The Avestan Vowels.
Michiel de Vaan's book, based on his 2002 Leiden doctoral thesis, is not for the faint-hearted Avestologist. It presents and analyzes a huge body of material collected from the Titus electronic text edition of the Avesta (checked against the printed editions), and it tackles a whole gamut of problems concerning Avestan diachronic phonology.
One of the book's virtues is its layout: each chapter is devoted to a related group of vowels, and concludes with a summary of the findings. A useful final chapter highlights the most important conclusions, provides an overview of the different chronological stages of development, and tabulates all the possible diachronic interpretations for the fifteen vowel signs and their combinations that are employed in the manuscripts. An appendix of proposed corrections to Geldner's edition is followed by a set of indices, which makes it easy to locate the discussion of any Avestan word. (1) It will be used for reference, and there are shocks for the Indo-Europeanist, such as the doubts cast on the long vowel of [TEXT NOT REPRODUCIBLE IN ASCII] 'liver' (pp. 68-69).
Perhaps the most interesting aspect of this study is the attempt to work out a relative chronology for the changes that resulted in the spellings of the archetype manuscript and its much later descendants, which provide our only source for the Avestan language. De Vaan concludes that Old Avestan (OAv.) had a system of only six simple vowel (i, i, a, a, u, u), and that the expansion of the inherited Indo-Iranian system belongs to the Late Young Avestan (YAv.) period (with an eleven-vowel system: i, i, e, [TEXT NOT REPRODUCIBLE IN ASCII], a, a, a (= a), a, o, u, u). However, many of the developments that give Avestan a very different appearance from Old Persian (not only i-epenthesis, u-epenthesis, etc., but also, for instance, *-hrt- > -s-, *-an, *-am > -an, -am) are situated in the "Post YAv." period leading up to the archetype, that is, during the centuries of transmission when Avestan was no longer a living language. The introduction presents the view that YAv. merely represents a later form of OAv., with morphological simplifications, but nothing that cannot represent a direct continuation of the same Avestan dialect, and this is reaffirmed with additional evidence from vowel developments in the conclusion. Within YAv. apparent inconsistencies in the reflexes of the inherited vowels, which have often been accounted for via dialect variation within the YAv. corpus, (2) are explained as conditioned variants or via analogies at different stages of the tradition. For example, *hu- regularly yields Av [x.sup.v]-, but the variant realization huu- is said to result from the restoration of hu- 'good' in compounds, and also from conditioning by a following -u- (pp. 565-68).
De Vaan thus follows in the footsteps of Beekes and Benveniste (3) in reconstructing something close to the Proto-Iranian vowel system for OAv. But in much of the detail he is very original, and his study represents a thorough and bold reappraisal of the whole subject of Avestan vocalism. Sometimes he seems over-eager to jettison results from the last century of scholarship that point to links between Avestan and Middle East Iranian languages. For instance, YAv. has been considered to show shortening of inherited long *-a- in the sequences *-ayV-, -avV- (the "Sogdian shortening" noted by Tedesco, Henning, and others), but de Vaan doubts the validity of such a sound change because there are counter-examples, and because he prefers to invoke morphological analogies to explain the inflections -aiia, -aiia, etc., of feminine stems in -a-, and -auuant- versus Sanskrit -avant- in pronominal adjectives (pp. 123-24). However, the cases that are left, although small in number, consist of isolated inherited lexemes where such a phonological shortening provides a single straightforward explanation. De Vaan admits "shortening at some stage" (p. 120) for asaiia 'without shade' (Skt. acchaya-), but for Henning's other example, nauuaza- 'boatman' (Skt. navaja-), he invokes dissimilation of a ... a in successive syllables (for which there are also counter examples, p. 137). Then the short vowel of raiia 'by wealth' (Skt. raya) is ascribed to paradigmatic analogy with rayim, etc., and that of vaiiu- 'wind' (Skt. vayu-) to remodelling on the basis of the verbal stem vaiia- "to blow" (p. 122). Such cases illustrate the sort of ad hoc explaining that occupies pages and pages of this long book. One of the frustrations of Avestan is that sound laws often cannot be shown to be exceptionless, and ad hoc explanations cannot always be avoided, but it is a question of when and for what is ad hoc explaining appropriate? If most changes to do with vowel quantities can be shown to belong to the post-YAv. stage, this should not exclude the possibility that the remnants of real linguistic developments that belonged to the living language may be preserved, albeit incompletely, because they have been overlaid by other changes of a different nature from the period of transmission.
