4443, 4450). La sous-section "Manuscrits et inscriptions" (pp. 228-233) enumere des editions de textes sanskrits bases sur des manuscrits nepalais ou des travaux qui se rapportent a des manuscrits nepalais de textes sanskrits. On a l'impression que le mot "nepalais" dans le titre de l'article ou de l'ouvrage est a l'origine de ces notices. Ainsi on trouvera une mention de deux articles concernant le Dharmasamuccaya (4446, 4457) mais les travaux de Lin Li-kouang sont passes sous silence. On aurait mieux fait d'omettre entierement ces notices et se borner a signaler des travaux sur des collections de manuscrits nepalais de textes sanskrits. Pour terminer quelques remarques de detail. L'ouvrage d'E. von Furer-Haimendorf ne contient pas 152 pages mais 748 pages (p. 13, VII). A propos de la bibliographie de Minaev un article est signale (p. 14, XXX). Une bibliographie beaucoup plus detaillee est donnee dans un ouvrage collectif: Ivan Pavlovic Minaev (Moskva, Izdatelstvo "Nauka", 1967, pp. 123-134). A. P. Barannikov a ecrit une bonne etude sur Minaev:

Page #11
REVIEWS 257 "Biografija I. P. Minaeva (1840-1890)" dans I. P. Minaev, Dnevniki putesestvij v Indiju i Birmu, 1880 i 1885-1886 (Moskva, Izdatel'stvo Akademii Nauk SSSR, 1955, pp. 17-32). La pagination des deux volumes de Minaev n'est pas indiquee (no. 142: Vol. I, V + 284 pp.; vol. II, 238 pp.). Le nom d'Edward Conze est ecrit partout E. Conze. No.1775 ne mentionne pas l'edition originale de l'ouvrage de Beal sur le Si-yu-ki (London, 1884). L'ouvrage de Watters (no. 1882) a paru en un seul volume dans une edition photomecanique (Dehli, Munshi Ram Manohar Lal, 1961). Parmi les bibliographies periodiques (p. 15) il y aurait eu lieu de mentionner l'Annual Bibliography of Indian Archaeology (Leyden, Kern Institute) qui contient une section sur le Nepal et le Tibet. Australian National University J. W. de Jong Ratnamayi Dikshit, Women in Sanskrit Dramas. Dehli, Meher Chand Lacchman Das, [1964]. iii, ii, 495 pp., xv plates. The fact that this lengthy book should have been written is itself evidence of the growth of feminism in India, as is the title of the ninth and final chapter, "March of Women through the Ages". Throughout her work Dr. Dikshit, a lady Sanskritist of Delhi University, focusses most of her attention on the question of the social and legal status of her heroines, and psychological and cultural aspects (e.g. the still debated question of whether respectable ladies in ancient India appeared 'topless' in public) are less emphasized. Her general conclusion is unexceptionable -- that the status of Indian women in society steadily worsened after the Gupta period, but has been improving in the last hundred years. To prove the later point she refers to certain modern Sanskrit plays, little known to the western Indologist, but evidently very interesting from many points of view. These are discussed in a lengthy chapter (VII) of seventy pages, and stimulate the reader's wish to study the plays themselves, because, though apparently still following all the rules of Sanskrit dramaturgy, some of them clearly reflect the social and political problems and tensions of near contemporary times. The female characters in the Bhaktasudarsanam of Mathura Prasada Diksita "definitely show the dawn of a new era" (p. 385). In the Naladamayantiyam of Kalipada Tarkacarya we see "man and woman on an equal status" (p. 390). The behaviour pattern of the women characters in the Samyogita Svayamvaram of Mulasankara Manikalala Yajnika "indicates the renaissance in the position of women during the present century" (p. 403). We might quote several similar examples of little known modern Sanskrit plays which appear to reflect contemporary problems. The curiosity of the social historian is stimulated, especially as the few quotations given in the footnotes show that the language of these plays is correct and enjoyable to read. This book is mainly significant as a reflection of the way in which modern Indians approach their classical literature. This is not to say that it is unscholarly or dull -- the author knows her subject well and writes interestingly. It is regrettable that so long a book does not contain a proper index, but only an index to the dramas referred to, which also serves as a bibliography. References to secondary works are very few indeed, and then are often misspelt (e.g. p. 214: Elliot and Dawson for Elliot and Dowson; pp. 27, 28: Camil Bulke for Camille Bulcke). A. L. Basham Ludwik Sternbach, Canakya-Niti-Text-Tradition, Volume II, Part I: Introduction; Part II: Canakya's Six Versions of Maxims: An attempt to reconstruct the Ur-text; Part III: B. Maxims of doubtful origin; C. Reconstructed fragmentary maxims (= Vishveshvaranand Indological

Page #12
258 REVIEWS Series, 29, 29a, 29b). Hoshiarpur, 1967 (Pt. II), 1968 (Pt. III), 1970 (Pt. I). XVI, 276 pp. (Pt. I); 1904 pp. (Pt II and III).1 Dr. Sternbach's grand publication of the Canakya-niti-text-tradition (CNTT) has been brought to completion with this second volume consisting of three parts. The first part, which was the last to appear, contains a long Introduction and 108 tables while the second and third parts are assigned to the reconstructed text. The main points of the Introduction may be summarized under the following headings. A. Manuscripts, editions of, and studies about, Canakya's collections of aphorisms (pp. 23-35): a history of the Canakya studies in modern times with an ample bibliography. B. Canakya's aphorisms in "Greater India" (pp. 35-67): on the spreading of Canakya's sayings in Burma and adjacent countries (especially on Pali Lokaniti), Java, Tibet (and Mongolia), the Niti-literature in Tamil, Ceylon, etc. with bibliography and tables. Cf. L. Sternbach, The spreading of Canakya's aphorisms over "Greater India" (Calcutta, 1969). C. Versified Cannakya-raja-niti (pp. 68-70): the author's touching upon a much debated problem concerning a statement of the Kautaliya Arthasastra itself that the work consists of 6000 slokas (I.1.18) and Dandin's reference to Visnugupta's dandaniti in 6000 slokas written for Maurya (Dasak, pt. 2, ed. Peterson, Bombay 1891, p. 52.1012). Without entering upon a detailed discussion Sternbach thus expresses his opinion on the coming into existence of the original collection of the so-called Canakya's aphorisms with the following words: "This text containing 6000 slokas, if it ever existed, could have been the ur-Canakya text. Until the time when this text be discovered we must base ourself on the theory that the original first collection of the socalled Canakya's aphorisms contained a choice of aphorisms from a treatise of polity attributed to Canakya, to which at a later date other sayings were gradually added." (p. 69). The present reviewer, however, does not believe in the metrical origin of the Kauf. Arthas. and agrees with the view that sees in Dandin's "6000 slokas" a wellknown way of fixing the extent of a prose work by using "sloka" as a unit of 32 syllables. D. Reconstruction (pp. 70-72): a section important for knowing what the author understands by ur-text. He tries to give "a probable ur-text of C.'s maxims, but not the ur-text of C.'s niti work if it ever existed. The present reconstruction is much broader, since it deals with the reconstruction of all moral, ethical and political maxims attributed to C." (p. 70). Sternbach continues to clarify his ideas about the ur-text: "This so-called ur-text is based on the six basic versions, as they have been reconstructed in Vol. I and contains all maxims which were included in one or another place in these 1 For Vol. I see IIJ, 9 (1966), pp. 301-307 where the names of the six basic versions together with their abbreviations are mentioned 2 Similarly, Canakya-raja-niti (Madras, 1963), pp. 17-19. Cf. Tsuji, IIJ, 9 (1966). p. 305, n. 3, but I believe now that Dandin referred to nothing else than the Kausaliya Arthasastra - The Canakya-Kausalya-problem should be considered in the light of T. Burrow's convincing argument that the author of the Kautaliya Arthasastra is distinct from Canakya, the minister of Candragupta Maurya (ABORI, 48/49, 1968, pp. 17-31). Cf. also Th. R. Trautmann, Kautilya and the Arthasastra (Leiden, 1971): The Kautaliya Arthasastra in its present form is a composite work in which at least three hands are discernible; the age is approximately estimated at 150 A.D. At any rate there is little chance for surmising a versified Dandaniti or Rajaniti as a source of the ur-Canakya. 8 Cf. Th. R. Trautmann, "A metrical origin for the Kautiliya Arthasastra?", JAOS, 88 (1968), pp. 347-349,

Page #13
REVIEWS 259 six versions. This text is called 'ur-text' on the understanding that each verse, each aphorism is a separate entity and the question of an organized text does not arise in this volume." (ibid.). The result is a collection of all the staznas of the six basic versions given in their original form and arranged in alphabetical order. The ur-text taken in this sense is quite different from the ultimate form of a certain text obtained by reconstruction as is familiar to us from Edgerton's 'Pancatantra Reconstructed'. On this point see further below. E. New texts (pp. 72-77): First treated is the Canakya-saptati edited by K. V. Sarma, Supplement to Vishveshavranand Indological Journal, III, 1 (Hoshiarpur, 1965). Sternbach denies the editor's opinion that this short text forms a separate version and recognizes it as a CN manuscript, for it is "a compilation of stanzas usually ascribed to C. and prevalent very often in the South" (p. 73). Next follow some remarks on the so-called notebook manuscripts (s. CNTT, I 1, p. XII) which look like notebooks of students in Sanskrit or those of school children and some of which are divided into adhyayas of equal or unequal length. Cf. Table LXIV-LXV (pp. 163-179). F. Presentation of the text (pp. 77-78, cf. also Preface, pp. V-VI): important section as it contains the definition of Sections A(= CNTT, II, 2), B, and C(= II, 3). Section A(= aphorisms 1-1119, in reality 1122 stanzas, see p. 77, n. 73) comprises the ur-text (or orginal text) of the individual stanzas of the six basic versions arranged in alphabetical order, while Section B (= aphorisms 1120-2103, in reality 960 stanzas, see p. 78, n. 75) contains those aphorisms of doubtful authority not reconstructed before in the six basic versions but found in the texts and manuscripts which formed the basis for the reconstruction of these versions, and Section C(= aphorisms 2104-2235, in reality 133 stanzas, see p. 78, n. 76) contains those of a fragmentary character. G Metres used in the reconstructed text (pp. 78-92): Besides the sloka (1841 out of 2215 stanzas), twenty-five different metres are met with and, as expected, the greatest variety is observed in the version CR, some aphorisms of which are known for their refined language and style. H. Presentation of the stanzas (pp. 92-93): explanation of references used in the critical apparatus under each stanza. The rest of the book (p. 95ff.) consists of tbales of concordance (I-CVII) showing the correspondence between the one hundred and thirteen main texts used for the reconstruction and the text as it is given in CNTT, II, 2 and 3. The last table (CVIII) is assigned to the stanzas ascribed in secondary sources to Canakya but not found in any Canakya text. Instead of discussing the appropriateness of the use of the term ur-text in the above mentioned sense, I would prefer to point out here the difficulty of reconstructing the ur-text of Canakya's aphorisms in its ordinary meaning. One usual way for reaching an original state of a versified work consists in collecting the stanzas common to all or the majority of the versions or texts concerned. Unless I have overlooked something, I could find only one verse in Section A which is found in all the versions, that is, no. 212, though even here the version CL is represented by the stanza M of Group II (CNTT, I, 2, p. 45). When one excludes CL in view of the fact that CL and Cv can be regarded as one unit (See IIJ, 9, p. 303), one obtains nine stanzas common to five versions, that is, nos. 174 (also in Pancatantra textus simplicior, Hitopadesa, 392 in Vetalapancavimsatika), 501, 527 (also in Pancatantra textus ornatior, Hitopadesa), 565, 646 (also in Hitopadesa, 746, 886, 1002 (also in Vetalapancav). Even if one contents oneself with the stanzas common to four versions, only thirty-five aphorisms come up to this standard, and not a few of them occur also in other Sanskrit works (e.g. Mahabharata, Manu, Pancatantra, Hitopadesa, Vetalapancav), CV Cv CN CS nos. 352, 355, 428, (for CN: Gr. II, B), 691,714,885. 959, 1058, 1076,

Page #14
260 REVIEWS CV Cy CN CS CR nos. 87, 142, 159, 431, 477, 507 (for CR: Gr. II, B), 828, 860, 890. CR nos. 6 (for CR: Gr. II, S), 227, 257, 625, 686 (intro ductory stanza), 779, 914. nos. 208, 269, 303, 610, (for CN: Gr. II, H), 678. CR no. 110. nos. 149, 665, 963. no. 260 (for CL: Gr. II, E). CL 80 Under these circumstances no one would think of reconstructing the ur-Canakya, that is, a collection of Canakya's aphorisms conceived as the ultimate source from which all the known versions can be derived. Is it, however, at all possible to assume an organized collection of Canakya's aphorisms in some form or other as the basis of later ramification? At first, I suppose, a mass of floating stanzas attributed to Canakya, a famous politician of yore, were spreading orally among literary circles. With the lapse of time the number of such verses increased, and they were then gathered and compiled into various collections in different times and places probably under mutual influence. In such a case the hypothesis of an ur-Canakya would be quite useless. Until a historical interrelation of the basic versions and the genesis and development of each version become more clearly known, one had better refrain from speaking of an ur-Canakya in a strict sense. Whether one may approve or not of Sternbach's designation, his collection of all the available Canakyan aphorisms is certainly an outstanding achievement involving a tremendous amount of scholarly labour and acumen. It should be appreciated not only as the starting point of future study of Canakya's compendia but also as an immense treasury of Sanskrit gnomic poetry in general. Tokyo N. Tsuji [2] Tsuji Naoshiro, Genson Yajuru-veda bunken - Kodai Indo no saishiki ni kansuru konpon shiryo no bunkengakuteki kenkyu -- (Existent YajurvedaLiterature - Philological Study of the Fundamental Sources of the Vedic Ritual -) (= The Toyo Bunko ronso, Series A, Vol. LII). Tokyo, The Toyo Bunko, 1970. xi + 211 pp. [1] H. S. Ananthanarayana, Verb Forms of the Taittiriya Brahmana (= Deccan College Building Centenary and Silver Jubilee Series, 60). Poona, Deccan College, 1970. XV + 368 pp. Rs. 30/-- Maitrayani Samhita. Die Samhita der Maitrayaniya-Sakha. Herausgegeben von Leopold von Schroeder. Erstes Buch. Wiesbaden, Franz Steiner Verlag GmbH., 1970. XLVI + 175 pp. DM 48.-; Zweites Buch. XII + 169 pp. DM 34,- Ibid., 1971. Kathaka. Die Samhita der Katha-Sakha. Herausgegeben von Leopold von Schroeder. Erstes Buch. XIV + 284 pp. DM 58,-; Zweites Buch. V + 193 pp. DM 34, Wiesbaden, Franz Steiner Verlag GmbH., 1970-1971. [4] [1] With the publication of his book on the Yajurveda Literature Professor Tsuji, the doyen of Vedic studies in Japan, has rendered a great service to Indian studies. Tsuji's book is not an introduction to the Yajurveda but a philological guide to the extant Yajurveda texts. It contains a systematic survey of the Yajurveda schools and their literature. The author studies the traditions about the transmission and the authorship

Page #15
REVIEWS 261 of the texts, their divisions, the relations to other texts, their style and date. Particular attention is given to grammatical and lexicographical peculiarities. The author is mainly concerned with the Vedic ritual; the Kalpasutras and especially the Srautasutras are carefully examined. As to the Upanisads, only philological problems are dealt with and no attempt has been made to study their content, their date and similar problems in any detail. The text of Tsuji's book occupies only 83 pages but it is followed by not less than 793 notes (pp. 85-200). As Tsuji explains in the preface, his book was originally written as a supplement to his doctoral thesis on The Relation between Brahmanas and Srautasutras which gained him the doctoral degree in 1943. His thesis was published in 1952 but without the supplement (Brahmana to Srautasutra no kankei, Tokyo, The Toyo Bunko, 1952; cf. L. Renou, JA, 1953, pp. 280-281). The supplement was based on lectures on Vedic literature, delivered at the University of Tokyo since 1935. In its original form it contained exhaustive references to the literature on the subject. The author has refrained from adding a complete bibliography of the publications which appeared during the last twenty-five years. However, nothing of importance has been overlooked. One must regret the fact that the author, who has published several important studies on Vedic ritual in English (On the relation between Brahmanas and Srautasutras, Tokyo, 1952, English summary, pp. 181-247; "The marriage-section of the AgnivesyaGshyasutra", Memoirs of the Research Department of the Toyo Bunko, 19, 1960, pp. 43-77; "Notes on the Rajasuya-section (IX, 1) of the Manava-Srautasutra", ibid., 23, 1964, pp. 1-34; 25, 1967, pp. 121-143), has preferred to publish this book in Japanese. However, it must be pointed out that the book can be consulted with great benefit even by a person who has no knowledge of Japanese. In particular, the very detailed and valuable bibliographical references, given in the notes, will be extremely useful to nonJapanese readers. Tsuji has an extensive knowledge of the history of Vedic studies. Already in 193 he compiled a bibliography of Caland's writings ("A Bibliography of the late Professor Willem Caland with special reference to Vedic studies", J. Rahder, Levensbericht van Prof. Dr. Willem Caland, Leiden, 1933, pp. 13-26). In 1939 he published a long article in Japanese on "Vedic studies, past and present" (Bukkyo kenkyu, III, 5, pp. 129-165; published in revised form in Veda to Upanisad, Tokyo, 1953); in 1948 also in Japanese "A Survey of extant Samaveda-Literature" (Gogaku ronso, I, pp. 1-37). Recently, Tsuji published a systematic bibliography of Renou's publications in Japanese: "A preliminary bibliography of important publications of the late Dr. Louis Renou", Toyo Gakuho, 49, 4, 1967, pp. 01-033). Of Tsuji's other publications in the field of Vedic studies one must mention his book on the Vedas and the Upanisads (Veda to Upanisad, Tokyo, 1953), which studies the philosophy of the Samhitas and the Brahmanas, the philosophy of the Upanisads and the ethical doctrines of the Veda. An introduction to Vedic literature from his hand appeared in 1967 under the title "The Dawn of Indian civilization" (Indo bunmei no akebono). It is not possible to mention here Tsuji's other publications which relate not only to Vedic literature but also to many other fields of Indian studies. A select bibliography of his writings, which Professor Tsuji has kindly sent to me at my request, lists a great number of books, articles, translations and reviews. Japanese readers must be considered fortunate to be able to read the works of such a distinguished scholar. A translation into English of his most important contributions to Vedic studies would be greatly welcomed by scholars in the West and in India. [2] Dr. H. S. Ananthanarayana's Index of the verb forms of the Taittiriya Brahmana consists of three parts. The first lists all forms by roots as in Whitney's Roots. In the second part all forms are given in alphabetical order with reference to their occurrence in the text; an asterisk indicates that the form occurs with a preverb and = indicates that the form occurs with a preverb, but detached from it. The third part contains

