Criticism and bibliography

Frankfurt am Main: Peter Lang, 2003. 107 p. (Europaische Hochschulschrieften. Reihe XXI: Linquistik. Bd. 255)*

© 2004

The reviewed work is the result of many years of research by R. Abduvaliyeva devoted to the study of sandhi in Sanskrit and their phonological and comparative historical interpretation. The author is primarily interested in external sandhi, which act at the junctions of words and at the end of phrases, as they can shed light on the solution of the problem of the end of a word in the Indo-European proto-language. This problem, according to R. Abduvalieva, is extremely important from the point of view of the characteristics of the general sound device of the proto-language, as well as in connection with the establishment of its syllable-morphemic character. In recent years, no research has been conducted on this topic, although comparative studies have a considerable amount of data on the end of a word in certain Indo-European languages. It seems that the numerous and diverse Ancient Indian sandhas, which can be used as a rather important source for reconstructing the end of a word in a common Indo-European language, are not sufficiently taken into account.

The special significance of this research is that the author relies not only on the classical works of European and Indian linguists of the XIX-XX centuries, but also on a large number of unpublished (handwritten) ancient Indian treatises, which are inaccessible to a wide range of indologists and linguists.

The reviewed paper aims to study sandhi from the point of view of modern phonological theories. At the same time, according to R. Abduvaliyeva, different types of sandhi in the ancient Indian language material are not a single system, which is determined by the structure of the language: they are represented both in texts and in grammatical descriptions of the Vedic language and Sanskrit by various modifications and variants related to different historical layers and epochs. Each type of sandhi is considered from several points of view: first, the sandhi rules are formulated, the terms associated with them are introduced and explained (as they are given in ancient Indian treatises and later European and Indian writings); then, all possible exceptions to the generally accepted rules are analyzed in detail to find out whether they are synchronous variants or archaisms, if possible an attempt is being made to open it the reasons for the appearance of the studied variant; in all cases, attention is paid to the historical aspect and, if possible, the relative chronology of different types of ancient Indian sandhi is established, their evolution in different historical periods-from common Indo-European to Indo-Iranian and then to Indo-Aryan. A kind of result of consideration of all types of sandhi is their phonological interpretation: sandhi cases are treated as varieties of neutralization (in the understanding of N. S. Trubetskoy and based on his classification of neutralization and types of phonemic oppositions); binary phonological oppositions are distinguished, and on their basis - differential features of phonemes. On this basis, an attempt is made to reconstruct the phonological system of Sanskrit, taking into account the data of ancient Indian sandhi.

I would like to emphasize that R. Abduvaliyeva's research was conducted on the basis of literary monuments. Fragments from the Rig Veda and the Atharvaveda were used in this work.

* R. Abduvalieva. The phenomenon of sandhi in Sanskrit. Frankfurt am Main: Peter Lang, 2003. 107 p. (Higher School publications. Series XXI: Linguistics, vol. 255).

page 184

the third book of the Mahabharata - " Forest Chapters "("The Story of Savitri"); fables from" Panchatantra"; episodes from Kalidasa's drama"Shakuntala". R. Abduvalieva supports the most complex cases of Sandhi with examples from Indian and European descriptions of Sanskrit, as well as from well - known textbooks; archaisms and non-standard variants-with examples from Atharva-pratisakhya, belonging to both its author and subsequent commentators. The choice fell on the Atharva-pratisakhya, since W. Whitney, when publishing it, brought correspondences to each sutra from other pratisakhyas, taking into account previously published treatises. Thus, he created for researchers a fairly complete and complete picture of the functioning and development of the ancient Indian phonetic system.

In the first chapter, the author sets out to describe the sandhi of vowels in the ancient Indian language as a dynamic phenomenon. Solving this problem will help to find out which types of sandhi in the above-mentioned monuments are new formations, which ones gradually lost their relevance and disappeared during this era, and which ones can be attributed to earlier historical periods. According to R. Abduvaliyeva, the coverage of these issues will serve as a basis for an adequate reconstruction of vocalism in the proto-language. At the same time, it takes into account not only Indo-Aryan data, but also data from ancient Iranian languages, takes into account a number of provisions of Indo-European comparative historical linguistics, as well as data from modern phonology and typology.

In my opinion, the following table is extremely useful and very interesting. on page 36) classification of vocal sandhi, belonging to ancient Indian phonetists. They identified four main types or groups of sandhi vowels: abhinihita, praslista, ksaipra, udgraha.