What were the causes, the nature, and the scope of the post-Avestan changes? Such questions could with profit have received some general discussion, especially in view of the interesting hypothesis (pp. 446-50) that the nom. singulars in [TEXT NOT REPRODUCIBLE IN ASCII] in late Yasts may be attributed to the pre-Middle Persian vernacular of the compilers. De Vaan appears to assume like other scholars that the OAv. compositions were chanted in a different way from those in YAv., but nevertheless "blind" changes affected the Avestan tradition as a whole. With regard to the period when YAv. was a living language, he builds on the work of Narten and Beekes to present a detailed picture of how YAv. vocalism interfered with that of the now canonized OAv. texts. The speakers of Early YAv., it is argued, replaced OAv. phonemes by their YAv. allophones (e.g., -aha- by [TEXT NOT REPRODUCIBLE IN ASCII], -ahm- by [TEXT NOT REPRODUCIBLE IN ASCII]) in cases where they had no exact equivalent in their own language, but then by a stage of Late YAv. such replacements ceased.
For the most recent part of the Avestan tradition de Vaan has also done some very useful work, collating the readings of several important individual manuscripts, and he has brought to light some striking facts such as the variation between -im and -im (pp. 265-66) in different parts of F1, a manuscript that also in other respects turns out to be very idiosyncratic in its spelling (cf., e.g., the possibly old spellings [TEXT NOT REPRODUCIBLE IN ASCII] in place of normal -arst-, pp. 528-29).
Almost every form which is cited in the book receives some discussion, and this is mostly useful, but de Vaan's arguments need to be examined critically, particularly when he is using opinions expressed by other scholars as a basis. One problem is that YAv. emerges from the discussion of some phonological questions as a language with a penchant for aberrant morphological innovations.
For instance, it is argued (pp. 144-46, 411-12) that presents in -aiia- and -iia- shortened the long vowel *-a- in the inherited first persons indicative and the first singular subjunctive inflectional endings (*-ami, -amahi, etc., subj. *-ani) as a result of paradigmatic levelling where an alternation *a / a was inherited. Quite apart from the difficulty that the inherited thematic subjunctive showed no such alternation but had a long vowel throughout, it is hard to believe that only these two types of thematic stem should have had their inflections remodelled in this way, while the other thematic classes show no such innovation. A phonological solution or solutions associated with the preceding yod seems more likely. (In the case of *-yami, *-yani > -iiemi, -iieni it is easier to believe, following Morgenstierne, that "i-mutation" did in fact apply to the original long vowel -a-; the third singular subjunctive forms in -aiti, -aite are not valid counter evidence, as in the first person forms the nasal could also have contributed to the narrowing of the preceding vowel.) Likewise the excellent chapter on the diphthongs falters via its suggestion (p. 349) that thematic 2nd dual indicative middle [TEXT NOT REPRODUCIBLE IN ASCII]--which regularly functions in YAv. as a 3rd dual inflection!--must owe its spelling -oi- to the analogy of 2nd pl. optative middle [TEXT NOT REPRODUCIBLE IN ASCII].
In spite of the sort of problems indicated above, it must be emphasized that many points of Avestan vocalism and their relative chronology have received convincing clarification, and above all a full collection of material has been made available. In particular, mention should be made of de Vaan's discussion of the sequences -aur-, -aour- and -aoir- (pp. 419-24), and his findings about the role of open and closed syllables in determining vocalic developments. The demonstration that the quantity of Avestan i and u normally reflects the original quantities, if allowance is made for a number of late conditioned changes, shows just how worthwhile it is to reinvestigate every aspect of this language's vowel system without any preconceptions. Whether most of de Vaan's findings gain general acceptance or not, his work should ensure that the debate on all aspects of Avestan vocalism is reopened, and that opinions based on insufficient evidence are revised.
1. The index of Avestan lists individual words, and so any form encountered in a text can be looked up immediately; but it does not follow the usual lexicographical principle of arranging forms under a lemma (verb roots, stems of nominal forms, etc.). In order to find, for instance, the series of verb forms from man- 'to think', the reader must know to look up mainiia, manta, [TEXT NOT REPRODUCIBLE IN ASCII], etc., which appear separated by unrelated words in the index.
2. Explanations of this sort have been frequently used by K. Hoffmann and his pupils, but many opinions of the "Erlangen school" that have achieved the status of dogma particularly via their inclusion in K. Hoffmann and B. Forssman, Avestische Laut- und Flexionslehre (Innsbruck, 1996) are courageously questioned by de Vaan.
3. Presumably because of an oversight, E. Benveniste, "Le systeme phonologique de l'iranien ancien," BSL 37 (1968): 53-64 is omitted from the very comprehensive and useful bibliography. Another omission is N. Oettinger, "Zur Entwicklung der ia- Prasentien im Jungavestischen," MSS 43 (1984): 175-81.
|Printer friendly Cite/link Email Feedback|
|Publication:||The Journal of the American Oriental Society|
|Article Type:||Book review|
|Date:||Oct 1, 2004|
|Previous Article:||The Mahavadanasutra: A New Edition Based on Manuscripts Discovered in Northern Turkestan.|
|Next Article:||A Journey in the World of the Tantras.|