Page #16
262 REVIEWS tables of the frequency of the verb forms, a list of alphabetically arranged verb forms with attached preverbs, a list of alphabetically arranged verb forms with more than one attached preverb, a list of alphabetically arranged verb forms with preverbs detached from them, references to the occurrence of frequently occurring forms: aha, juhoti and bhavati, a list of alphabetically arranged roots in each verbal system and finally a list of ten passages containing uncertain verb forms. The three editions used for the index are those published in the Bibliotheca Indica, the Government Oriental Library Series and the Anandasrama Sanskrit Series (cf. M. B. Emeneau, Union list, nos. 207-209). One must be grateful to Dr. Ananthanarayana for having undertaken the laborious task of compiling this index. The University of Texas and the Deccan College deserve our gratitude, the first for having given a grant enabling the completion of this work, the second for having included it in the Deccan College Building Centenary and Silver Jubilee Series. [3, 4] Leopold von Schroeder's editions of the Maitrayani Samhita and the KathakaSamhita are indispensable tools for Vedic studies. Both works have been out of print for many years. One must be very grateful to the Franz Steiner Verlag for having undertaken a photomechanic reprint. According to the announcement of the publisher, the reprint of the Kathaka will also include Richard Simon's index. Book one and two have already appeared and the third book and the index are due shortly. Of the four books of the Maitrayani Samhita the first and the second have so far been published. The publishers have rendered an eminent service to Vedic studies by bringing out this excellent reprint. Professor Klaus Janert has contributed a "Vorbemerkung" which is to be found in the beginning of each volume. Janert does not mention the critical remarks made by several scholars on von Schroeder's editions. In his above-mentioned book Tsuji remarks that there are many passages which have to be corrected (cf. notes 373 and 476). It would have been useful to have a list of all the emendations which have been suggested since the original publication of these two editions. The most important studies, which contain emendations of the text of the Maitrayani Samhita and the Kathaka-Samhita, are mentioned by Tsuji. One must hope that the reprint of von Schroeder's editions will prompt a scholar to undertake a critical study of the text of both Samhitas, taking into account the emendations already proposed by previous scholars. Australian National University J. W. de Jong Hymnes de Abhinavagupta. Traduits et commentes par Lilian Silburn (= Publications de l'Institut de Civilisation Indienne, fasc. 31). Paris, Institut de Civilisation Indienne, 1970; Depositaire exclusif: Edition E. de Bocard. 107 pp. 30 F. Melle Lilian Silburn a deja consacre plusieurs travaux au Sivaisme du Kasmir (Publications de l'Institut de Civilisation Indienne, fascicules 5, 8, 15, 19, 29). Cette fois-ci elle nous donne une traduction, abondamment annotee, de huit hymnes d'Abhinavagupta. Sept de ces hymnes furent publies pour la premiere fois par M. K. C. Pandey en 1935. La Bodhapancadasika fut publiee en 1918 dans le volume XIV (non XIII comme le dit p. 2, note 3) des Kashmir Series of Texts and Studies. Dans l'introduction Melle Silburn etudie les idees philosophiques et mystiques d'Abhinavagupta. On y cherchera en vain une discussion de l'authenticite de l'attribution de ces hymnes a Abhinavagupta. Dans l'ouvrage de M. Pandey (Abhinavagupta, deuxieme edition, Benares, 1963) on trouvera

Page #17
REVIEWS 263 seulement quelques indications. La Bodhapancadasika est citee dans le Malinivijayavartika comme ecrite par l'auteur meme (Pandey, op.cit., p. 33). Sobhakaragupta qui a ecrit un commentaire sur le Bhairavastotra l'attribue a Abhinavagupta (ibid., p. 263; K. L. Janert und N. Narasimhan Poti, Indische und Nepalische Handschriften, Teil 2, Wiesbaden, 1970, p. 133). Deux autres hymnes, l'Anuttarastika et la Paramarthadvadasika, sont cites respectivement par Jayaratha et Ramyadevacarya (Pandey, pp. 7172), mais ces citations ne mentionnent pas le nom de l'auteur. En ce qui concerne l'Anubhavanivedana M. Pandey dit expressement qu'il l'attribue a Abhinavagupta en raison de la seule tradition (pp. 72-73). Les deux premieres stances sont citees dans la Hathayogapradipika, et l'ouvrage semble plutot se rattacher a l'ecole du Hathayoga qu'au Sivaisme du Kasmir. Selon M. Pandey, Abhinavagupta cite souvent ses propres stotra (p. 28). Il serait donc possible de reunir ces citations pour voir lesquels de ces stotra qui lui sont attribues sont cites par Abhinavagupta. D'autre part, il faudrait etudier la terminologie de ces stotra, non seulement des points de vue philosophique et mystique mais aussi du point de vue philologique. On se demande si Abhinavagupta a une preference pour certains mots que l'on peut retrouver aussi bien dans ses ouvrages philosophiques que dans ses hymnes. Melle Silburn ne s'occupe guere de problemes d'ordre philologique. Dans une note elle remarque a propos du mot kalana: "Doit-on comprendre kalana comme une activite subtile et kriya comme l'activite ordinaire?" (p. 67, n.4). Ce meme mot kalana se rencontre a deux autres endroits: Anuttarastika, stance 6, sankatarkakalarkayuktikalanatitah. Melle Silburn traduit: "Reste par-dela l'imperfection propre aux angoisses du doute." Les mots yukti et kalana sont omis dans cette traduction. Dans stance 10 de la Paramarthadvadasika on trouve pratihatakalana que Melle Silburn rend par "sans imperfection". Ces traductions ne nous renseignent pas sur la valeur exacte du mot kalana. On aimerait bien savoir si ce mot se rencontre dans les ouvrages philosophiques d'Abhinavagupta et, si oui, dans quels contextes. Dans l'introduction Melle Silburn exprime sa profonde gratitude au Swami Lakshman Brahmacarin avec lequel elle a lu et relu les hymnes d'Abhinavagupta. Sans aucun doute, la traduction de ces hymnes doit beaucoup a la tradition qui existe au Kasmir. Toutefois, il faut constater que l'interpretation, donnee par Melle Silburn, ne convainc pas toujours. Par exemple la stance 9 du Bhairavastava dit: tvam priyam apya sudarsanam ekam durlabham anyajanaih samayajnam. Melle Silburn traduit: "Des qu'elle a pris possession de Toi, le Bien-aime, l'accomplissement du sacrifice unique, celui de l'egalite, si ardu pour d'autres, lui est aise." Cette traduction interprete samayajna comme sama-yajna, et non comme samaya-jna. Le commentaire n'explique pas cette interpretation surprenante. La stance 10 de la Paramarthadvadasika debute ainsi: ye ye ke 'pi prakasa mayi sati paramayyomni labdhavakasah kasa hy etesu. Melle Silburn traduit: "Quelles que soient les apparences manifestees a la Conscience, on les atteint en moi, supreme firmament; car ces rayons qui sont en elles (en leur specificite)". Melle Silburn interprete kasa comme signifiant 'rayon' mais, dans une note, elle ajoute que ce mot a aussi le sens de brin d'herbe. On ne voit pas tres bien comment, dans un meme endroit, un mot peut signifier a la fois 'rayon' et 'brin d'herbe'. D'ailleurs, les dictionnaires n'ont pas enregistre un mot kasa, signifiant 'rayon' (cf. PW et pw). Melle Silburn n'explique pas ses raisons pour rejeter l'analyse de kasa en ka + asa. Il faut remarquer que les deux manuscrits, utilises par M. Pandey, ont kasametesu et kasamhyetesu. La lecon kasa hy etesu est une correction, faite par lui. On peut se demander s'il ne faut pas lire, avec un hiatus irregulier, kasa etesu. L'Introduction et les commentaires, qui suivent les traductions de chaque hymne, montrent que MelleSilburn s'est plongee profondement dans l'experience mystique des penseurs sivaites. Son experience est d'une grande valeur pour la comprehension de leur pensee. Toutefois, on ne peut pas perdre de vue les rigueurs des methodes philologiques. L'ouvrage de Melle Silburn laisse le philologue sur sa soif. Esperons que,

Page #18
264 REVIEWS dans ses prochains travaux, elle tiendra davantage compte de ses desirs a l'egard de l'edition et de la traduction de textes.1 Australian National University J. W. de Jong Venisamhara. Drame sanskrit edite et traduit par Francine Bourgeois (= Publications de l'Institut de Civilisation Indienne, fasc. 33). Paris, Institut de Civilisation Indienne, 1971 ; Depositaire exclusif: Editions E. de Boccard. 243 pp. Le Venisamhara a joui de la faveur des auteurs de traites dramaturgiques. Dans l'Inde il en existe plusieurs editions et traductions mais il n'y a qu'une seule edition due a un savant occidental, celle de Julius Grill qui date de 1871.' Melle Francine Bourgeois donne une traduction tres reussie de ce drame dont le style kavya n'est pas facile a rendre dans une langue occidentale. C'est un grand plaisir de lire cette version elegante qui suit d'assez pres les tournures des longues periodes sanskrites. Comme il ressort de l'analyse serree de la structure du drame et des notes de la traduction, Melle Bourgeois prend un grand interet a montrer comment le Venisamhara correspond aux exigences dramaturgiques. Elle reussit a faire ressortir l'importance de ce drame a cet egard. Je n'ai pas pu comparer la traduction de Melle Bourgeois avec d'autres traductions. A de rares endroits on ne peut pas accepter ses interpretations. Je signale les cas suivants: p. 124.4 bahubalao - grandes armees; p. 138.11 kusumido via taru - des abeilles couvrant un arbre; p. 150, st.14 bodhayami - pourrais-je te ranimer (160.29 pratibodhayitum - consoler); p. 154.3 degekasesapravalo -- seule force qui subsiste; p. 184.20 onirakranda - silencieux. Malheureusement, l'edition du texte laisse beaucoup a desirer. Bien que le titre de ce livre indique que le texte a ete edite, Melle. Bourgeois presente un texte sanskrit sans aucune indication de ses sources. La bibliographie mentionne quatre editions, mais on ne sait pas si Melle Bourgeois a suivi principalement une de ces editions ou a emprunte des lecons a toutes les quatre. Puisqu'elle n'avait pas l'intention de donner une edition critique du texte comme le montre l'absence de notes textuelles, elle aurait pu se contenter de reproduire le texte d'une de ces quatre editions et en faire mention dans l'introduction. J'ai note un assez grand nombre de corrections a apporter au texte. P. 38 chaya a. nanyum-manyum; p. 41, n. 3 Yodhisthita-Yudhisthira; p. 56, st. 2 tapasasya-tapayasya; p. 66 chaya f.osacakam-osucakam; p. 68 chaya c. okantio-okantio; p. 88 chaya b. pranastham-pranastam; p. 94, st. 7 dvadasarka-dvadasarka, sauryaraseh-sauryaraseh; p. 104, st. 22 snstim-smrtim, st. 24 adyasvatthamaadyasvatthama; p. 106, st. 25 adhismam-abhismam; p. 106.6 ketaparikasya-kytaparikarasya; p. 114, st. 39 carena-carana; p. 116.16 'paksapaty-opaksapaty; p. 126.8 duhsasana-duhsasana; p. 128.6 aksamah-aksamah; p. 130, st. 6 'sya vinayam-'sy avinayam; p. 132 chaya 1. 1 gathao-gadhao; p.136.1 osamghanao-samdhanao (de meme dans la chaya); p. 144 chaya b. samaraturyavarah- samaraturyaravah; p. 146, chaya a. et c. degjivadeg-jyadeg; p. 148, st. 10 oyauvananaramyasobham-yauvanaramyasobham, chaya b. kubaras-kubaras; p. 150.22 jayakarsi-jayakarksi, st. 13 krtvnayonyamkrtvanyonyam, nivstau-nirvstau; p. 156, st. 1 duran-duran; p. 164, st. 11 pravisatapravisata; p. 170. 4 aropya-aropya; p. 170.6 anyor-anayor; p. 184.6 sukhad-mukhod (il faut corriger la traduction); p. 186.20 kesambarao-okesambarao; p. 202.21 kuncara-kunjara; st. 18 mamanuko-mamanujo; p. 222.14 vairi-vairi, st. 39 nisanna 1 Dans les textes sanskrits il faut corriger les fautes d'impression suivantes: p. 23, stance 4: pratibimbatam, lire pratibimbitam; p. 48, stance 3: visvagaye, lire visvagate; p. 56, stance 4: vittamadhya', lire vittamadyao. 1 Corriger p. 22: Kritish mit Enleitung en kritisch mit Einleitung.

Page #19
REVIEWS 265 nisanna, yodhah-yodhah; p. 224.10 anabhavatu-anubhavatu; p. 226.25 degvibramahovibhramah. Australian National University J. W. de Jong Manuscript Remains of Buddhist Literature found in Eastern Turkestan. Fascimiles with transcripts, translations and notes. Edited in conjunction with other scholars by A. F. Rudolf Hoernle. Vol. I, parts 1-2. St. Leonards, Ad Orientem Ltd.; Amsterdam, Philo Press, 1970. xxxvi + 412 pp., 22 plates. Dutch Guilders 84. Hoernle's Manuscript Remains, published in Oxford in 1916, have been out of print for many years. The reprint, published jointly by Ad Orientem Ltd. and the Philo Press is extremely welcome. The Manuscript Remains contain a great number of texts in several languages. The Khotanese texts, edited by S. Konow (pp. 214-356) and Hoernle (pp. 400-402), have been edited again by H. W. Bailey (cf. the list of editions of Khotanese texts in L. G. Gercenberg's Xotano-sakskij jazyk, Moskva, 1965, pp. 16-29). At present, the Manuscript Remains are still of great importance for the Sanskrit fragments, edited by Hoernle (pp. 1-84), F. W. Thomas (pp. 85-138), H. Luders (pp. 139-175) and F. E. Pargiter (pp. 176-195) and the Sanskrit vocabulary to the texts, which was compiled by Hoernle (pp. 196-212). Moreover, Konow's edition of the Khotanese version of the Aparimitayuh Sutra also contains the Sanskrit text and the Tibetan translation (for other editions of the Sanskrit text see Yamada Ryujo, Bongo butten no shobunken, Kyoto, 1959, p. 157). Almost all Sanskrit fragments have been identified by the editors. One unidentified fragment (pp. 121-125, edited by Hoernle) has been identified as a fragment of the Suryagarbhasutra by Sylvain Levi (cf. JRAS, 1917, pp. 610-611). Another fragment, also edited by Hoernle (pp. 166-175), corresponds to passages of the Mahavagga and the Anguttaranikaya as pointed out by Hoernle. One can add to these Pali texts Theragatha 640-643. Passages, corresponding to the fragment are to be found in the Chinese versions of the Madhyamagama (Taisho, Vol. I, pp. 612c-613a), the Samyuktagama (ibid., II, p. 62a-b) and the Dharmaguptakavinaya (ibid., XXII, pp. 844c-845a). Hoernle has not tried to restore the missing passage in line four of the reverse: ksinajati ... (5) smad-bhavam prajanati. One must compare Mahaparinirvanasutra, ed. E. Waldschmidt, Teil II (Berlin, 1951), pp. 160-162: ksina me jatir usitam brahmacaryam katam karaniyam naparam asmad bhavam prajanami. The Vinaya fragments, edited by Hoernle (pp. 4-16), cannot be identified with passages from one of the Vinayas. However, Hirakawa Akira has shown that they are very similar to passages in Vinaya texts of the Sarvastivada (Ritsuzo no kenkyu, Tokyo, 1960, pp. 76-85). As far as I know, the only text, which so far has not been identified at all, is a fragment of a Mahayana sutra, edited by Hoernle (pp. 97-103). It is to be hoped that the publication of this excellent reprint will lead to further study of the texts which it contains. Australian National University J. W. de Jong Benjamin Bergmann, Nomadische Streifereien unter den Kalmuken in den Jahren 1802 und 1803. Erster Theil, Riga, 1804, 352 pp., 12 facs.; zweiter Theil, Riga, 1804, 352 pp.; dritter Theil, Riga, 1804, 302 pp.; vierter Theil, Riga, 1805, 356 pp. 12 x 19 cm. -- Reprint. Mit einer Einfuhrung von Siegbert Hummel. Oosterhout, Anthropological Publications; New York, Humanities Press, 1969. VIII + 342 pp. 22 x 33 cm. Dutch guilders 86,

Page #20
266 REVIEWS Bergmann's Nomadische Streifereien belongs to the small group of books which are of lasting importance to Mongolian and Lamaist studies. Bergmann lived for more than a year among the Kalmyks. He was a very good observer without any prejudice. The second volume contains a detailed description of Kalmyk life and customs: "Die Kalmuken zwischen der Wolga und dem Don. Ein Sittengemalde" (pp. 1-322). Other ethnographical observations are to be found in the thirty-seven letters, written by Bergmann during his stay with the Kalmyks (vol. I, pp. 32-138; vol. IV, pp. 216-355). Some brief information on Kalmyk medicine, chronology and astrology is to be found in Vol. II (pp. 324-340). For the history of the return of the Torghuts from Russia to China in 1771 Bergmann's "Versuch zur Geschichte der Kalmukenflucht von der Wolga" (Vol. I, pp. 140-246) remains one of the important sources, because he was able to obtain many details from an eye-witness, M. S. Weseloff, who was carried off by the Torghuts as a prisoner. Bergmann was very interested in the religion of the Kalmyks. The third volume contains a treatise on their religious rites: "Die Religionsdienst der Kalmuken" (pp. 72-184) and the "Ideen zu einer Darstellung des tibetanisch-mongolischen Lehrsystems" (pp. 22-70) in which Bergmann describes the Lamaist religious system. Bergmann realized that it was not possible to study the literature, the philosophy and the religion of the Kalmyks without a knowledge of Mongolian and Kalmyk. It is interesting to note that he was also well aware of the importance of a knowledge of Tibetan for the study of Lamaism: "Um das mongolische Religionssystem gehorig kennen zu lernen, wird nicht bloss erfordert, dass man die gewohnlichen mongolischen Dialekte, und die mongolische Schriftsprache, sondern auch die tangutische oder tibetanische Sprache kennen lerne." (Vol. I, pp. 18-19). Bergmann's advice has too often been left unheeded by Mongolists in later times. Several important texts have been translated by Bergmann: thirteen stories of the Siddhi kur (Vol. I, pp. 248-351), Kalmyk anecdotes (Vol. II, pp. 342-352), the Mirror of the World (Yertuncu-yin toli, cf. W. Heissig, Mongolische Handschriften, Blockdrucke, Landkarten, Wiesbaden, 1961, Nr. 3) (Vol. III, pp. 186-230), two chapters of the Gesar epic (ibid., pp. 232-284), the Mongolian version of the Visvantara Jataka (cf. Heissig, op.cit., Nr. 159) (ibid., pp. 286-302), the Geu cikitu (cf. Heissig, op.cit., Nr.158) (Vol. IV, pp. 14-180) and a chapter of the Kalmyk epic, the Dzangar (ibid., pp. 182-214). Bergmann's book has for many years been a bibliographical rarity. We must be very grateful for the publication of this reprint. By reproducing four pages of the original edition on one page it has been possible to publish the four volumes of the original edition in one large-sized volume. An introduction by Siegfried Hummel has been added to the reprint (pp. V-VIII). Hummel's knowledge of Tibetan and Mongolian seems to be rather superficial. His etymologies of Mongolian words are surprising. Yirtincu is explained as the Mongolian pronounciation of Tibetan 'Jig-rten-bcu "Zehn Behalter des Zerstorbaren, Die zehn Welten, eigentlich die zehn kosmischen Richtungen - Universum"; oron as the Mongolian pronounciation of Tibetan grong "Stadt".1 The information, given by Hummel, is sometimes incorrect. For instance, the Dzangar is said to have originated in the eighteenth century. The Dzangar contains elements of different origin. Probably it obtained its present form only at the end of the eighteenth century but some parts must be much older (cf. Nikolaus Poppe, "Das mongolische Heldenepos", Zentralasiatische Studien, 2, Bonn, 1968, pp. 198-199). For the return of the Torghuts Hummel refers to Sven Hedin's Jehol but no mention is made of the recent study by C. D. Barkman ("The Return of the Torghuts from Russia to China", Journal of Oriental Studies, II, Hongkong, 1955, pp. 89-115). No information is given on Bergmann himself apart from the indication that he comes from Latvia and that his complete name is Benjamin Furchtegott Balthasar von Bergmann. It 1 For the etymology of yertincu see F. W. Cleaves, HJAS, 17 (1954), pp. 89-90, n. 16.