The first type includes the sandhi of the final-e and-o before the initial short a-of the next word. The result, as you know, will be the loss of the named initial vowel, while in writing, the zero sound will be transmitted using an auxiliary sign-avagrahi. A different result is given by combining the final-e and-o with all the other initial vowels of the next word, so they fall into another group and together with all possible sandhi of final diphthongs make up the fourth type - udgraha sandhi. The behavior of the final-a (both short and long) in all combinations with other initial vowels is regulated by the praslista sandhi rules, which also include combinations of final -/ with initial /-, final-and with initial and -, and final syllabic-r with initial syllabic r - (regardless of whether the initial vowels are combined with other initial vowels). they are long or short). The ksaipra sandhi group combines all other cases of changes in the final -/, -and and syllabic-r. Such a systematization of vocal sandhi has not previously been introduced in Sanskrit textbooks, although it could be useful for anyone who studies it.

In the first chapter, all four types of sandhi vowels are discussed in great detail, their nature, various modifications, and some hypotheses about their origin are proposed. As applied to each individual type, its phonological interpretation was proposed. Of particular interest is the frequency of each type of sandhi. Such calculations, as R. Abduvalieva herself notes, undoubtedly require special research, but still some idea of this can be obtained from her statistical analysis of the fifth chapter of the "Story of Savitri": praslista sandhi-114 cases; abhinihita sandhi-31 cases; ksaipra sandhi-25 cases; udgraha sandhi-0 cases. Such an impressive number of cases of sandhi vowels already in epic Sanskrit-the transition stage from Vedic to classical Sanskrit-indicates that in the ancient Indian language there was a very obvious tendency to avoid vowel gaping.

The analysis of phonemic neutralization occurring in sandhi vowels allows the author of the reviewed work to make an assumption that there were phonological oppositions in Sanskrit, which were determined by the following differential features: "vocality-consonance" (confirmed by the cases of abhinihita sandhi, ksaipra sandhi and udgraha sandhi); "longitude-brevity" (derived from the cases of praslista sandhi);"sharpness - non-sharpness", or relative increase of a vowel (derived from cases of praslista sandhi); "flatness-non-flatness", or relative decrease of a vowel (derived from cases of praslista sandhi). At the same time, the author makes assumptions about which of the features in each named pair are marked. So, in the first pair, the sign "vocality" is considered marked, in the second - "longitude". Two types of sandhi (praslista sandhi and udgraha sandhi) are indicated-

page 185

They indicate the equipolent nature of the "sharpness" and "flatness" signs. Analysis of neutralization cases in the case of udgraha sandhi suggests that in the pair of "flatness-non-flatness" signs, "flatness" ¹ will be marked .

At the end of the first chapter, R. Abduvaliyeva comes to the conclusion that different types of sandhi were formed in different historical periods: abhinihita sandhi dates back to the Indo - Iranian period, and all other types-to the Indo-Aryan period.

In the second chapter, the main regularities of functioning and development of sandhi consonants in ancient Indian are analyzed, and an attempt is made to establish their relative chronology. The author identifies four types of external sandhi consonants: 1) assimilation by the presence of voice (by "sonority-deafness"); 2) sandhi associated with loss of aspiration; 3) sandhi of final s and - r; 4)sandhi of final nasal. According to the statistical analysis performed, the frequency of occurrence of certain types of sandhi consonants in the fifth chapter of the " Story of Savitri "is as follows: sandhi of final nasal - 280 cases; sandhi of final s and-r - 177; assimilation by" sonority-deafness " - 25; loss of aspiration - 2 cases.

The analysis of phonemic neutralization resulting from sandhi consonants allowed the author to identify the following semantic distinguishing features: 1) "sonority-deafness"; 2)" aspirated-non - aspirated"; 3)" diffuse-compact"; 4)"central-peripheral". Marked phonemes are recognized as having the features that are named first in each pair.

The relative chronology of different types of sandhi consonants is given in Table 12 (p. 91). According to R. Abduvaliyeva, the sandhi of the final s and-r dates back to the general Indo-European era; in the Indo-Iranian period, sandhi of nasal consonants and assimilation on the basis of "sonority-deafness" appeared; later than all the others, i.e., in the Indo-Iranian era, sandhi associated with the loss of aspiration appeared (they are known in comparative studies as "Grassmann's law").

In the final part, R. Abduvalieva considers it necessary to draw on the data of some phonemic neutralizations that arise due to a number of internal sandhi, since, in her opinion, the previously identified differential features are not sufficient for a complete and consistent description of the entire inventory of Sanskrit phonemes. Thus, we introduce (in addition to those listed earlier) the distinguishing features of "discontinuity-continuity", "nasality-non-nasality", and justify the need for separate consideration of "vocality-non-vocality" and "consonance-non-consonance". In total, 11 pairs of differential signs were proposed that are necessary and sufficient (from the point of view of the author of the book) to describe the Sanskrit vowel and consonant system. Table 14 (pp. 96-97) gives an idea of how ancient Indian phonemes are characterized by all these features.

At the end of the book there is a rather extensive list of literature on this topic (p. 100-105) and a short glossary (54 units in total) of ancient Indian phonetic terms (p. 106-107), which raises a number of questions regarding the principles of selection and presentation of the included lexical material, as well as the absence of many of the terms mentioned in the work (in particular, there are no specific names for various types of sandhi, which, in fact, was the subject of the reviewed monograph). Individual terms are rather vaguely defined both in the dictionary and in the main text of the book. This remark applies, for example, to the term pluti - "elongated, stretched, longer vowel". In the introduction, this type of vowel is mentioned more than once, including in connection with vo-

¹ I would like to point out that the use of certain terms in the text of the monograph, whose authorship is indisputable and is well known to all linguists, is not entirely correct. R. Abduvalieva does not hide the fact that she relies on the classic work of N. S. Trubetskoy "Fundamentals of Phonology" and uses in a number of cases both his author's classifications of oppositions and neutralization of phonemes, and the terms suggested by them. N. S. Trubetskoy, speaking about marking-unmarking or equipolence, refers these definitions not to the features of phonemes, but to phonemic oppositions. In other words, if you keep the author's intention, then you should not say "marked feature": a phoneme that is characterized by the presence of this feature can be marked, and accordingly, not this or that feature can be unmarked, but a member of the phonemic opposition that does not have this feature. It seems that this remark is not a formal quibble, but, on the contrary, evidence of deep respect for the author's idea and word, as it was proposed and defined by N. S. Trubetskoy himself.

page 186

R. Abduvalieva notes that different ancient Indian sources indicate a different number of vowels in Sanskrit - from 13 to 23; almost all treatises accept the existence of short, long and elongated vowels, while the number of pluti is recognized as different in different sources.

What these special vowels are, where they appear, and what their length is-all this remains unclear in this work. Nevertheless, a very specific answer to this question is also contained in one of the modern books proposed in the list of references. Thus, T. Y. Elizarenkova writes: "Vedic vowels are short (hrasva-) and long (dirgha-). A short vowel contains one mora (matra-), and a long vowel contains two moras. The pratishakhyas also mention super-long vowels (pluti -) containing three moras. These vowels are rare, usually in the verb of an interrogative sentence, so the length should be attributed to intonation, and not to the paradigmatic characteristics of the vowel itself " (T. Ya. Elizarenkova. Grammar of the Vedic language, Moscow: Nauka Publ., 1982, p. 65).

Those parts of the monograph where ancient Indian sandhi is described in the light of N. S. Trubetskoy's theory of phonemic neutralization raise extremely many questions, and R. Abduvalieva's arguments about which of the members of a particular opposition should be recognized as marked or unmarked are particularly doubtful. Let me remind you that in" Fundamentals of Phonology " N. S. Trubetskoy does not offer any clear methodology for determining the marking or unmarking of a particular member of the binary opposition. Therefore, more detailed explanations of R. Abduvalieva herself would be useful, than she is guided by, recognizing flat, voiced, central and other signs as marked. The most vulnerable in this respect are oppositions on the "sonority-deafness" basis, since it would also be necessary to convince readers that for Sanskrit this opposition is privative, and not equipolent, and that it generally makes sense to talk about labeling in this case (and in any case, it is not obvious what exactly "voiced " phonemes are marked members of such oppositions).