Page #21
REVIEWS 267 would probably be possible to find more information about him in German libraries. The Neue deutsche Biographie (Vol. II, Berlin, 1953, p. 88) mentions in an article on Ernst Gustav Benjamin Bergmann his father Benjamin Furchtegott Bergmann (1772 ; Pastor in Rujen. Historiker und Sprachforscher). Probably this Bergmann is the author of the Nomadische Streifereien. Australian National University J. W. de Jong Franz Altheim und Ruth Stiehl, Geschichte Mittelasiens im Altertum; Berlin, W. de Gruyter, 1970; 811 S; 10 Abb. (S. 755-764). Eine Geschichte Mittelasiens im Altertum, wie im Titel verheissen, ist ein echtes Desiderat und bisher nie geschrieben worden. Wie gleich nachzuweisen, leider auch hier nicht. Voraussetzung ware die geographische und historische Abgrenzung und die Periodisierung des Mittelasienbegriffes mit nachfolgender einigermassen adaquater Abdeckung des Gesamtinhaltes, also ein Konzept mit Erfullung. Die Gliederung in neun "Bucher" (1: Zarathustra; 2: Ostiran bis auf Alexander den Grossen; 3: Alexander der Grosse; 4: Alexanders Nachfolge im Osten; 5: Die Griechen in Ostiran und Nordwest* Indien; 6: Die Parther; 7: Ausgang der Griechen in Ostiran und Indien; 8: Die Kusan und ihre Zeit; 9: Awesta. Zusammenfassung) tauscht eine gewisse Abdeckung vor, aber die Lucken sind allenthalben erkennbar: so fehlen China (Kap. 20: "Serer und Phauner" ist kein Ersatz!) und die Sasaniden in Ostiran, um nur zwei grossere zu nennen. Die einzelnen Kapitel wiederum decken den Inhalt der "Bucher" nicht ab, sind vielmehr zwar sehr geschickt gewahlte und oft sehr geistreiche Abschnitte eklektizistischer Pragung, in keinem Falle aber sorgsam portionierte und in Hinblick auf das Ganze abgestimmte Teile. Was bleibt, ist ein ausserst buntes Mosaik von Beitragen, deren Zusammenhang mit dem Haupttitel keineswegs immer klar erkennbar oder beweisbar ist. Niemand wird z.B. die Bedeutung Zarathustras in diesem Zusammenhange bestreiten, aber dieser und das Awesta nehmen unerhorterweise zwei von neun Buchern ein. Sinnigerweise sind sie getrennt (Buch 1 und 9), wie ubrigens auch der von Buch 2 bis 7 gegebene und noch vertretbare Zusammenhang durch Buch 6 (Parther) getrennt ist. Wie immer lasst A. in diesem Werk gleichfalls verschiedene andere Autoren zu Wort kommen. Ausser der nie im Detail definierbaren und sicher nicht geringen Leistung von St., deren Ungreifbarkeit die notwendige Kritik am Hauptverfasser erschwert, finden sich Beitrage von Janos Harmatta, Dieter Harnack, Roch Knapowski, Franz F. Schwarz, Zuhair Shunnar, Oswald Szemerenyi und (+) Erika Trautmann-Nehring (Bildteil), also klingende wie zunachst unbekannte Namen, auch diese gewiss nicht ohne Verdienst. Schwerlich haben sie von Anfang an gewusst, in welches Gesamtkonzept so ungleicher Formulierung sie spater eingegeliedert wurden. Auch dem Rez. ist es einst so ergangen. Im Vorwort begrundet A. das Werk damit, seine 1947/48 erschienene Weltgeschichte Asiens im griechischen Zeitalter sei seit mehr als einem Jahrzehnt vergriffen und auch auf dem antiquarischer Buchermarkt nicht mehr zu haben. Indessen ist das Wiederaufwarmen alterer Beitrage und deren verschiedene Mischung in nur zum allergeringsten Teil wirklich "neuen" Werken seit Jahrzehnten das Charakteristikum des Hauptautors. Er halt es hierin mit der Witwe Bolte bei Wilhelm Busch. Die von E. Merkel verfasste Bibliographie Altheim'scher Schriften gibt daruber zuverlassige Auskunft, wie A. einmal selbst lobte. Den Eingeweihten ist dies langst bekannt, dem Buchermarkt offenbar nicht. Es ist unmoglich, bei so disparater Themenlage und angesichts eines mangelnden echten Konzeptes auch nur stichprobenweise pro Kapitel ins Detail zu gehen, wofur auch der Raum einer Besprechung nicht reichen wurde. Dafur seien einige paradig. matische Notizen zur Arbeitsweise des Autors gebracht; Kap. 21: "Anfange der

Page #22
268 REVIEWS Parther", S. 453: dem Verf. sind samtliche neueren Arbeiten zur parthischen Numismatik (A. und B. Simonetta, Le Rider, um nur drei der bedeutendsten zu nennen) vollig fremd. Sein letzter Gewahrsmann ist W. Wroth (BMC Parthia) aus dem Jahre 1903 (sic!). Die Folgerungen uber die Aufteilung der Munzen der ersten parthischen Konige, uber die Abfolge der Konigstitel etc. sind daher unbrauchbar. Besonders hubsch ist nebenher auch der "Radmantel" des Reiters (S. 455), ohne Zweifel eine uber die Nachrichten Justins hinausgehende Bereicherung. Kap. 29: "Reges Tocharorum Asiani", S. 637ff. versucht A., aus Heraos (aus dem er einen aus den Munzlegenden nicht nachweisbaren Heraios macht) einen Sakauraker zu machen. Zu diesem Zweck schlagt er bei P. Gardner BMC, Greek and Scythic Kings of Bactria and India nach, erschienen 1886, was der Reprint 1966 verschweigt, kennt nicht oder verschweigt z.B. Zograf's Arbeit uber die Munzen der Heraos (1937) und die weitere neuere, meist sowjetische Literatur dazu, ignoriert den Emissionszusammenhang zwischen den Drachmen und den (epigraphisch und palaographisch hier allein gultigen) Tetradrachmen, macht aus der Legende Canab ein Cakab, um den Sakauraker zu erzwingen, und aus unbestrittenem, da unbestreitbarem Koccanov oder Koppanov ein Koyyanov oder Koyyakov (von P. Gardner 1886 irrig so gelesen), um den Kusan wegzubekommen, der drin enthalten ist. Derlei geht einfach nicht und ist fern jeder Verantwortung. In Kap. 32: "Chronologie der Kusan und der Hephthaliten" greift A. meinen numismatisch gefuhrten und bisher unbestrittenen Nachweis von vier aufeinanderfolgenden Wellen iranischer Hunnen (Dokumente zur Geschichte der Iranischen Hunnen in Baktrien und Indien, Harrassowitz, Wiesbaden 1967) auf S. 690 ff. vehement an, ohne indessen den Gegenbeweis fuhren zu konnen. Uber die Behandlung der Numismatik S.O. Es ware schwer entschuldbar, wurde ich diese Zeilen in eigener Sache benutzen, aber: der Vorwurf, das Material z.B. sowjetischer Museen nicht benutzt zu haben, geht ins Leere. A. selbst hat dessen Nichtachtung (s.o. Heraos) bewiesen. Im ubrigen kann man nur benutzen, was geboten wird: die Bestande russischer Museen sind mit wenigen Ausnahmen nicht publiziert, in Innerasien uberhaupt schwer zuganglich, auch R. Ghirshman, dem dies Buch sinnigerweise gewidmet ist, konnte in seiner einstigen Arbeit Les Chionites-Hephthalites (sic!) (Kairo 1948), die der nunmehr ersetzte Vorlaufer der meinen war, obwohl selbst geborener Russe und mit Sprache und Literatur des Landes hinlanglich vertraut, keine anderen Ergebnisse fur das sowjetische Gebiet Zentralasiens erbringen, wovon A. schweigt. Naturlich erneuert A. seine langst widerlegte Behauptung, die Hephthaliten seien dem Ursprung nach "Turken". (S. 697). Das hinterlassene grosse Hunnenwerk von 0. (nicht E., so S. 693) Maenchen-Helfen, eben im Druck, wird Altheims Hunnentheorien endgultig zerstauben und uns weiterer Muhen entheben. Dank fur Erschliessung neuer Quellen ist bei A. nicht zu finden, weder hinsichtlich meiner "Dokumente" noch sonst. Kritik wird liebevoll auch im Detail geubt: wir wollen gerne hinkunftig die Kusan so und nicht, wie Altheim H. Humbach und mir ankreidet, Kusan schreiben, wenn es ihn nicht stort, dass er selbst es bisher in zahlreichen Werken stereotyp so getan. Die historischen Auslassungen uber das Datum des Kaniska werden darum nicht richtiger: Ohne sich der grundlegenden historischen Verschiedenheit bewusst zu werden, nimmt A. in Parallele zur PostYazdgard-Ara einen Beginn der Kaniska-Ara mit dem Tode des Konigs an (statt mit dem Regierungsbeginn, wie doch die an die 50 indischen Inschriften ausweisen, deren erste sofort einem zweiten Kaniska zuzweisen waren, was das numismatische Material als Quelle glatt ausrangieren wurde) und kommt (recht versteckt auf S. 701) auf sein Datum 219/20, obwohl er S. 687 sieht, dass mein Ansatz auf rund 230 n. Chr. durch die von mir als baktrisch erkannten Tochi-Inschriften uber die Rechnung von J. Harmatta auf 232 n.Chr. als sicher erwiesen wird. Wenn A. sich auf eine Ausfuhrung von Harmatta 1966 (S. 687, Fn. 13 belegt) stutzt, dieser hatte auch sogleich gesehen, dass in keinem der Falle bei den Tochi-Inschriften eine Bilinguis vorlage, so ist festzuhalten, dass Harmatta in Acta Ant. Hung. XVII/3-4 (1969) dies nicht bestritten hat (1.c. 368:

Page #23
REVIEWS 269 "... the dating i.e. die Datenidentitat) ... seems to be justified"), was A. entweder nicht weiss oder verschweigt. Harmatta benutzt die Gleichung ubrigens 1.c. 374. A. hatte die scharfsinnigen Arbeiten von Harmatta auch sonst besser auswerten konnen. Mit diesen Proben aus der allgemeinen und aus der besonderen Verfahrensweise des Verf., die allen Regeln strenger historischer Disziplin Hohn spricht, ist leicht ein Bild davon erreichbar, woraus das ganze Werk besteht. Der Autor ist einst angetreten, im Gegensatz zum Wirken der "Ritter der vereinigten Broselkrume" (Toynbee) ein weitschauendes Bild der Antike entwickeln zu helfen. Er hat darin manche bewundernswerte Leistung erbracht, fur die wir ihm auch jetzt noch Dank schulden und die nicht vergessen werden darf. Unberuhrt von jeder ernsten und sachlichen Kritik lasst sich auch einst Grosses aber nicht in die Gegenwart ziehen. Forschen heisst, nach einer geistreichen Definition, sehen, was jeder sieht, und denken, was keiner denkt. Altheim hat ihr lange, wenn auch nicht ausschliesslich, entsprochen. Sehen aber, was nicht ist, und nicht sehen wollen, was ist, ist in ihr nicht enthalten. Universitat Wien Robert Gobl Walther Hinz, Altiranische Funde und Forschungen; mit Beitragen von Rykle Borger und Gerd Gropp. Quart. 272 Seiten (einschliesslich 157 Tafeln, 34 Abbildungen, 1 Kartenskizze, 2 Farbtafeln). Berlin, Walter de Gruyter & Co., 1969. Aussere Grundlage dieser Publikation sind zwei Iranreisen des Vf. 1967 und 1968. In ihnen hat er nicht nur Bekanntes uberpruft, sondern auch manches Neue gefunden oder fur rasche Publikation gesichert und hier vorgestellt. Der Dank der Fachwelt gilt ihm hiefur von vornherein, denn es ist der Wille klar erkenntlich, durch Zusammenfassung des Verstreuten ebenso zu klaren wie durch Anfertigung zahlreicher besserer und ganz neuer Fotos, auf denen endlich manche umstrittene Details besser ausgemacht werden konnen. Die zahlreichen Bildtafeln des Buches gehoren daher auch zu dessen bestem Teil. Die einzelnen Kapitel umspannen einen weiten Interessenbogen des Vf. und ergeben, wie hernach ersichtlich, zwei Schwerpunkte: einen elamischen und medischachamenidischen und einen sasanidischen Teil. Von keinem Rez. kann unbedingt erwartet werden, dass er allen Abschnitten gleichermassen gerecht wird. Auch dem Vf. ist nicht alles gleichmassig gelungen. Abgesehen vom Methodischen, uber das noch zu reden sein wird, ist klar ersichtlich, dass er im ersten Teil weit besser daheim ist als im zweiten, im Philologischen besser als im Archaologischen. Unter dem begreiflichen Wunsch, Gefundenes und Gesehenes moglichst rasch herauszubringen, hat die echte Durchdringung der Materie, fur die eine wenn auch gute Kenntnis der Sekundarliteratur keineswegs ausreicht, haben Terminologie und Diktion sichtbar gelitten. Das ist bedauerlich und erschwert auch eine rechte Wurdigung, weil gerade dort, wo etwas zu sagen ist, soviel zu sagen ware, dass es den Rahmen einer Rezension bei weitem sprengen musste. Einige dieser Themen werden denn auch -- darunter eines vom Rez. - von Anderen erneut und voller aufgegriffen werden mussen. Kap. I. Eine neugefundene altelamische Silbervase (S. 11-44). Sie ist 1966 gefunden und das bisher einzige Obiekt dieser Art. Den Rez., dem sie (nach Foto) auf den ersten Blick verdachtig war, was er hier -- obschon bekanntlich kein Elamist -- deponieren mochte, hat zu lesen beruhigt, dass die Echtheit der Vase schon bei der Vorlage auf dem V. Internationalen Kongress fur Iranische Kunst und Archaologie in Teheran 1968 in Zweifel gezogen worden ist. Ob der Rettungsversuch von R. Ghirshman (Art. Asiae 30/1968, 243ff.), dessen Meriten ihn oft genug nicht vor Irrtumern bewahrt haben, hier zahlt, bleibe dahingestellt. Schwerer wirkt das Urteil des Verf., der sich fur die Echtheit entscheidet.

Page #24
270 REVIEWS Wenn H. (S. 12) sagt: "kein Falscher vermochte eine derartige Inschrift zusammenzustellen - und ware er selbst ein Elamist", so verlangt er damit Glauben in einem sehr weiten Ausmass und wirft eine Grundfrage des modernen Falscherbetriebes auf, die hier nicht angeschnitten oder beantwortet werden kann. Festzuhalten sind indessen die hier weiter bekraftigten und unbestrittenen Verdienste des Vf. um die Entzifferung der elamischen Schrift, der er die entscheidenden Wendungen gegeben hat. Nach H. kennen wir heute etwa 55 annahernd gesicherte Zeichen, wahrend in seiner Schatzung noch 50-60 fehlen. Wir sind also auf weitere Funde angewiesen. Kap. II. Eine neue Xerxes-Inschrift aus Persepolis (S. 45-52) Die im Jahre 1967 gefundene Inschrift ist - und darin enttauschte sie -- blosse Kopie eines Teiles der unteren Grabinschrift des Darius in Naqs-i Rustam (aber nach H. "offensichtlich von einer Urkunde der grosskoniglichen Kanzlei, moglicherweise von der ursprunglichen Niederschrift des Darius-Diktates"). Nur der Name des Xerxes ist eingesetzt und eine Schlussformel angefugt. Immerhin war es durch sie moglich, die meisten Textunklarheiten der an diesen Stellen beschadigten Darius-Inschrift auszuraumen. Das "Epigonentum des Xerxes" + beurteilt nach der Vorstellung geistloser Kopie einer der aussagetrachtigsten Inschriften seines Vaters -- scheint ein etwas zu abfalliges Urteil. Es fragt sich, da die Inschrift doch die Maximen der Herrscherqualitat und -verantwortung in einmaliger Diktion enthalt, ob nicht gerade die Absicht des Xerxes darin bestand, dieses selbstgeschaffene "Grundgesetz" zur Tradition zu machen und damit uberdies eine dem Vater gleiche Grosse zu beanspruchen. Was hatte manbei intendiert gleichem Gedankeninhalt -- nach einem kaum ubertreff baren Standardtext, der uns noch heute in voller Lebendigkeit anspricht und packt, wirklich Neues erwarten durfen? Dass Xerxes uns hier keine neue achamenidische Inschrift beschert hat, ist wohl bedauerlich, aber ob das zu einem Verdikt ausreicht, ist mehr als fraglich, es so simplifiziert auszusprechen, bedenklich. Kap. III. Die dreisprachige untere Grabinschrift des Darius (S. 53-62) Dabei hat Rykle Borger (Direktor des Seminars fur Keilschriftforschung an der Univ. Gottingen) die akkadische Version, Hinz die altpersische und die elamische Fassung bearbeitet. Damit sind -- ein beachtliches Faktum -- erstmals alle drei Versionen zusammen publiziert und kommentiert. Die Xerxes-Inschrift (s. Kap. II.) hat dabei ihren Nutzen erwiesen. Kap. IV, Medisches und Elamisches am Achamenidenhof (S. 63-94) Der Vf. stellt in Untersuchungen an den achamenidischen Reliefs, besonders an jenen in Persepolis, fest, dass Reiten, Fahren und Sitzen des Herrschers Sache eines medischen Hofmarschalls mit medischem Personal war, was m.E. durchaus uberzeugend dargetan wird. In der Hoftracht glaubt der Vf. starken elamischen Einfluss feststellen zu konnen. Hier ist er vielleicht manchmal zu weit gegangen. Dass z.B. der Streitwagen elamisch ist, ist kaum glaublich (so auch K. Schippmann in seiner Rez. des Werkes in OLZ 66/1971, Nr. 5/6, 235). Dass Elamer auf ihm fahren, ist noch kein Beweis, dem der Vf. ubrigens ganze zehn Zeilen widmet. Zum Gestus des medischen Hofmarschalls auf dem sudl. Schatzhausrelief in Persepolis (Tf. 19) fehlt S. 63, Anm. 4 jeder Hinweis auf die langeren Erorterungen, die ihm und dem Thema der Proskynesis etwa F. Altheim und F. Schachermeyr u.a. gewidmet haben. Da hier Gerate des Feuerkults zwischen Konig und Marschall stehen, ist die Auffassung des Vf. in Fn. 4 (Schutz des Feuers vor Verunreinigung) und nicht im Haupttext (Schutz des Herrschers vor Belastigung durch den Atem!) richtig. Auch das haben

Page #25
REVIEWS 271 Andere vor dem Vf. gesehen, was festzustellen gewesen ware. Zu S. 73 (Tf. 23): dass alle Figuren der Tafeln 21, 22 und 23 stehend und nicht gehend gedacht sind, ist schon durch die rasch ausgreifende Schrittstellung des Pferdes auf Tf. 23 (Parallele auf Tf. 24) widerlegt. Kap. V. Die Volkerschaften des Persepolis-Reliefs. (S. 95-114) Bekanntlich sind dort 23 Volkerschaften abgebildet. Die beste Konfrontation der bisherigen, sehr divergierenden Ansichten findet sich bei G. Walser, Die Volkerschaften auf dem Relief von Persepolis (=Teheraner Forschungen 2, Berlin 1966). Nur rund die Halfte der Volkerschaften ist als einhellig identifiziert anzusehen. Der Vf. verringert diese Zahl zwar durch Neuzuweisungen (12=Ionier; 15= Areier), ist aber sichtbar bemuht, Hypothesen durch Fakten aus dem Unsicherheitsbereich herauszuheben, sodass seine Vorschlage ernstlich erwogen werden mussen, weil seine Methode gangbar erscheint. Die folgenden vier Kapitel sind sasanidischen Felsreliefs gewidmet und bilden den zweiten Schwerpunkt des Buches. In ihm engagiert sich der Vf. besonders, ohne dass erkennbare echte Ambition und Voraussetzungen miteinander immer Schritt hielten. Hauptanliegen des Vf. ist die Deutung und Identifikation der dargestellten Personen. Er hat zu diesem Zweck folgerichtig und systematisch bestmogliche Aufnahmen zu machen getrachtet und hierin liegt, wie schon eingangs betont, ein bleibendes und echtes Verdienst. Er kennt weiters die Sekundarliteratur gut, und vor allem sind die Angaben der Res gestae divi Saporis, also der Sapur-Inschrift in Naqs-i Rustam eine seiner wesentlichsten Stutzen. Es ist unmoglich, auf alle Argumentationen im Detail einzugehen. Beachtliches und vollig Unannehmbares stehen in dichter Folge nebeneinander. Kap. VI. Die Felsreliefs Ardashirs I. (S. 115-144) Der Vf. veroffentlicht -- gewissermassen als Beitrag zu einem Corpus, das er nicht liefern kann, wobei man sich fragt, ob sich nicht doch ein solches lohne und wirklich heute so schwer zu erstellen sei (was kein Einwand gegen den Vf. ist!) -- samtliche funf Felsreliefs des Reichsgrunders Ardaser I. (224-241). Uber Methoden, Terminologie und Diktion usw. sei am Schluss die Rede. Hier nur einige Notizen: Relief I, S. 119 (zu Tf. 52/53 -- Kampf mit Artaban V.): Der Konig tragt weder Krone noch Mitra, sondern nur die Lockenfrisur, wie sie die Munzen zeigen (R. Gobl, Sasanidische Numismatik, Braunschweig 1968, Tab. I, Av.-Typ V, 3. Krone). Sie wird nur wie im Zugwind des zu Pferd Dahinsturmenden nach hinten weggezogen und deformiert. Der Konig tragt ein Diadem, wofur der Vf. trotz Herzfelds, Erdmanns und des Rez. Arbeiten uber die sasanidische Krone, die immerhin eine recht anerkannte Terminologie geschaffen haben, "Ring" setzt. Ahnlich verfahrt der Vf. mit den auf den Hauben und Haubenkronen von Prinzen und Wurdentragern ublichen Tamgas, die der Vf. (auch spater, bes. Kap. IX) "Wappen" nennt. Der Ausdruck Tamga ist ein zwar spaterer, aber immer noch der treffendste Terminus fur diese Wurden- und Hoheitszeichen vornehmlich der iranischen Volker. Sie kommen zweifelsfrei aus dem nomadischen Bereich, waren ursprunglich Besitzmarken und werden erst uber regierende Clans zu echten Hoheitszeichen. Da der hier eingefuhrte Begriff Schule machen konnte, muss eindrucklich festgestellt werden, dass der Ausdruck "Wappen" bekanntlich der europaischen Heraldik entstammt und auf anderem Hintergrund gepragt ist. Was der Vf. meint, lasst sich nach den strengen Regeln nur als "Schildfigur" bezeichnen, was es im Sasanidischen aber eben nicht ist. Zu S. 127 (Tf. 68): Die Bedenken des Vf. konnen beseitigt werden: wie W. B. Henning im Hb. d. Orientalistik IV/40 nachgewiesen hat, gibt es von Artavazd keine Munzen. Die ihm zugeschriebenen Drachmen sind (BMC pl. 36, 14-15; Petrovicz Tf. 23, 12 u. 15) nach

Page #26
272 REVIEWS der parthischen Rv.-Legende solche des Artaban V, was der Vf. ubersehen hat. Interessant sind die Ausfuhrungen uber die "Puscheln", die an den Pferdeschabracken hangen. Die Vermutung von Renate Rolle (Gottingen) auf (ursprunglich) SkalpHaare besiegter Feinde, eine zentralasiatische Sitte, die sie fur die Scythen mit Herodot IV, 64 belegen kann, hat etwas fur sich. Das Problem (der Vf. lasst die Deutung offen) musste unter Heranziehung nachsasanidischen Materials monographisch losbar sein. Relief von Salmas: der Rez. bedauert, dass der Vf. mit keinem Wort seine lange und unbeantwortet gebliebene briefliche Stellungnahme zur ersten Publikation des Vf. in Iranica Antiqua 5/1965, 148-160 erwahnt, worin u.a. das Relief als mogliches Halbfabrikat bezeichnet wurde, was ganz andere Aspekte ergabe. Im ubrigen zeigt die Zeichnung von Ker Porter 1819 fur das rechte Personenpaar die Uberreichung eines Ringes (Kranzes), was folgerichtig auch fur das linke Paar gilt. Von einer Kolah kann keine Rede sein. Kap. VII. Das sasanidische Felsrelief bei Darab (S. 145-172) Kap. VIII. Die romischen Kaiser der Shapuhr-Reliefs (S. 173-188) Beide Kap. mussen in einem gesehen werden. Fur Darab verdanken wir dem Vf. erstmals gut brauchbare Aufnahmen. Der Vf. will die These von MacDermot (Roman Emperors in the Sassanian Reliefs, JRSt 44/1954, 76ff.), der erstmals die Deutung der Personen auf den funf Siegesreliefs des Sapur I. auf die drei Kaiser Gordianus III., Philippus I. Arabs und Valerianus I. (fruher galt nur die kniende Hauptfigur als Valerian) dergestalt umstossen, dass er uberall Philippus I. fur Valerianus I. e vice versa einsetzt. Der Rez. bedauert, dem Vf. hier nicht folgen zu konnen. Er hat in einem Seminar in seinem Wiener Institut 1970/71 die Frage genau untersucht und muss grundsatzlich MacDermot (gegen eigene fruhere Ablehnung) Recht geben, wozu von ihm eine Publikation unter Heranziehung aller, hier teilweise arg vernachlassigter Quellen zu erwarten ist. Vgl. auch die zusammenfassende Kritik nach Kap. IX. Kap. IX. Karders Felsbildnisse (S. 189-228) Zugrundeliegt ein- und der Rez. zweifelt nicht daran, dass er richtig ist - fundamen- . taler und folgenschwerer Schluss des Vf. Dieser identifiziert Karter uber dessen Relief in Naqs-i Radjab auch auf dem Sapur-Relief in Naqs-i Rustam, kommt damit auf das Tamga des Karter, das er "Scherenwappen" nennt und vermag uber dieses auf drei weitere Reliefs sicher den beruhmten Grossmobad zu identifizieren, da feststeht, dass der Trager des Tamgas in allen Fallen eine und dieselbe Person sein muss. Unverstandlich ist, warum der Vf. Karter in Naqs-i Rustam erst beweisen zu mussen glaubt, der ahnlich wie in Naqs-i Radjab (dort fast neben) uber seiner Inschrift steht. Entschieden muss aber gegen die naive Deutung des "Scheren"-Tamgas (vom Vf. wegen seiner Form in einer modernen Gedankenassoziation so genannt) Einspruch erhoben werden. Der erblickt in der Schere "einen sinnbildlichen Bezug auf Karders Entscheidungsbefugnisse: als hochster Richter besass er die Macht zum "Durchschneiden" auch der schwierigsten Probleme" (S. 191). Dies ist eine Simplifizierung der Interpretation, gegen die das Durchhauen des gordischen Knotens noch ein recht kompliziertes Verfahren genannt werden muss. Mit solchen Scheren kann die Iranforschung ihre schwierigsten Probleme jedenfalls nicht durchschneiden! Was zugrundeliegt, ist nach Meinung des Rez. eher ein stilisierter Hase, wie das sasanidische Siegelbilder und manche Teppichfiguren nahelegen, in denen der ausgezeichnete Persienkenner allerlei Verwandtes entdecken konnte, was hier unerwahnt bleiben muss. Ahnlichen Tierhintergrund haben das Tamga der Kusan-Dynastie (Hirsch/frontal) und jenes der Alxon-Hunnen (Stier/frontal) u.a.m. Die Zuruckweisung der Deutung des Rez. (S. 194, Fn. 5) auf die Kronprinzen auf

Page #27
REVIEWS 273 dem Relief des Varhran II. in Naqs-i Rustam ist nicht annehmbar, solange nur Meinung gegen Meinung gesetzt wird. Wenn die Figur neben dem Konig die Konigin ware, musste sie die fur sansanidische Damen unabdingbaren Zopfe tragen, die indessen nicht sichtbar sind. Kap. X. Einige neuentdeckte Inschriften aus sasanidischer Zeit (S. 229-263) Gerd Gropp (Deutsches Arch. Institut, Abt. Teheran) bearbeitet hier sehr verdienstvoll und grundlich neue Inschriften, auf die Hinz ihn hingewiesen hat. Unter ihnen sind besonders bemerkenswert eine kurzere Fassung der Haggiabad-Inschrift von Naqs-i Rustam (Pfeilschuss Sapurs I!) in Tang-i Buraq (sasanidisches und parthisches Pehlevi), eine Inschrift aus Eqlid aus dem Jahre 6 des Yazdgard (vermutlich III.), die auch durch vertikale Zeilenfuhrung bemerkenswert ist, ferner eine griechische Inschrift am Relief des Sapur I. in Naqs-i Rustam. Bei dieser ist unverstandlich, dass Gropp die griechische Version der KbZ-Inschrift weder heranzieht noch erwahnt, bei der er fur seine Ausfuhrung der Gleichung mit Pahlavik und Parsik doch alle gewunschten Parallelen gefunden hatte. Warum er nur die (allerdings ausgezeichneten) Fotos nimmt und keine Abklatsche angefertigt hat, erfahren wir nicht. Zu dem zweiten Teil des Buches scheinen einige Bemerkungen grundsatzlicher Art notwendig: Der Vf. geht in der Behandlung sasanidischer Felsreliefs eigene Wege. Sein Identifizierungswille ist anerkennenswert (die Nennung zahlreicher Prinzen und Wurdentrager in den Inschriften fordert diesen schliesslich heraus), verleitet ihn aber offenbar und unverstandlich zu methodischen Sunden. So etwa, wenn der Pariser Cameo mit der Gefangennahme Valerians uberhaupt nicht herangezogen, ja nicht einmal dort genannt wird, wo dies unabdingbar ist, namlich erstens S. 146 zum Relief von Darab (wo die ganze Erorterung der "Helmkappenkrone" in ein anderes Licht getreten ware) und zweitens ebenfalls in Kap. VII und dann in Kap. VIII zur Identifizierungsfrage der drei Kaiser. Der Vf. musste den Cameo ablehnen; weglassen kann er ihn nicht. Es ist evident, dass auf diese Weise auch die Glaubwurdigkeit der zahlreichen gedankenvollen und erwagenswerten Anregungen und Interpretationen des Vf. ernstlich gefahrdet wird. Grundeinstellung, zahlreiche Termini und Einzeldeutungen sind verfehlt. Der Vf. hangt glaubig am Wahrheitsgehalt der KbZ-Inschrift an allen Stellen, ubersetzt (S. 179) beim Tod Gordians das Ettavrpn mit "fiel", wo doch der Ausdruck intentionell farblos gehalten ist (=wurde dahingerafft), begnugt sich fur die Version des Todes mit A. Maricq (S. 173/174: "Recherches" 1952) und erwahnt nicht die weit glaubhafteren romischen Nachrichten (wonach Philippus den jungen Gordian unter dramatischen Umstanden ermorden liess), so als ob es keine moderne kritische romische Geschichtsschreibung gabe; fur die Numismatik gilt bei ihm, was ahnlich O. MaenchenHelfen einst fur das Chinesische sagte, namlich es wurde nicht gelesen, aber missbraucht. Statt einen ausgewiesenen numismatischen Fachmann zu konsultieren, was fur die Triumphreliefs seit dem (eher abschreckenden) Beispiel von MacDermot unabdingbar ist, oder selbst die Reichsmunzen des Valerianus anzusehen, begnugt er sich mit dem (hinsichtlich des Ausdrucks Provinzpragungen) missverstandlichen Hinweis von H. von Gall, der selbst kein Numismatiker ist, auf R. Delbruecks Werk uber die "Munzbildnisse von Maximinus bis Carinus" (S. 182) fur Valerians Bart. S. 146, um nur weniges herauszugreifen, finden wir "Korymbosreif" statt Diadem, eine "Tunica" bei Sapur; wir sehen einen "Dolch oder Kurzschwert" herabhangen, indessen deutlich ein Dolch (mit Klinge innerhalb der Hose) steckt; wir vermissen die Benennung des in der Hand des Grosskonigs befindlichen und S. 149 richtig gedeuteten flachen, rechteckigen Gegenstandes als naturlich einer "tessera"; S. 178 finden wir ein "Band der Herrschaft". Die Liste liesse sich beliebig verlangern. Sie zeugt von grenzenloser Unbekummertheit, die ubrigens auch in der teils in der heutigen Iranistik unublichen, teils unhomogenen, teils auch inkongruenten Schreibweise sasanidischer Namen sicht

Page #28
274 REVIEWS bar wird. Wenn dieser Stil der Forschung Mode wird, wird in Hinkunft in der deutschsprachigen Iranforschung vielleicht zwar mehr geschrieben, aber weniger gelesen werden konnen. Der Rez. bedauert, dass er der Tatsache dieser standigen Mischung von beachtlicher Belesenheit, Gelehrsamkeit und grosser Sachkenntnis mit Unzulanglichkeit, von Ambition, redlicher Muhe und enormen Fleiss mit ausgesprochener Fluchtigkeit, an so ausgedehnten und so diffizilen Sachverhalten praktiziert, mit keinem anderen Ausdruck Rechnung zu tragen weiss. Das Werk ist ohne Zweifel ein echtes und trotz der zahlreichen Schwachen bedeutendes Stimulans der Forschung, aber auch eines ihrer Menetekel. Robert Gobl Muhammad Mokri, Le Chasseur de Dieu et le Mythe du Roi-Aigle (Dawra-y Damyari). Texte etabli, traduit et commente avec une etude sur la chasse mystique, le temps cyclique et des notes linguistiques. Wiesbaden, Otto Harrassowitz, 1967. 161 pp. + 52 pp. Persian and Gurani text and 2 pp. plates. DM. 58.--. This book, dedicated to E. Benveniste, is the first volume of the Beitrage zur Iranistik edited by Professor Georges Redard of Bern. It contains the first Chapter of an exposition of the Creed of the Ahl-i Haqg, written in Gurani, the sacred language of this sect. The edition is based on a recent copy of the Kalam-i Khazana (also called Kalam-i Perdiwari), made in 1934, but authorized by a member of the religious directorate of the sect. In Gurani opinion this is an important and authoritative work on the teachings of the Ahl-i Haqg, which sect - "Men of God" (Minorsky); "Fideles de la Verite" (Mokri) -- is widely spread among the Kurds of the province of Kirmanshah and also has its followers elsewhere in Iran, in Irak and in Turkey. The Ahl-i Haqq understand by Kalam a collection of sayings in poetic style spoken by God, the five Angels of His retinue and a few other prominent religious personalities when engaged in discussion at the time of their various incarnations in the seven epochs of world history, the culmination of which was reached in the fourth, the epoch of Sultan Sihak in the eighth century of the Muslim era. This Sultan Sihak is honoured as the founder of the Ahl-i Haqq religion. The Kalam-i Khazana contains 26 kalam, the first of which, called Dawra-y Damyari, "Epoch of the Hunt", forms a separate whole comprising 221 strophes of 4 hemistichs. The structure of the kalam is this, that after an opening verse spoken by God again and again a number of angels - not always all of them nor always in the same orderbegin to speak and utter one or more strophes. Except at the end, after strophe 204, the angels appear under the names which were theirs in the period of their incarnation in the eighth century A. H. The hunter (damyar) is the Angel Gabriel, here called Binyamin. Being created before all other creatures he is nearest to God and, therefore, in charge of attending to the observation of the pact concluded between God and the Angels to the effect that God pledged Himself to manifest Himself to the human race in the course of the centuries. Comparable to a hunter who with a net or a lasso (dam) in his hand follows the trail of the royal bird of prey (shahbaz), Gabriel-Binyamin, who on account of his priority of creation is pre-eminently fit for this task, traces the theophanies in the various epochs of the history of mankind. This span of time, an interlude between pre-eternity (azal) and eternity (abad), is the period during which men and angels, passing through a thousand incarnations of an average of fifty years each, may attain the perfection they need to see God at the end of the world. It is obvious that the teachings of this sect are alien to Islam as it is professed in Iran.

Page #29
REVIEWS 275 They bear a marked impress of gnosticism and testify to the continuity of autochthonous ideas as well as to the presence of notions that impart a syncretistic colouring. Therefore, the explanatory notes added to the translation of the text are far from superfluous; they elucidate much that otherwise would have remained an unsolved puzzle to the uninitiated reader. The plan of Dr. Mokri's book is as follows. After an introductory chapter dealing with the manuscript and the text he provides us in the second chapter with copious data concerning the net or lasso; the hunter; the ever returning phrase nayna tananis, "in that place he spread it (his net)"; the following of the trail; the royal eagle and the myth of the King-Eagle, and other symbolic birds in Iranian folklore and literature. One will find here many things made known previously in papers read by the author at the International Congresses of Orientalists in Munchen (1959) and Moskou (1963). In a third chapter the author finally deals with cyclic time and the different modalities of time to wit the exoteric aspect of history and its epochs, and the esoteric aspect of the timeless Reality vaulting it. The annotated translation of the text is followed by linguistic and grammatical notes; a glossary of substantives and one of proper names; a list of the places mentioned in the text; a detailed index, and finally a bibliography of works consulted, comprising no less than 60 titles, among which several manuscripts. The author's own publications in this field are mentioned separately on p. 2 of the Introduction. The work concludes with a short introduction in Persian and the Gurani text of the Dawra-y Damyari, printed in Arabic character. The author may be complimented with this important contribution to our knowledge of the religious ideas, the folklore and the sacred language of the Ahl-i Haqq. He has supplied valuable new materials, and added a good deal of information to what was already known from other sources as, for instance, the publications of De Gobineau, Minorsky and Ivanow. Noordwijk G. W. J. Drewes J.-P. de Menasce, Feux et Fondations pieuses dans le droit sassanide. Paris, Klincksieck, 1964, in-8deg, 62 p. L'unique ouvrage de droit existant en pehlevi. le Matigan i hazar Datistan "Livre des mille Jugements", n'a pas encore ete edite; il doit l'etre prochainement, en U.R.S.S., par Anahit Perikhanian. Le droit sassanide ne peut se comprendre a partir du droit romain, mais seulement du droit hellenistique. Le P. de Menasce a extrait de ce texte difficile tout ce qui regarde les feux, en donnant une transcription et une traduction impeccables de tous les passages. Il y a ajoute la traduction d'extraits du Datistan i denik et de la Rivayat d'Emet concernant les notions de sardarih familiale et de sturih. Il procede ensuite, sur la base de ces documents, a l'etude de cinq questions: la nomenclature des feux, le personnel des feux, la sardarih des feux, la sturih des feux et les fondations pour l'ame. Il y a peu de fautes d'impression: p. 49, milieu, lire Maguan Andarzpat; p. 51, ligne 10 du bas, aux trois quarts; p. 52, ligne 29, on conclura; p. 62, ligne 14, un secours; note 33, ligne 4 du bas, ruvanakan. Liege J. Duchesne-Guillemin

Page #30
276 REVIEWS A[ndras] Rona-Tas, Tibeto-Mongolica: the Tibetan loanwords of Monguor and the development of the archaic Tibetan dialects (= Indo-Iranian monographs, 7). 232 pp. Budapest, Akademiai Kiado; The Hague, Mouton, 1966. The extraordinary importance of this monograph for the study of Tibetan historical phonology, as well as for the history of the Mongol languages, has already been brought to the attention of scholars concerned with these subjects in a number of reviews; in addition to my own notice in Language, 44 (1968), 147-68, the following reviews have come to my attention (and there may well have been others that have escaped me): M. I. Vorobjeva-Desjatovskaja, Narody Azii i Afriki, 1966, 5, 189-92; R. K. Sprigg, BSOAS, 30 (1967), 216-17, G. Kara, Acta Orient. Hung., 20 (1967), 377b-81b. Furthermore, I have had the favor of a long personal communication from Professor Rona-Tas, dated Budapest 30th September, 1968, in which he has been good enough to comment at some length on a number of points that I raised in my review published in Language. It may, therefore, be appropriate to utilize this additional short notice of this important work in order to comment briefly upon some of the points raised in the reviews that have appeared, and also some of the items discussed by Professor Rona-Tas in his letter. Mme Vorobjeva-Desjatovskaja's review aims more at introducing the Russian reader to the contents of the monograph than at a critical appraisal of its contents; hence she does not go deeply into any points of detail. Nevertheless, in the course of summarizing the contents of the monograph, she does focus attention on one important matter that I must confess had escaped my attention until I saw it stressed in her review. This is the passage (pp. 187ff.) in which Rona-Tas deals with pitch and 'tone' in the modern Tibetan dialects. When he writes," [Y. R.) Chao (Love songs, p. 27) and Miller (Writing, p. 2) distinguished the pitch as a phonemic quality from tone, the first of which is the level of intonation, the second the sinking-rising or the change of the relative level of intonation", I am afraid he does not give a completely accurate view of what I attempted to do in my treatment of the suprasegmentals of modern spoken Tibetan, or, for that matter, an accurate account of what Y. R. Chao attempted to do in his pioneering study of the Lhasa language. Both of us were working in the same way, along what would now be generally considered hopelessly old-fashioned Bloomfield-Bloch-Trager lines of 'phonemics'; tone, pitch, movement up, movement down, anything that the voice did that 'was not 'obviously' a 'vowel or a 'consonant' was a 'suprasegmental'; and we grouped these 'suprasegmentals' into distinctive 'tones', or 'tonemes', according to the same body of methodological assumptions by means of which we grouped two or more 'phones' into 'phonemes'. Whether what we were doing was 'right' or 'wrong', and whether it should continue to be done that way today, would involve questions of methodology and problems in linguistic theory rather far removed from the interests of this journal, and - in these post-Chomsky days - probably outside the competence of the reviewer. But for the sake of the student who may one day wish to re-open the investigation of such matters, the point should be made -- and without the review of Mme VorobjevaDesjatovskaja the problem would most likely have continued to escape my attention - that, most regretfully, I find Rona-Tas' description of my attempts to 'phonemicize' the Tibetan suprasegmentals a little far from the mark. But I am sure that we would both agree that, in this point in the history of Tibetan studies, what we need above everything 1 "Studies in Spoken Tibetan, I: Phonemics", JAOS, 75 (1955), 46-51. 3 Y. R. Chao (Jaw Yuanrenn) and Yu Dawchyuan, Love songs of the sixth Dalai lama Tshangs-dbyangs-rgya-mtsho (= Academia Sinica, National Research Institute of History and Philology Monographs, Series A, No. 5) (Peiping, 1930).

Page #31
REVIEWS 277 else is not additional theoretical or analytic work on the pitifully small amount of available data, but new field-work with native speakers, and the accumulation and publication of new linguistic data, covering every facet of the language, the tones (or 'pitches', or whatever) included. The review by R. K. Sprigg is little more than a description of the contents of the monograph, and raises no interesting points. His insistence that "W(ritten) T(ibetan) Ih- is and presumably always has been" a digraph for a voiceless [1] (p. 217b) is no doubt correct enough, and Rona-Tas' postulation of a shift O(ld) Tibetan) Ih-> hl(p. 128) should probably be omitted from a history of Tibetan phonology, but the point is minor, and more a question of the symbols and notation we use for our formulations than it is a problem in the history of the language. Interestingly enough, several of the points in Rona-Tas' monograph that I thought it worthwhile to comment upon in some detail in my 1968 review were also signaled out for special attention in the review by G. Kara that appeared the year earlier; it is perhaps needless to add that the two notices were, of course, prepared independently of each other. Thus, Kara comments especially upon Rona-Tas' etymologies #26 and #29, both of which also attracted my attention (cf. my p. 163). What Kara has to say about #26 is important, and should be taken into consideration in any future study of the problems presented by this etymology; it still seems to me that the possible effect of a Chinese intermediary form on the history of the lexical items involved should be given further consideration. In the case of #29, I am afraid that honesty now compels me to say that neither my comments in my 1968 review, nor Kara's remarks, actually shed much light on this still vexing etymology. Another item of fortuitous coincidence between our two reviews is to be found in the treatment of the developments of original Tibetan initial n- (Rona-Tas, p. 123-4; Kara, p. 381a; Miller, p. 165). In this connection, I now wonder if, once again, an intermediate stage of linguistic history involving loans through and from Chinese might not with profit be invoked in order to throw some light on those Tibetan dialects in which original initial n-appears to have shifted to initial zero (though, as I pointed out in my review, the Central Tibetan form for 'camel' is not one of these). Here I particularly have in mind the zero-initial forms of modern Mandarin. "This initial, in the majority-type pronunciation, does have a slight consonantal-type obstruction in the form of a frictionless velar or uvular voiced continuant... A (large) minority of speakers use a glottal stop or a pure vocalic beginning for all words with a zero initial. A very small minority of speakers begin such words with a consonantal ng-..." This is, of course, the feature of modern Chinese pronunciation that accounts for the *continental Sinological transcription of the syllables that are, for Wade-Giles, an, as ngan (Tch'ang-ngan, etc.). Middle Chinese had both initial glottal stop and initial n-, as well as initial zero (what Karlgren calls 'smooth vocalic ingress'), but in the history of the forms involved, there has been much analogic shifting back and forth among these categories. Perhaps we have here to deal with an areal feature, borrowed from one language family into another language family, unrelated but contiguous; or perhaps it was simply the forms that were borrowed, and re-borrowed. Finally, it is interesting to note that Rona-Tas' remarks on the semivowel u attracted Kara's attention (pp. 380b-381a) as they did my own, and with good reason, since they are important, and deal with a critical feature in the historic phonology of Tibetan as well as of Mongol (I return to this point shortly below). : Yuen Ren Chao, A grammar of spoken Chinese (Berkeley and Los Angeles, 1965), pp. 26-7. 4 Berhard Karlgren, Grammata Serica (Stockholm, 1940), p. 49.

Page #32
278 REVIEWS In his personal communication of 30th September, 1968, Rona-Tas was kind enough to comment in detail on several points that I had raised in my Language review; and while it will unfortunately not be possible to enter here upon a detailed treatment of all these, I would still like to take the opportunity of the present notice to comment briefly upon a few of these items. Rona-Tas suggests in his letter that my remarks concerning his etymology #199 (cf. my review, p. 163) are somewhat wide of the mark; and at this point, while I still feel that a number of complex early Altaic connections between forms is involved here, I must agree with him that my own etymological suggestions, in my review, actually oversimplified this very difficult problem, and thus tended to obscure rather than to clarify the issue. In the case of Tibetan gur 'tent' and its kin, we are surely in the presence of an ancient Altaic mot voyageur, the peregrinations of which have embraced a range of time and space so vast that they will probably always remain beyond the grasp of our comparative grammar and its methodology. In the case of #222, Rona-Tas agrees that "the preradical of dpar 'form, pattern, printing-block' could be theoretically something similar to bskal-pa or bca-'chin" - in other words, nothing more than an orthographic flourish, and totally without importance for the history of the form in Tibetan. "But", he continues, "in this peculiar case I see in the dialectal reflexes /xyar/ an argument in favour of the dp-initial which gives regularly xy- while p- does not." Again, I am afraid the somewhat oversimplified presentation of my review (p. 162) tended to obscure rather than to clarify the history of this important word. It was an oversimplification to write, as I did, that there is no evidence for the d- of dpar in the Tibetan dialects. The dialects show evidence for an earlier initial consonant cluster in this word. But whether the elements were d and p, as the received orthography would have us believe, is another question. I continue to find it difficult to disassociate this Tibetan form from Middle Chinese *p"an 'id.'. Even if the Chinese form was not borrowed directly into Tibetan (which is still what I believe happened), its shape may very well have influenced the configuration of the Tibetan form, as well as its eventual semantic extension--for I think that we must all agree that there are very few historical facts as well established as the fact that printing is a Chinese invention, not a Tibetan one! The final -r going with Chinese -n has excellent parallels in a number of Chinese loans into Old and late Old Japanese; details would take us too far afield, but the materials noted long ago by S. Yoshitake, BSOS 7 (1935), 940-41, come immediately to mind in this connection. In the case of the initial, of course, it would have been the combination of *p followed by the labial semivowel -- for *pw is as much a consonant cluster as dp or gp or any other sequence of two non-vocalic segments - that was originally borrowed into some early Tibetan dialect; either in this dialect or in some other one closely related to it, the form was shifted to initial x*-; and later still, the orthographic combination dp was employed to write this initial of the Tibetan form, the form that of course directly underlies the Monguor forms xuor and xuar cited and studied by Rona-Tas. The point is worth making again, and I would hope here with more clarity than in my original review, if only to obviate any possibility of survival for Shafer's thesis that the initial dp- of Tibetan dpar is to be taken as an indication that the technique of wood-block printing originated not in China, but in Tibet. And of course the existence of WT dpe 'pattern, model, type, form, custom, example', that is borrowed as Monguor xue 'parable, comparison, story', cannot but have played an important role in the development of the Monguor and other non-Tibetan versions of this word. About #464 (cf. my p. 153), Rona-Tas writes me, "the problem with reGul 'winter' is the final -1, and it seems to me that this is due to a contamination with Mong. egul. If so, then egul has influenced an earlier Monguor *reGun and not a Tibetan rgun. Naturally the Mahavyutpatti form is of the highest importance and shows that the Monguor *reGun goes back to a Tibetan rgun which can be ascertained as early as the Mahavyut

Page #33
REVIEWS 279 patti." This is in response to my citing the form rgun from Mhy. 9282, a form that seemed to me to render unnecessary the hypothetical form *reGun that Rona-Tas reconstructed to account for the shape of the initial in the Monguor form. To be sure, the final -1 of the Monguor form does present a problem in associating the form with its Tibetan etymology, but I did not comment upon it because the problem seemed to me to be completely dealt with in Rona-Tas' footnote 65 to his p. 77, where he helpfully reported that in response to his inquiry to ascertain whether or not the form in question might simply be a misprint (it can be cited from only a single passage in a single text, at least from among published materials), Professor Schroder had replied that this and all other examples of this word in his still unpublished materials "are in the genitive: regunni". From this I concluded that the Monguor form to be considered here is not actually reGul but reGun, and that the form reGul was nothing more than an ad hoc creation of an artificial isolation form by Schroder's informant(s), either made up ad lib, or with the final -1 indeed due to contamination with Mongol egul, but at any rate not appearing in the texts and hence not meriting further consideration. In the meantime, it is interesting to note that another citation of WT rgun 'winter' has now come to light, and that from an unexpected quarter. It is to be found in the Paris - Societe asiatique Ms. (of K'ang-hsi date) of the Hsi-fan-kuan i-yu, now edited, published, and studied in detail by Nishida Tatsuo.5 The form is item #100 as numbered by Nishida, and appears on p. 85 of his publication; the Paris Ms. has rgyun in Tibetan script, but the transcription of the Tibetan pronunciation into Chinese characters makes it clear that a form [rgun) was intended, and as Nishida suggests, the -y- in the writing rgyun is surely nothing but an example of dittography, anticipated from WT rgyun-du'always', which is #121 (on the same p. 85 in Nishida's edition). Nishida, who certainly now knows more about these Ming-Ch'ing interpreters' and translators' vocabularies than any other scholar in the field, claims that the Paris Ms. represents a "15th century Literary Tibetan from the Amdo region", the language that he calls his "Seibango A [i.e., Hsi-fan-yu A]". The appearance of the form rgun in this source, where its pronunciation is attested by a (for once) unambiguous Chinese transcription, represents a valuable falling-together of several diverse sources of data. In #560, my review questioned the proposed pattern of semantic development; Rona-Tas writes, "the semantic development to make dry, lean, meagre' > 'to lose weight, to be meagre' remains clear for me; to be meagre, thin is e.g. in the case of meat synonymous with dry, at least in some languages, among them in Hungarian." His remark is valuable, but it shows again in rather sharp relief the insecure areas we all tread when we become involved in 'patterns of semantic development', for what will seem perfectly obvious to one person (most usually because of some relationship, actual or supposed, between forms in his own language) will never occur to another (whose own language may not have any parallel related forms). It is an area of linguistic speculation in which, most unfortunately, almost anything is equally possible, and almost anything becomes equally probable, if we think about it longer. This is what I intended to point out in my original review, nothing more; but I still think the point is worth stressing. And much the same still holds true for #772, which etymology RonaTas was also good enough to mention in his letter. 5 Nishida Tatsuo, Seibankan yakugo no kenkyu, Chibetto gengogaku josetsu (= Kaiyakugo kenkyu sosho, I) (Kyoto, 1970). This publication represents a continuation of Professor Nishida's work with these materials, begun with the publication of his paper "Juroku seiki ni okeru Seikosho Chibettogo Tenzen hogen ni tsuite -- Kango-Chibettogo tangoshu iwayuru Heishubon 'Seibankan yakugo' no kenkyu", Kyoto Daigaku Bungakubu Kenkyu Kiyo, 7 (1963), 85-174. As late as my 1968 review of the Rona-Tas monograph, I was still misreading "Seibankan' as 'Saibankan', as for example on p. 167 of that review, a long-standing error of mine that was finally pointed out to me by Professor Nishida when I last saw him in Kyoto, in the summer of 1969.

Page #34
280 REVIEWS Finally, Rona-Tas' letter again directs my attention to a number of important problems, each too involved to be considered here in anything resembling a complete manner, but concerning which I would like to comment briefly as follows: 1. In connection with my remarks (pp. 154-5) on the possibility of influences by way of non-NE Tibetan loans, Rona-Tas writes as follows: "In Buddhist terminology learned borrowings and in everyday business life the Tibetan koine had surely an important influence - not to speak about pilgrims. I was surprised, nevertheless, (to find in my study) that most of the Monguor forms showed neither literary nor Central Tibetan or Lhasa items as their immediate source." This is an important point, and very well-taken; and the remarks in my review (which still, I think, are of value) should certainly be read with this statement in mind. 2. In connection with my remarks (pp. 156-7) on the problems presented by the labial semivowel x in Monguor loanwords from Tibetan, Rona-Tas expresses himself as remaining unconvinced by my arguments, and suggests that "we have here to do with an areal phenomenon". This seems to me a most important suggestion, and one that should be followed up by future students of the problem, though I still do not see why the development of a labialized velar phoneme from an earlier labial-plus-velar sequence should present any particular difficulties of interpretation. 3. In connection with my remarks (p. 158) on vowel-harmony assimilations in Tibetan forms underlying Monguor loans, Rona-Tas writes as follows: "The possibility of reflecting Tibetan vowel harmony features by Monguor forms is made difficult by the fact that vowel harmony is an essential part of the phonological system of any Mongolian language, so naturally also of Monguor. The so-called 'breaking' can also be taken into account in such cases as araDag, murGuo, suro, etc." The comments in my review should have paid more considered attention to the implications of the Mongol phonological canon than they did, and the point that Rona-Tas makes is of importance in any future consideration of these issues. 4. In connection with my remarks (p. 159) on the importance of the problem of morpheme-boundaries in Tibetan compounds, Rona-Tas points out that his etymology #285, Mgr. lisGa 'work, deed, custom' can only be explained in the light of a morpheme-division different from that of WT las-dka', var. las-ka, i.e., as li-sGa< *la(s)-ska, an important example that had until now escaped my attention. But on my suggestion with regard to the morpheme-segmentation of WT rdo-rje 'vajra-sceptre', Rona-Tas writes, "I do not think that a segmentation rdor-je can be justified." Since the suggestion that the forms in the living Tibetan languages point in the direction of an earlier rdor-je, rather than toward the segmentation of the form in the received orthography, WT rdo-rje, is one that I have tentatively put forth in the literature several times since 1955 (references in my review, p. 159), a few additional lines here in support of such a view may not be out of order. The problem is, to be sure, essentially concerned with the internal morpheme boundaries of the form, in historical terms, but it does not stop there, since the initials of each of the two morphemes, and especially of the first, are equally critical in the problem. Briefly, the most important considerations that enter into the question are the following: (a) If the Tibetan form that historically is responsible for Mgr. Duor Dzi, var. Dor DZi 'vajra-sceptre' is correctly represented by the received orthography of WT rdo-rje, either the initial of the first morpheme, or the initial of the second morpheme, must be anomalous in terms of its historical development. In the Monguor loans, original Tibetan r- before -C- either is retained as such, or otherwise appears in transparent guise as s- or s-: if we center our attention solely on forms with original initial rd-, and -o- vocalism (and hence exactly parallel to the phonological configuration of the form at issue), we find WT rdo/,, Mgr. $D- (Rona-Tas, p. 91), WT rdo-'bum, Mgr. reD- (p. 77), WT rdog, Mgr. ar D- (p. 40), WT rdon-rgan, Mgr. rD- (p. 75), and WT '/dor-mo, Mgr. $D- (p. 91). This is why Rona-Tas writes, "OT r-- [appears as)

Page #35
REVIEWS 281 ar- in the first syllable [in a single form]: $- before non-palatal or non-palatalized radicals in every position ...; [and] - before palatal or palatalized radicals ..." (p. 138). But the Monguor form for 'vajra-sceptre' has only initial D(u)s, not *reD-, or any other overt indication of an original r- before -C- in its representation of the initial of this form. (b) Nor do any of the dialect forms that can be cited for 'vajra-sceptre give any hint of an original r-before-C-in this word. Those forms that can be cited, and that usually are given (as for example, in Rona-Tas, p. 45, #93), are not dialect reflexes of the form 'vajra-sceptre'; they are dialect reflexes of the word for 'stone', e.g. Bal. rdoah, Pur. rdoa, Lad. rdoa, all 'stone'. They go, of course, with WT rdo 'id.'. The same is true, it is worth noting, of the fifteenth-century linguistic data represented by Nishida's 'Seibango A', where WT rd- before -o- is regularly and uniformly represented in the pronunciation-transcriptions into Chinese characters in the Paris - Societe asiatique Ms. as [rd], with the single exception of the form 'vajra-sceptre', where we find anomalous [d]: rdo # and page in Nishida WT #53, p. 83 #905, p. 118 rdo(-mthin) #483, p. 100 rdo (rin-po-che) #936, p. 120 rdog(-cig) #283, p. 92 rdo-rje Pronunciation-transcription [rdo] [rdo] [rdo] [rdo] [do-rd3e] (c) On the basis of these data, either of two equally satisfactory conclusions is possible: (a) the etymological identification of the first morpheme in the Tibetan word for vajra-sceptre' suggested by the received Tibetan orthography, by which this first morpheme is taken to be identical with the word for 'stone', is not, in historical fact, to be maintained; or, (B) rdo-rje came into Monguor, and into the 'Seibango A' language, and so also into most (if not all) the modern Tibetan dialects (resp. languages), not as a regular genetic inheritance but rather as a loan from some earlier dialect, now otherwise apparently lost, in which the r-element in the initial rС- had been simplified to zero; but this would necessarily have been a dialect quite different historically from that responsible for the bulk of the Monguor loans. (d) But in the meantime, and regardless of whether (a) or (1) above is elected, the phonological canons of the modern dialects known to me, and particularly of modern Central Tibetan and Lhasa, determine that a historical development from a rdor-je segmentation is regular, while a development from the received rdo-rje segmentation is anomalous. (e) Even apart from all this, that a sequence rdor is something more than an abstraction arrived at solely on the basis of linguistic evidence, may easily be demonstrated by a consideration of the shape of the morpheme WT rdor, which appears as a combiningform for rdo-rje, as for example in phyag-rdor 'vajrapani', the compendious equivalent of Mhv. 649, phyag-na rdo-rje 'id.'. Nor may it be suggested that this is a mere lexicographer's ghost; for in fact, Rona-Tas (pp. 45, 95) twice cites yaraypu 'id' from the Tangutsko-Tibetskaja okraina, II, of G. N. Potanin (where it appears in text J, 'from an unnamed Tangut'). A simplification phyag-na rdor-je > phyag-rdor is regular; phyag-na rdo-rje > phyag-rdor is anomalous; hence, I opt for the former. This, incidentally, remains true whether we think in terms of an original rdor-rje > rdor-je, or an original rdor-je with no prior assimilation and simplification -r-rj-> -r-j-. And so both varieties of evidence, linguistic and philological, point in the direction of final -r in the prior morpheme, rather than an open-syllable in -o with initial r- in the next morpheme. (f) Finally, on this important word, I would in addition be tempted to point out

Page #36
282 REVIEWS that the -u- of Mgr. DuorDzi, otherwise unexplained, goes quite regularly with the *-y- that I ventured to reconstruct in *rdva 'stone' on the basis of totally independent evidence (ZDMG, 106 [1956], 351; cf. Rona-Tas, p. 45, n. 12), were it not for the fact that, with such an explanation, we would return full-circle to the traditional identification of the first morpheme in 'vajra-sceptre' with the word for 'stone', and at this stage at least I am not particularly anxious to support that identification, for the reasons set forth above. If such an identification is historically correct, then my 1956 reconstruction has obvious value in explaining the otherwise anomalous -u- of the Monguor form; but if it is not, then the coincidence of the forms is fortuitous. 5. In connection with my remarks (pp. 160-62) on problems in phonemic theory and terminology, Rona-Tas now points out to me in his letter something that I am sure both of us can only heartily agree upon--namely, that our field is now and probably always will be a long way from having a uniform phonological methodology for historical problems, reflected in a uniform terminology, and that if we wished to engage in the creation of new linguistic terms, the field of historical Tibetan phonology would provide if anything a too-rich area for such endeavors. Fortunately, I note that neither of us shows much inclination to indulge in such activities, both wishing to reserve our time and energies for more productive work. In commenting on my remarks relating to the phonemic role, within the original Tibetan orthography, of the graph generally transliterated h, Rona-Tas writes in his letter, that "... some reason has to be given for the fact that the Tibetan orthography - which was not phonemic - chose the same symbol for different phonemes"Much of this discussion clearly hinges upon what each of us means by 'phonemic' - and just as clearly, each of us means something rather different - but in terms of what phonemic' means to me (and that, I fear, is not very much different from what it meant to Bloomfield and Bloch, nor in any way 'advanced' over their views), the Tibetan orthography was phonemic, and, as I wrote in 1968, "the decision to lump together two phonetic entities in complementary distribution, one an initial voiced velar spirant, the other a homorganic nasal initial in consonant clusters, under a single phoneme, WT h, is a phonemic decision as old as the Tibetan script". Ordinarily, I do not believe there is much point in multiple reviews of a single title by the same hand; but I have broken my own rule on this occasion in part because the kind offer of the editors of the IIJ provided an opportunity to follow-up on at least a few of the many important problems treated in this monograph, particularly in the light of the other reviews, and especially in the light of my subsequent correspondence with Professor Rona-Tas himself. But mostly, I have been motivated by my high regard for this monograph, and my appreciation of its great scientific value for the entire field of Tibetan linguistic studies. One cannot but agree most heartily with Professor Kara's encomium, which I take the liberty of citing in conclusion, as representing most accurately my own views: "En fin de compte, on ne peut que souhaiter la naissance de nouveaux travaux pareils aux Tibeto-Mongolica dont l'importance est incontestable pour les recherches sur la linguistique comparative des dialectes tibetains, terre presque inculte auparavant. Il n'est pas douteux, le livre est un nouveau poteau indicateur a la frontiere tibeto-mongole." (Acta Orient. Hung. 20 [1967), 281b). Seattle Roy Andrew Miller

Page #37
REVIEWS 283 M. Andronov, A Standard Grammar of Modern and Classical Tamil. Translated by the author from the original Russian (U.S.S.R. Academy of Sciences - Institute of Oriental Studies). Foreworded by S. K. Chatterji and T. P. Meenakshisundaran. Madras, 1969. xii, 342 pp. Rs. 15.00. Unter den zahlreichen Werken, mit denen sich M. S. Andronov um die Dravidistik verdient gemacht hat, ragen Dravidijskie jazyki (Jazyki Narodov Azii i Afriki.) (Moskva, Akademija Nauk SSSR -- Institut Narodov Azii, 1965) und Grammatika Tamiljskogo jazyka (Moskva, Akademija Nauk SSSR -- Institut Narodov Azii 1966) besonders hervor. Es ist ausserordentlich zu begrussen, dass letzteres Buch jetzt durch eine gute englische Ubersetzung einem grosseren Interessentenkreise zuganglich ist. Die englische Ausgabe hat daruber hinaus den Vorzug, dass durchweg die Original-Tamil-Schrift zusammen mit einer "breiten" phonetischen Umschrift, aus der man die phonemische Schreibung selbst unschwer ableiten kann, verwendet worden ist. Das Erscheinen einer modernen Tamil-Grammatik kommt einem echten Bedurfnis entgegen. Die alteren Arbeiten, wie die von A. H. Arden, A Progressive Grammar of Common Tamil (Madras, 1934) und H. Beythan, Praktische Grammatik der Tamilsprache (Leipzig, 1943), reichen bei weitem nicht aus, so gut sie an sich sein mogen. Es handelt sich bei ihnen weniger um die zum Teil verstandlicherweise veraltete Methode der Sprachbeschreibung als vielmehr darum, dass sie der verwirrenden Vielschichtigkeit des Tamil nicht gerecht werden. Das Tamil ist, ahnlich wie zahlreiche andere indische Sprachen, gewissermassen drei- oder sogar vierdimensional: Es sind nicht nur die verschiedenen Sprachepochen (alt, mittel, modern) und raumlich gegliederte Dialekte zu unterscheiden, sondern eine immens grosse Rolle spielt auch der Unterschied zwischen Schrift- und Umgangssprache sowie der der einzelnen sozialen Schichten des Volkes (Brahmanendialekt usw.). Arden bringt nur "Common Tamil", Beythans Basis ist etwas umfangreicher; er fuhrt auch die alten literarischen Formen (deutlich gekennzeichnet durch #) an; aber voll berucksichtigt diese Vielschichtigkeit erst Andronov in seiner trefflichen Grammatik. Gewiss liegen zur Zeit bereits verschiedene Sonderarbeiten uber einzelne soziale und raumliche Dialekte, auch uber die Besonderheiten der Sprache bestimmter alter Texte vor, aber es fehlte bislang eine ubersichtliche Zusammenfassung in einer handlichen Grammatik. Der Verfasser ist nicht den hypermodernen Methoden vieler jetziger Linguisten gefolgt, was kein Fehler sein durfte. Die Hauptsache ist schliesslich die adaquate Beschreibung der Sprache, und die ist hier geleistet worden. Neben der Berucksichtigung der Vielschichtigkeit gruppiert und ordnet Andronov den Stoff zum Teil anders als seine Vorganger, bringt zahlreiche neue Gesichtspunkte hinein und beleuchtet manche Einzelheit dieser interessanten Sprache erstmalig. Sehr begrussenswert wird es der Leser finden, dass alle Daten durch Beispiele aus den verschiedensten literarischen Werken belegt werden; dadurch wird die Darstellung ungemein lebendig und die Lekture des Buches zur Freude. Im folgenden wollen wir verschiedene Punkte der Grammatik beruhren und einige erganzende Bemerkungen einflechten. 1. Bezuglich der Vokale folgt Andronov der Tradition sowie auch der Auffassung verschiedener moderner Dravidologen und setzt lange Vokale (a i ue o) gesondert an. Da aber Geminaten in der Sprache zahlreich sind (kk, tt usw.), ware es gunstig, auch bei den Vokalen Geminaten anzunehmen, wie es schon J. R. Firth (im Anhang bei Arden) tat: /aa ii uu ee oo). Die Fuge muss dann gegebenenfalls bezeichnet werden: maraa-ati oder maraa'ati (p. 57). Als "speech-sounds" fuhrt Andronov zusatzlich die nichtphonemischen retroflexen Vokale i ie e und uberkurze Vokale i u (besser ware die Schreibung u) an. i wird von Beythan nicht erwahnt. 2. Sehr eindrucksvoll ist das luckenlose System der Verschluss Reibelaute mit je drei

Page #38
284 REVIEWS Gliedern pro Phonem. Dazu kommt jeweils ein Nasal. Die Nasale sind ausser phonemicsh: pt t t ck b d d d jg BS r s S X mn n n n n [r] fur intervokalisches /t/ fehlt bei Beythan. Vielleicht handelt es sich hier um eine ganz junge Entwicklung, die aber nicht ausnahmslos eingetreten ist; es findet sich in dieser Position auch [d]. Bei /k/ waren nicht nur die Allophone [k g x] zu vermerken, sondern, wie Beythan angibt, noch zusatzlich das palatale (c) und das stimmhafte [y]; auch stimmhaftes [6] kommt vor. Letzteres ist auch die gewohnliche Aussprache fur (Grantha-) h, das auf keinen Fall mit x transliteriert werden sollte (p. 9), weil x die passende Schreibung fur aaytam ist, ein Graphem, das Beythan weniger gut durch k wiedergibt. - Als Transliteration schreibt Andronov fur den alveolaren Verschluss/ Reibelaut der Tradition folgend j, als Phonem setzt er hingegen sehr richtig /t/ an (p. 32). Besser ware in beiden Fallen die Schreibung mit t. Vorzuschlagen ware auch die Phonemansetzung /r/ statt /z/ (und konsequenterweise dann [8] statt [r]). /n/ und p/sind nach Andronov Phoneme, da intervokalisch beide Laute vorkommen, z.B. vallunan 'starker Mann', panai 'Weinpalme', vgl. p. 20f. - So ergabe sich ein ganz einheitliches Phonemsystem: lie a o u/ /p t t c k/ /m non n / vr pyl Dazu /aa/ usw. wie /tt/ usw., /ai au/ wie /nt/ usw. e, o. Doch scheint es, dass man aussondern kann und mit /m/ oder /n/ gleichsetzen, vgl. mar@m (p. 25) = mara (p. 26) = mara (p. 28) 'Baum'; av@n = avi 'er'. Dann hatte das Umgangs-Tamil letztlich nur ein Phonem, das das literarische Tamil nicht kennt, namlich /ae/ (da z = aeae ist). 4. Sehr wichtig ist der Abschnitt uber die Betonung, die Andronov genau untersucht hat. Sie ist im wesentlichen frei, lokalisiert, quantitativ-dynamisch. 5. Ebenfalls bedeutsam ist das Kapitel uber Morpho- und Syntaktophonemik, von Andronov "Liaison" genannt. Trotz der meisterhaften Bearbeitung wird man nicht in allen Punkten dem Autor zustimmen. P. 44 z.B. wird angegeben:1 + t = nt; 1 +t= nt;1 + tt = tt; 1 + tt = tt (beim Verb). In anderen Fallen (p. 43) gilt aber 1 + t = tt; 1+t=tt. Diese Doppelentwicklung anzunehmen ist unnotig, wenn man bei den Verben ansetzt 1 + nt = pt; } + nt = pt. Die betreffenden Verben gehoren nicht zur Klasse I, wie Beythan und andere anfuhren, sondern zur Klasse II (vgl. Beythan p. 83), also z.B. kol 'toten', kopteen 'ich totete' aus kol-nt-een, gebildet wie ati 'wissen, atinteen 'ich wusste' (ati-nt-een). 6. "Euphonic sounds and syllables (increments)" werden p. 51 angegeben, namlich a, att, am, att, aNG, u, tt, y, v und andere. Ausser u, y und v handelt es sich aber wohl um besondere ursprunglich bedeutungstragende Morpheme, fur die die Benennung *Zwischenglieder" (Beythan) gut passt. Ausser u, y, v und vielleicht auch a sind sie sicher nicht eu-phonisch. Der Obliquus zu maram 'Baum', maratt- (p. 52), ebenso der sondern eher ein ursprunglich demonstratives, artikelahnliches t(u), vgl. a-t(u) 'es'.

Page #39
REVIEWS 285 Zwischenglieder an den Sandhi im Umgangs Andronov unter *m-t wurde dann -uber *n-t (vielleicht durch Stufenwechsel, der allerdings hier unklar bleibt), zu tt, *t-t zu tt: *maram-t(u) > *maran-t(u) und schliesslich maratt(u), ganz entsprechend der Entwicklung *naat-t(u) > naatt(u). Vortrefflich bei Andronov ist indes, dass die Zwischenglieder alle ubersichtlich zusammengestellt worden sind. Sehr wertvoll sind ferner die Angaben uber den Sandhi im Umgangs-Tamil. 7. Die Morphologie ist meisterhaft und sehr ausfuhrlich behandelt. Andronov unterscheidet zehn Redeteile; neben den ublichen noch personliche Nomina, imitative und Echo-Worter. Beim Nomen bietet u.a. die Tabelle der Nominalbasen pp. 72-74 eine gute Ubersicht. Die Anordnung der Deklination geschieht nach strukturellen Grundsatzen, nicht nach den herkommlichen, die durch die Sanskrit-Grammatik beeinflusst sind. So wird zwischen Komitativ und Instrumental durchweg geschieden (p. 79), und die eigentlich syntaktischen Verbindungen mit utaiya, -iliruntu usw. sind nicht mit aufgenommen worden, was durchaus richtig ist. Einen Hinweis auf diese Formen hatte man allerdings gern gesehen.' 8. Sehr charakteristisch fur das Tamil sind die personlichen Nomina (pp. 122ff.), im Tamil kutippuvinai 'heimliches Verb' genannt, z.B. nallaa! 'sie, die gut ist oder 'eine gute Frau', teevariir 'ihr, die ihr (ahnlich) Gotter(n) seid'. Beythan ubersetzt diese Ausdrucke mit 'sie ist gut' bzw. 'ihr seid Gott' = 'du o Gott' (p. 123; p. 127). Es handelt sich um alte Wortsatze, wie sie auch in Mundasprachen vorkommen. Die personlichen Nomina bilden das Bindeglied zu den Adjektiven, die ursprunglich Nextrum-Formen, 3. Person Pluralis von personlichen Nomina waren, wie Andronov ausfuhrt. 9. Beim Verbum geht Andronov in Ubereinstimmung mit den einheimischen Grammatikern von der Form des Verbalnomens auf (t)tal aus. Dann unterscheidet er bei den Verben jeweils zwei Stamme, von denen der eine suffixlos ist, der andere k oder kk aufweist. Auf diese Weise konnen z.B. die von Beythan und anderen als (k)ki(n)t angesetzten Prasenssuffixe in (k)k + int aufgelost werden. Auch auf die Einteilung der Verben in Klassen verzichtet Andronov auf Grund dieser Interpretation des Materials. Ob diese Auffassung sich durchsetzen und anerkannt werden wird, bleibt abzuwarten. Hilfreich fur den Leser ware es gewesen, wenn der Autor das Problem der Klasseneinteilung wenigstens kurz erortert hatte. Durch die ausfuhrlichen Konjugationsmuster pp. 215-243 wird man indes weitgehend entschadigt. 10. Von hervorragendem Wert ist das Kapitel uber die Syntax. Die Quellenangabe bei den Beispielsatzen ist hier naturlich besonders nutzlich. 11. Die Textubersetzungen sind sehr akkurat durchgefuhrt worden. Eine Ausnahme sei erwahnt: etirikal aranmanai veelaikkaararilee cilarai vacamaakki 'his enemies, having won over the palace servants...' (p. 183). Besser: "Die Feinde, die einige der Palastdiener bestochen hatten". Abschliessend sei betont: Wer sich jetzt mit dem Tamil beschaftigt, wird sicherlich zuerst zu Andronovs hervorragendem Werk greifen. Dem Autor gebuhrt unser aller Dank, dass er uns diese wichtige indische Sprache in ihrer Vielschichtigkeit nahergebracht hat. Heinz-Jurgen Pinnow Current Trends in Linguistics, ed. by Thomas A. Sebeok, Volume 5: Linguistics in South Asia. The Hague-Paris, Mouton, 1969. XVIII + 814 pp. This fifth volume of the well-known series, dedicated to South Asia, shows once more the immense importance of Current Trends in Linguistics. That the master plan could be realized so far is largely due to the generous financial support of three U.S. federal agencies. The period covered by this survey is in general from about 1947 to 1966,

Page #40
286 REVIEWS although a few references to publications of 1968 occur. The large majority of the contributors belong to the generation born in the twenties and the early thirties, the older generation being represented by Renou, Burrow, Clark and Emeneau only. Although there naturally is an undeniable difference in quality and in general character between the various chapters, most of them are of a high standard. For future research this general survey of what has been done and is being done in the field, with frequent indications of the gaps in our knowledge and the most urgent problems that wait for a solution, will be indispensable as a starting point. The extensive bibliographical references will no doubt prove extremely useful. The circumstance that this book is the first to give a comprehensive account of the recent amazing development of Dravidian studies contributes much to its importance. Studies in the field of the Munda languages cannot compare with those in Dravidian because of their smaller number, their cultural background and the amount of work done (or, at least, published). Still, this work, too, has only started in the last few decades and the general survey here given is one of the first to appear in print. The book is divided into four parts. Part One ("Indo-Aryan Languages") consists of fourteen chapters, Part Two ("Dravidian Languages") of six, Part Three ("Other Language Families") of three only, while Part Four ("Linguistics and Related Fields in South Asia") comprises as many as eleven chapters. Only a few of these thirty-four chapters can briefly be mentioned here. In Part One Thomas Burrow, to whom the task of surveying the linguistic publications on Sanskrit was assigned, begins his contribution with the words "The progress of Sanskrit studies in the last twenty years is characterised by no striking new developments." He then gives an almost exhaustive and fair report on publications in the various departments of phonology, morphology, syntax and lexicography. The last category is perhaps treated somewhat more elaborately than might seem necessary in view of the exhaustive and detailed bibliography (pp. 19-35, as long as the article), which gives a most useful enumeration of all the publications in this field. In "Comparative Indo-Aryan" Gordon H. Fairbanks gives in seven pages (pp. 36-45) a historical survey of the earlier stages of comparative studies in Middle and New IndoAryan languages, starting from Beames (1872-1879) via Jules Bloch (1933) to modern developments, but the function of this chapter in the general plan of this book is not sufficiently clear. Ernest Bender's account of Middle Indo-Aryan could be brief: (about three pages, but with a "selected bibliography" of four pages). Out of the eleven chapters on the New Indo-Aryan languages an arbitrary choice must be made. Vladimir Miltner's chapter on Hindi (pp. 55-84) is particularly useful on account of its enumeration of the many Russian contributions, most of which are not easily accessible to scholars working outside the East European bloc. Besides it gives a reasoned account of the numerous Indian publications written in Hindi and other Indian as well as Western languages. In conclusion he quotes K. Ch. Bahl as saying that "the grammatical treatises on Hindi lack any scientific coherence". Kali Charan Bahl's article on Panjabi (pp. 153-200) contains a theoretical treatment of high standard of the verbal system. Incidentally he remarks (p. 175) that "the investigation of compound verbs seems to have reached a dead end" and (p. 181) that "there is no way to understand the nature of modification of the meaning of the main verb by means of a subsidiary verb without first knowing what the meaning of the main verb (i.e. the lexical content in our sense of the term) really is". In his conclusion (p. 195f.), he states that scholars "have never understood the difference between the colloquial and literary aspects of the language. Consequently the scientific investigation of Panjabi is beset with discussions on language vs. dialect, spoken vs. written language, colloquial vs. literary standards, linguistic descriptions vs. nonlinguistic or extralinguistic considerations, and so on... The grammars of the language describe it only superficially.

Page #41
REVIEWS 287 The dictionaries are not only unable to catch up with the numerous literary usages that are coming into the language every day, they have left out a whole lot of lexical stock of tadbhava origin. Historical and comparative studies are based on second hand sources, and dialect studies are almost nonexistent. A serious scientific investigation aiming at sufficiently deep and exhaustive description of the language in any one of these aspects, is yet to begin." As regards M. W. Sugathapala de Silva's chapter on Sinhalese (pp. 235-248) special mention may be made of his account of Vedda and Rodiya. On p. 246f. he states that "Vedda and Rodiya are not dialects of Sinhalese. Vedda is a creole with much Sinhalese influence, and Rodiya is a secret language in which non-Sinhalese lexical items are used in Sinhalese structures. Any dialect survey of Sinhalese should treat these two speeches separately. Sufficient work has already been done on Vedda, but Rodiya needs further investigation. As most "secret" words used by the Rodiyas are very different from Sinhalese or Tamil, it would be interesting to search for the origin of such words." The same is true of Vedda, which according to de Silva is "a creole based on an older Vedda language with Sinhalese as the second contributing factor". Some additional data for Vedda are communicated by Hettiaratchi on p. 743. Both languages would seem particularly important for a correct linguistic analysis of the so-called Munda language Nihali (see below, and cf. Kuiper, Nahali, p. 114, where a paper by Wilhelm Geiger on a "Gaunersprache" in Ceylon is quoted as a parallel to Nahali). Braj B. Kachru, in his very instructive survey of "Kashmiri and other Dardic languages" (pp. 284-306), states (p. 286) that "The question of the final affiliation of the Dardic family of languages has not yet been answered." In his opinion there is not "much authentic linguistic evidence" in support of the views of Grierson or Morgenstierne. This sceptical attitude of a scholar who is intimately acquainted with the material and the problems it raises should be a warning to all those who try, on the basis of the available material, to form an opinion on the historical relations of the Kafir languages to the Indo-Iranian family. As Morgenstierne was the first to point out (NTS, 13 [1945), p. 235) the question may be raised if not certain specific Kafir characteristics point to a very early separation of Kafir from the proto-Indo-Iranian group. (It should be observed that even the notion of a common Indo-Iranian period in Western Asia has been questioned. See Kaj Barr, Illustreret Religionshistorie (redigeret af J. A. Asmussen og J. Laessoe), p. 242, who assumes two successive waves of immigration from South Russia, first of the Proto-Indians and later of the Proto-Iranians. If, on the other hand, archaeologists like H. D. Sankalia are right in dating the arrival of the Aryans in India about 2000 B. C. (e.g., Munshi Indol. Fel. Vol., p. 233) this would put the period of a common culture too far back to account for the many parallels in the poetic diction of the Rigvedic poets and Zarathustra (Bernfried Schlerath, AwestaWorterbuch, Vorarbeiten, II, p. 148ff.). See also Morgenstierne, NTS, 13, pp. 234-238.] As far as Kachru is concerned, it should be kept in mind that he is referring to "authentic linguistic evidence", which in his opinion is lacking, and that according to him "Morgenstierne's work ... leaves much to be desired" (p. 285 n. 2), a judgment that would probably not have surprised Morgenstierne himself in view of the very uncommon circumstances under which he at times had to collect his material. Anyway, Kachru's reservations deserve notice since several Indo-Europeanists have drawn far-going conclusions from some data communicated by Morgenstierne (see the references in IIJ, 10, p. 103 n. 4). Kachru concludes his survey with the statement that "the research in Kashmiri and other Dardic languages has made practically no serious progress in the last two decades". Part Two is devoted to the Dravidian Languages (pp. 309-408). Bhadriraju Krishnamurti reports on "Comparative Dravidian Studies" (pp. 309-333), an indispensable

Page #42
288 REVIEWS survey for all Dravidologists. The Dravidian family is said to consists of "ca. 22 languages spoken by about 110 million people in South Asia". He arrives at this number in spite of the fact that he takes Ollari and Gadaba as different languages. In his historical survey a tribute is rightly paid to L. V. Ramaswami Aiyar. What this scholar has achieved, managing to remain in contact with what was being done outside India in spite of the relative isolation in which it was his fate to work, shows that he was an outstanding and original scholar. Krishnamurti's judgment of the work of that generation is reasonable and fair. He even finds something to praise in E. H. Tuttle's publications. In his excellent survey of recent publications (primarily those by Emeneau and Burrow) occurs his thought-provoking suggestion of reconstructing a ProtoDravidian laryngeal (or h-type of sound) in the demonstrative and interrogative stems *iH-, *uH-, *aH- and *ya-leH- to account for the unexpected vocalism in Kui, Kuvi and Brahui and for h in Kuvi. [As for the glottal stop in Kuvi, Kui (dialectally), Konda and Gondi (dialectally) Burrow and Bhattacharya, IIJ, 6 (1962-63), p. 243, consider it to be "presumably of fairly recent origin", ancient Tamil having anticipated this development by its aytam which, accordingly, is not historically connected with the Central Dravidian developments.] Krishnamurti points to the forms gi for i, ga for a and ge for e in low class Telugu of Telangana, for which he assumes *H > 8, parallel to Old Tamil eHku 'steel' > colloquial mod. Tamil eggi (p. 320). As his notation eHku shows, Krishnamurti identifies the Old Tamil aytam with his reconstructed PDr. phoneme H/. He, however, overlooks the fact that the distribution of the OTa. aytam is entirely different from that of his reconstructed /H/. Mod.Ta. eggi with its curious intervocalic [g:] may be hard to explain in detail but there can be no doubt that it represents a particular development of /k/, which may or may not have been conditioned by the preceding aytam. Since the aytam must have disappeared from living speech at a very early period, it is not easy to imagine a straight-line development from OTa. ehku to eggi. The Tranquebar Dictionary gives eku by the side of ehku (instead of *ekku, which would be the normal representation). However that may be, the colloquial form eggi gan hardly prove anything in this connection, no matter how OTa. ehku, without cognates in Dravidian, can be explained. On the other hand, demonstrative pronouns in general constitute a particular category in that, as a result of their function, they are open to various kinds of emphatical reinforcements. Cf., e.g., Sanskrit a + sau, Latin *hod + ce > hoc, Old French (ecce + hoc >) co, mod. French ce, French celui > celui-ci, celui-la, Boer-Dutch hier-die 'this' (the old demonstrative die 'that' having taken over the function of the article de, itself a demonstrative in origin), etc. So far the theory of a Dravidian laryngeal (cf. L. V. Ramaswami Aiyar's article in Indian Antiquary, 59[1930), p. 197ff.) would not seem to rest on a solid foundation. With the aytam, which only occurs in long stops (cf. Ta. ehku- 'to pull with fingers (as cotton)': Ma. ekkuka-'to card cotton', Ka. ekku- 'to dress cotton, card wool) it has certainly nothing to do. In the discussion of the personal pronouns Krishnamurti refers to his explanation of *nam as standing for n + yam 'you and we' (p. 321), which he has since elaborated in the Emeneau-Festschrift, p. 194. Such a dvandva-compound of two pronouns, however, would be a unique phenomenon in Dravidian word-composition. In the sub-grouping of South-Dravidian Krishnamurti suggests an improvement on Emeneau's diagram in assuming that Kodagu split off from the Tamil (-Malayalam) branch at a later date than Toda and Kota (p. 326). As for Malayalam, the traditional view that it has split off in the Middle Tamil period (a view still represented by Krishnamurti's diagram on p. 327) is only a very rough approximation of the real historical process. See A. Govindankutty, IIJ, 14, pp. 52-60. As for the genetic relationship of Dravidian with other language-families, Krishnamurti's criticisms of some arguments of Altaic scholars (p. 328) are fully justified. "At the same time", he adds on p. 329, "it must be admitted that there is a fairly large

Page #43
REVIEWS 289 amount of morphological and lexical evidence to show that there are good chances of some original relationship having existed between these two families." Although the exact purport of these words is not quite clear, since he here confounds Uralian, Altaic and Ural-Altaic, one may readily agree that, if it will be possible to demonstrate any relationship with another family, this will indeed most likely be with Uralian. [Incidentally, if a relationship can be proved, this is a relationship tout court, not an original one which has existed.) Attempts to connect Dravidian with Caucasian, Korean, Egyptian or Sumerian are rightly passed by in silence (p. 329). The chapter on "The non-literary Dravidian languages" by Murray B. Emeneau (pp. 334-342) will remain of lasting historical interest because it describes in detail the various stages of a fascinating period in the development of Dravidian studies by a scholar who has himself taken a prominent part in the process of detecting and describing the so-called 'tribal languages'. Although the collection of data began as early as 1789 (for Malto), the epic period of exploration is exactly the one covered by this book, that is, the twenty years from 1947 onwards. As Emeneau points out, "The American blank period in which descriptive workers were ideologically opposed to comparative work, does not have its counterpart in Dravidian studies." So most of the recent work has been both descriptive and comparative. The full account of the effort to identify languages which were often known by more than one name is fascinating. A well-known instance is the Gadba tribe, the central section of which speaks the Munda language Gutob (= Gadaba in Grierson's Linguistic Survey), while the marginal section speaks a Dravidian language, now called Gadba (with the dialect Ollari). The story of the disentanglement of the intricate problem of Naiki (= Naiki, LSI.) and Naiksi (= "Bhili of Basim", LSI.) is told on p. 340 (see also Krishnamurti, p. 309 n. 1). At the end Emeneau considers the question whether all the languages of South Dravidian have been discovered. Gerard Diffloth strikes Irula from the list, but thinks that the Betta Kurumba may possibly speak an independent language. (On the other hand Kamil Zvelebil has since claimed for Irula the status of a separate language.) Emeneau further points to D. N. Shankara Bhatt who, in the Emeneau-Festschrift, describes Koraga as "a new Dravidian language" (See also the same, Linguistic Survey Bulletin, 7, 8, 9]. It may be a long time yet before a general agreement on the exact number of Dravidian languages has been reached. Emeneau enumerates twenty-one as belonging "with certainty" to this family (p. 334) but more fieldwork is needed before in all cases the question "independent language or dialect?" will have been answered. The first of the chapters on separate literary languages, naturally dedicated to Tamil, has been written by Kamil Zvelebil (pp. 343-371, including a bibliography of eight pages). He characterizes the years between 1946 and 1966 as "the era of Vorarbeiten" (p. 344 f.). In this connection it may be observed that among the universities where attention is being paid to Tamil (pp. 344, 362) the Dutch ones might have been mentioned along with the German, Japanese and Polish ones. The phonemic inventory appears to range in the various descriptions from 11 to 44 consonant phonemes. Incidentally, in the diagram of Bright and Ramanujan's inventory (p. 349 n. 9) S and s have obviously changed places. In the survey of synchronic and diachronic studies Zvelebil suggests a new explanation for the still enigmatic present tense morpheme kinr(p. 354 n. 24), viz. connection with kil- 'to be able'. The older theories are passed by in silence. Dialectology, Sociolinguistics, Syntax, Lexicology and Applied Linguistics are discussed under separate headings. In Vadasery I. Subramoniam's account of Malayalam (pp. 372-381) the achievements (since 1947) in the field of lexicography are stressed. Work on the history of Malayalam has been dominated by the question as to whether Malayalam is a dialect of Tamil or an independent offshoot of proto-Dravidian - a question that could only be raised by non-linguists. Much important work on the dialects is buried in unpublished theses of Kerala University. On p. 377 Subramoniam states that "Malayalam is a collection

Page #44
290 REVIEWS of dialects which are heterogeneous in nature". This is a problem that may perhaps be of great importance for our reconstruction of the history of the West Coast dialects but which still waits further exploration. One would have liked to hear somewhat more about the relation between the standard language and the dialects (Subramoniam refers on p. 377 to "the standard forms of Malayalam" but the question as to what exactly is a standard form does not seem to be hotly disputed), and particularly about "the tribal languages of Kerala, such as the language of the Kaadars of Cochin and the Muthuvans", mentioned as being different from the dialects (p. 378). This laconic statement raises questions which remain unanswered. In conclusion, Subramoniam states that "the inadequacy of existing theoretical models is occasionally felt when they are applied to Malayalam" and expresses the hope that the younger generation may be able "to formulate an independent and comprehensive model or an alteration of the existing models for language analysis". At the end of a comparatively brief chapter on Telugu (pp. 382-393) George Kelley states that "relatively little of value on Telugu has appeared in recent years, except for the extensive work of Krishnamurti... Progress has been slow because few scholars are at work in the field." He gives a clear description of the three regional dialects and the problem of the social variants (formal and informal styles among educated speakers, educated and uneducated varieties of speech). A Telugu dialect Dictionary of occupational terms is mentioned as a unique kind of lexicography in India (but what about Grierson's pioneering work Bihar Peasant Life?). Still shorter is H. S. Biligiri's chapter on Kannada: about five pages text, with a bibliography of nine pages. ary facilities. Onon was stolen in Inde specially to be Part Three. "Other language Families" (pp. 411-477) deals with Austroasiatic. TibetoBurman and Iranian languages. The chapter on "Munda and non-Munda Austroasiatic languages" by Norman H. Zide (pp. 411-430) deals with Munda, Nihali, Nicobar and Khasi. Since this is the first time (apart from Pinnow's older Kharia-Sprache) that a survey of the whole field (with references to recent, mostly unpublished papers) is given, it is especially to be regretted that the original version of this contribution was stolen in India. The present version was written without library facilities. One may marvel that the final result is still better than one might have expected. This is largely due to David Stampe's help in providing bibliographical references. Still it is disappointing that this first account of the Munda studies should be so unsatisfactory and defective. Sudhibhusan Bhattacharya's brief survey "The Munda Languages and South-East Asia" (Bulletin, Indian Institute of Advanced Studies, Simla, July 1970, pp. 23-31) does not complement it although it may be consulted alongside with it. After giving a Stammbaum diagram (in which Central Munda is grouped with South Munda, see Emeneau-Festschrift, pp. 370-377) and referring to Pinnow's studies on the relationship with other Austroasiatic languages (which is now, at last, generally accepted), Zide gives a reconstruction of the proto-Munda consonant system. As for the PM. morphology, Proto-Munda is said to have had "predominantly monosyllabic morphemes". Much depends on the extent of "predominantly". There are so many disyllabic root morphemes which Munda has in common with other Austroasiatic and even Austronesian languages (some of which may be found in Orientalia Neerlandica, p. 376f.) that it might be preferable to ascribe the predominantly monosyllabic root structure to an earlier historical stage than proto-Munda. The following sub-chapters deal exhaustively with phonology, morphology and syntax, lexicology (and semantics), etymology (with some well-founded criticisms of some explanations of Sanskrit words from Munda), vernacular literature and miscellaneous topics. In addition to the numerous unpublished American papers H. J. Pinnow's important monograph Grundzuge einer Phonetik des Mundari, Berlin, 1954 (cited in Kharia-Sprache, p. 460) should have been mentioned.

Page #45
REVIEWS 291 On p. 427 Zide devotes half a page to this reviewer's paper on "Consonant variation in Munda", which cannot be passed by without some comment. The passage begins with the words: "Kuiper's paper on consonant variation in Munda has a typological burden and follows up earlier work of his own and of other Dutch scholars, e.g. Gonda on Austronesian. He questions the applicability of the comparative method ... so that clean-cut reconstructions cannot be made, as they presumably can be made for IndoEuropean. Kuiper provides a large corpus of variant forms in Munda, and in discussing them rejects such standard explanations for the variation as dialect mixture. In my opinion the variants he cites probably come from a variety of sources: conditioned changes (e.g. consonant cluster simplification), dialect mixture, and expressive derivation..." With regard to this passage I should like to make the following comments: (1) the main object of the paper, which was presented to a Conference on Indo-Pacific linguistics, was simply to point out the existence of a problem that deserves to be studied because it constitutes a serious difficulty in comparative studies. The explanation of expressive deformation had been adopted (as was clearly stated) from Pinnow. Still, the main purport of the paper was not to propose an explanation but to present material, a point that Zide apparently missed. The limited value of material culled from dictionaries and second-rate grammars was obvious. (2) Whether or not one chooses to call the paper typological, anyway, it cannot be said to follow up earlier work of my own since I clearly dissociated myself from my earlier publications as far as the interpretation of the material was concerned. Nor am I aware of any connection between this paper and studies of other Dutch scholars". The paper does definitely not reflect a national habit of advocating unorthodox views. (3) I did definitely not question the applicability of the comparative method in general. On the contrary, it was pointed out that for a large portion of the vocabulary this method does hold good but that only a certain sector of it had to be put apart, a fact that had been fully recognized by Pinnow, p. 21. (4) Zide's words to the effect that clear-cut reconstructions "presumably can be made for Indo-European" are only intelligible to me if they express a doubt on his part as to the applicability of the classical method of historical reconstruction even to Indo-European. It is true that, ever since the nineteenth century, some scholars have been aware of the fact that etymological word-studies in Indo-European often present more intricate problems of phonemic correspondences than might be guessed from the exposition of these correspondences in the current handbooks. (See, e.g., the discussion, with references, in Mnemes charin, Gedenkschrift Paul Kretschmer, I (1956), p. 222f.). The fact that Indo-European comparative linguistics has nevertheless become the model for comparative studies in general is probably to be attributed to the circumstance that the reconstructed proto-Indo-European was the language of a rigidly organized "Herrenschicht" and as such, for sociological reasons, different from the language of jungle tribes. G. Fortune's lecture on Ideophones in Shona (London, 1962) is revealing for the kind of transformation, words can undergo in certain societies, and more particularly for the circumstances and the way in which this may happen. For these phenomena in the Austroasiatic area see, besides Pinnow, Kharia-Sprache, p. 20 ff., also J. A. Gorgonjev, Grammatika khmerskogo jazyka (Moskva, 1966), p. 69ff. and David Stampe, IJAL, 32 (1966), p. 397. The term "expressive" is only a rough indication of this phenomenon. (5) On p. 414 Zide reports on Pinnow's listing "the regular and semi-regular correspondences", thereby implicitly accepting the inapplicability of the traditional comparative method to part of the lexical material -- a fact which Pinnow had correctly noted. By accepting the existence of "semi-regular correspondences", however, one has returned to the nineteenth-century theory of sporadic sound-laws, which amounts to denying the possibility of a scientific treatment of the linguistic material. (6) In the paper under discussion it had been pointed out that an attempt to explain the "variation" by means of the "standard explanation" of dialectmixture had led to the necessity of assuming an incredibly high number of strongly

Page #46
292 REVIEWS divergent dialects with bizarre criss-cross borrowings in e.g. Santali for which so far no essential dialectal split has been reported. If one nevertheless prefers to stick to the "standard explanation" it should be noted that for such cases as Mu. Ho buti : So. pudi 'navel such an explanation lacks any foundation in the facts. Consonant cluster simplification would presuppose for proto-Munda a word-structure entirely different from, e.g., the Austronesian type, for which I fail to see any indication. So there remains, for the present moment, as the most promising explanation the "expressive derivation", the very explanation proposed by Pinnow and accepted in the paper under discussion, I am afraid that Zide's report will fail to give the reader an insight into the real problem while creating false impressions about the extent to which dialect borrowing can be applied. On the other hand, his experience as a field-worker with the difficulty of distinguishing between p:k and b:g in Gutob is the very kind of information that we need. The problem of Nihali (Kalto), discussed on pp. 427-428, is one of the most intriguing. After Konow and Grierson had incorrectly assigned it to Munda, Robert Shafer was the first to point out unidentifiable elements in it (1940: Hary, J. As. Studies, 5, pp. 346-371, cf. 1954: Ethnography of Ancient India, pp. 10-12). After the publication of fresh and more reliable material by Bhattacharya in 1957 (Ind.Ling., 17, pp. 245-258), Burrow referred to Nahali in the next year as a remainder of pre-Dravidian and preMunda (1958: Bull. Ramakrishna Mission Institute of Culture, no. 19, p. 8). Pinnow, while making several attempts to explain it as a Munda language, was for a long time non-committal (1959: Kharia-Sprache, p. 1 n. 6 vs. p. 45 etc.; 1960a: published in Studies in Comparative Austroasiatic Linguistics (1966), pp. 187-191; 19606: IIJ, 4, p. 86 n. 23). In 1964, however, he was more outspoken in Linguistic Comparison in SouthEast Asia and the Pacific (p. 151: "It is at any rate not Munda"; p. 152: "We may perhaps come closest to the truth if we assume that Nahali possesses an isolated nonAustroasiatic substratum that has been partially replaced by an Austroasiatic stratum which has also provided Nahali with its inflection." These words were clearly written independently of this reviewer's monograph, published in 1962, consisting of an analysis of the grammatical system and the available lexicographical material (about 500 words!), in which he had tried to prove that Nahali probably was an argot, the two oldest layers of the vocabulary being an unidentifiable language (ca. 24 per cent of the vocabulary) and (in a few per cent) an Austroasiatic language which cannot, in the present state of our knowledge, with certainty be called Munda. In 1966 David Stampe reported in IJAL, 32, p. 395, on a Nihali lexicon by Aasha Kelkar Mundlay (still unpublished), comprising some 2000 items, of which sixty to seventy per cent are recent borrowings from Indo-Aryan or Korku. "Of the rest, a few have older Dravidian or Munda sources, but most are not identifiable." This more extensive material, accordingly, confirmed entirely the conclusions of the earlier monograph. The latter, however, failed to convince Pinnow, who in 1965 (Indo-pacific Linguistic Studies, I = Lingua, 14) on the one hand characterized Nahali as "a language the status of which is still much disputed" (p. 4) but on the other hand concluded that "the personal pronouns of the disputed language Nahali can also be classified with those of the Austroasiatic family, even though they are rather markedly distinguished from the personal pronouns of the other groups", adding that Nahali and Nicobarese may possibly "be more closely connected than was hitherto thought to be the case" (p. 18). Finally, in 1966, he ended up a review of the monograph on Nahali with the conclusion: "Der grundlegend austroasiatische Charakter des Nahali schalt sich so nach und nach immer mehr heraus" (OLZ, 61, col. 496). In the present state the central problem would seem to be that of the origin of the few Austroasiatic elements in the vocabulary. It is on their occurrence that Pinnow founds his right to attempt to connect the Nahali verbal (1960) and pronominal (1965) systems with those of Austroasiatic. My provisional attempt at an analysis of the case-endings and the pronouns did not confirm this

Page #47
REVIEWS 293 assumption of an Austroasiatic provenance. In that case the lexical elements can only be very early borrowings from a source whose nature it remains to determine more exactly: Old Munda or para-Munda? Nih. te(')- 'to eat', tu- 'to embrace' can, it is true, have been taken from a Munda language now lost, but isolated cases like Dhimal unku 'cauli, rice' (if correct, see Nahali, p. 51) are a warning against rash conclusions. In the present state of our knowledge it would seem essential to recognize that there are some questions which cannot be answered until much more material is available. These facts have here been summarized because Zide, for the reason mentioned above, was unable to give a coherent account of this problem. Of the monograph of 1962 it is only said that the author "demonstrated that Nihali has borrowed heavily from IndoAryan and especially Dravidian, but he did not suggest that Nihali is Indo-Aryan or Dravidian". The last words are not singularly to the point because no one has ever thought of ascribing Nihali to either of these families. The only thing Zide could do was to state that he had long considered Nihali a Munda language but that he had changed his mind. His observations contain nothing new except the confirmation that Nihali is actually an argot, as had been suggested in 1962. There is obviously no point in criticizing Zide. His survey is a silent warning to all those who undertake a journey with a manuscript in their brief-case and omit to leave a carbon-typed or xeroxed copy back home. Roy Andrew Miller wrote the chapter on "The Tibeto-Burman Languages of South Asia" (pp. 431-449) and D. N. Mackenzie that on "Iranian Languages" (pp. 450-477), the latter not limited to India as it also comprises Old Iranian and Middle Iranian. So much of this interesting account is somewhat out of place in this book. In Part Four, "Linguistics and Related Fields in South Asia" (pp. 481-752), there is, first (pp. 481-498), Renou's masterly survey on Panini ("who ... is less interested in describing than in characterizing and analyzing", p. 493) and J. F. Staal's contribution "Sanskrit Philosophy and Language" (pp. 499-531). Of general interest for every linguist is "General Linguistics in South Asia" by Ashok R. Kelkar (pp. 532-542), a fascinating picture of the present situation in India. To his final critical remarks (p. 538ff.) every linguist will no doubt subscribe. He ends up with the words: "We have a long way to go. We can begin (...) by acquiring what we have inherited -- from our own past as well as from the rest of the world." It is encouraging to see that the generation of Kelkar, V. I. Subramoniam and others is well aware of "the Indian penchant for synthesis or desire to displease nobody" (which "often results in cheerful conflation of disparate or even conflicting elements"), of the "lackadaisical" editing of journals, and the difficulties arising from "the stratified power structure of lhe South Asian academic community" (p. 539f.). "Toward a phonological Typology of the Indian linguistic area" by A. K. Ramanujan and Colin Masica (pp. 543-577) is the most thorough attempt so far to test the areal theory for the whole subcontinent. This important study, which contains nine maps of India with various isoglosses, leads to the conclusion that "there appears to be, thus, a number of multi-familial Sprachbunds in the Indian area" (p. 577). Some doubts, however, may be expressed regarding the idea that diffusion is due to the "prestige of one community over another as the chief social factor causing adoption of traits and bilingualism" (p. 543f.), even if "prestige" is broadly interpreted as "importance to the borrower". Some other factors must sometimes be involved in the process. If the theory of a prehistoric diffusion of Dravidian traits in proto-Indo-Aryan is correct, other possible factors instead of prestige must be taken into consideration (in addition to the assumption of a social stratification in the pre-Vedic society which is only very feebly and indirectly reflected in the priestly literature). The Malay loan-words formerly used in "colonial Dutch" are also an interesting example of linguistic reflexes of a social symbiosis where prestige is inadequate as an explanation, mber of multi-famion regarding the idea factor causing acope.si

Page #48
294 REVIEWS On p. 620 Nur Yalman's criticisms of Dumont's views on kinship terminology should be noted. "Official Language Problems and Policies in South Asia" by Jyotirindra Das Gupta (pp. 578-596) will probably be of lasting importance as a detailed historical record of the development of language policy in India, of the conflicts between politics and linguistic realities and the defeat of Hindi as the national language. His conclusion is: "The complexity of the problem of official language in South Asia admits of no easy solution or judgment." Of the other papers I only mention "English in South Asia" by Braj B. Kachru (pp. 627-678), a masterly synthesis of a fascinating subject. Some more illustrative material might have enlivened the exposition (e.g., "your worthy self" for the "polite diction", listed without examples on p. 651) but the author has rightly concentrated upon the theoretical aspects. The same subject has since been dealt with by H. J. Vermeer in a monograph. The book ends with three chapters on linguistic studies in Pakistan, Ceylon I (Sinhalese) and Ceylon II (Tamil). D. E. Hettiraratchi gives some additional information about the Veddah speech (p. 746f.). The last chapter is one of the weakest. Zvelebil's article on Ceylon Tamil of 1966 (IIJ, 9, pp. 113-138), the most extensive study so far, was apparently not yet accessible to the author but in the sub-chapter"Tamil and ProtoDravidian" some words should at least have been said about the intriguing problem of the historical relations between Ceylon Tamil and "West-Coast Tamil" (= the western Tamil dialects and Malayalam). Instead the author expatiates on his own theory of a relationship between Dravidian and Sumerian. All in all, a book extremely rich in information on all fields of South Asian linguistics, and an indispensable tool for every linguist specializing in New Indo-Aryan or NonAryan languages. The number of misprints and misspellings is excusable in a book of this size and nature but higher than one would expect. F. B. J. Kuiper