For a number of reasons, the classification of Sanskrit phonemes proposed by R. Abduvalieva is puzzling. First, there is no argument at all in the book about the number and composition of phonemes in ancient Indian2 . Most often, the author actually talks about graphic signs, as if the number of letters and the number of phonemes are the same (see, for example, Table 13). Secondly, it is undesirable to connect together, without any comments and explanations, the signs of acoustic and articulatory. In my opinion, in principle, it is not permissible to mix them and consider them in one row: either an acoustic classification is given, or an articulatory one. Third, it is not entirely clear what is behind the "central - peripheral" attribute. Fourth, some acoustic concepts and terms are interpreted incorrectly. Thus, discussing the phonological nature of the sandhi of the final - i and-i, when they turn into glides before subsequent vowels - respectively - u and-v (p.50), R. Abduvalieva for some reason considers it obvious that there is a neutralization on the "vocality" trait and that sonants lose this trait. However, since the publication of the classic work of G. Fant, M. Halle and R. Jacobson "Introduction to Speech Analysis" (by the way, presented in the bibliography), it is known that sonants are both vocal and consonantal. So, it is not at all obvious that in the case of the appearance instead of - i and-and, respectively, - y and-v, we are dealing with the loss of the vocal trait. Apparently, R. Abduvalieva was let down here by the fact that two completely different classifications were mixed, and she either talks about these features from the standpoint of acoustic classification, or means purely articulatory differences between vowels and consonants. Fifth, by the end of the book, you can't help but wonder why all these arguments about different types of neutralization, about marked and unmarked signs, are given at all. what does this have to do with establishing, for example, the relative chronology of ancient Indian sandhas? maybe this sheds light on the problem of the end of a word in common Indo-European, which was raised at the beginning of the monograph? etc.

² The same argument does not seem superfluous or inappropriate. I would like to understand on the basis of what R. Abduvalieva recognizes certain units as phonemes, how she solves the question of the criteria for their selection. This kind of reasoning is given, for example, in the above-quoted "Grammar of the Vedic Language" by T. Ya.Elizarenkova or in the "Grammatical Essay on Sanskrit" by A. A. Zaliznyak.

page 187

R. Abduvaliyeva probably found answers to these questions for herself, but unfortunately, I didn't manage to do it, although I really wanted to.

As a result, the following impression was left. There is a detailed description of the various types of Sanskrit sandhi and exceptions to them, with a sufficient amount of illustrative material, based on ancient literary monuments and grammatical treatises, which can be useful both for linguists interested in this problem and for anyone who simply studies Sanskrit. This is on the one hand. There is an attempt to describe these sandhas using more modern terminology, to reinterpret them in the light of twentieth-century phonological theories, but all separately. This is from a completely different angle. Moreover, in the text of the monograph, there is also a kind of comparative historical padding between them, in which R. Abduvalieva asks questions about reconstruction, tries to establish to which historical stratum this or that phenomenon can be attributed, etc. In my opinion, in these parts it attracts unnecessarily little material and information from other Indo-European languages, although this does not prevent it from building hypotheses about the time of occurrence of certain sandhi.

R. Abduvalieva's final conclusions (p. 99), which conclude the entire study, are quite unexpected:

1. The phenomenon of attenuation of sounds at the end of a word refers to the general Indo-European state, namely, to the late Indo-European proto-language of the collapse era. Similar phenomena are observed in some Indo-European languages, and the tendency in the proto-language to "degrade" (the author's term) the end of the word in some languages is even more pronounced over time. This was due, in turn, to an apparently increased tendency to merge (or fuse) the root and various affixes into a single, closely fused inflectional word.

2. In ancient Indian, internal sandhi is in many cases very similar in nature and final results to external sandhi, which can be interpreted as a reflex of ancient independence, autonomy of morphemes.

3. An analysis of the actual sandhi variants found in texts and grammatical publications revealed the dialectic of the contradiction between attempts to avoid vowel gaping and the desire for it as a reflection of the restructuring, reorganization of the late Indo-European type into the Indo-Aryan language type.

It is rather difficult to call these conclusions new and represent the final result of this essay: firstly, these are all common points, already known Indo-European facts, for the sake of another voicing of which there was no point in conducting such a study; secondly, in my opinion, R. Abduvalieva's monograph was not written entirely about this (just look, what tasks are set at the beginning of the book). "Final conclusions" are almost axioms of modern comparative studies. Having formulated them in this way (with some claim to novelty and to their own discovery), the author rather reduces the overall impression of this work than improves it.

These critical comments certainly do not detract in any way from the numerous advantages of the book under review.

Permanent link to this publication: