- The Dravidian languages by Bhadriaju Krishnamurti
Krishnamurti, one of the foremost and most prolific Dravidianists, has produced a book which by any standard is a tour de force, both in breadth of coverage and attention to detail. It will be especially welcome to students of comparative-historical linguistics who are not Dravidianists, as well as scholars of South Asian prehistory, as it brings together materials which have been in many cases difficult to access and provides up-to-date discussions of numerous important issues in the field.
The book combines a focus on comparative Dravidian with a wealth of descriptive material, covering (in differing levels of detail) the twenty-six known Dravidian languages. In addition to descriptions of phonology, morphology, syntax (the last focusing on South Dravidian, particularly the four literary languages Tamil, Malayalam, Kannada, and Telugu), and writing systems, a picture of Proto-Dravidian (PD) is presented that combines all the available knowledge, along with an appendix listing about 700 reconstructed PD forms. The chapters dealing with comparative Dravidian (4–9), which form the core of the book, trace the changes that took place between PD and the modern languages, with frequent references to the evidence for subgrouping and clues to chronology. Much of this is based on K’s own work, including a long list of important contributions (some forty items are listed in the bibliography) on questions of subgrouping, PD gender and number, phonological changes (including the reconstruction of a PD laryngeal H), personal pronouns, comparative syntax, and language modernization, among other subjects. The discussions of linguistic features, both descriptive and comparative, are generally lucid in spite of being frequently quite complex. The book also covers earlier and current work in comparative Dravidian, the culture of PD based on reconstructed vocabulary (6–16), the geography and demography of modern Dravidian languages, relations between Dravidian and Indo-Aryan, suggested affinities with languages outside South Asia, sources of the Dravidian lexicon, the history of work on genetic subgrouping in Dravidian, and the chronology of the stages of PD.
Much of course remains to be done in the field of comparative Dravidian, as K acknowledges (502), and much remains controversial. A strength of this book is its many detailed discussions of the conflicting views of other scholars on a number of points. There is space here for only a few brief remarks on these issues. On subgrouping, K presents the now generally accepted tripartite view, which he himself has had a large role in shaping, with North, Central, and South Dravidian (ND, CD, and SD) as the main branches, the last divided into SDI (Tamil, etc.) and SDII (Telugu, etc.). Although K has apparently laid to rest the notion that Brahui is a separate branch of Dravidian (500), there still remains a question regarding the stages of the breakup of PD. While acknowledging that a two-way split is inherently more probable than a three-way one, K’s rejection of an original bipartite division of Dravidian (ND _ CDSD) is based on the argument that ‘there is lean evidence to set up a common stage of South and Central Dravidian’ (494). While there may be no exclusively shared innovations involving the entire CD-SD complex, it is clear that there was some interaction between CD and SDII, as shown by overlapping isoglosses (494–501), whereas there is no evidence for interaction between ND and the other branches. Thus while not strictly belonging to a single subgroup, the SD and CD languages apparently remained in some sort of contact at a time when ND languages were no longer in the picture.
This raises the question of the validity of reconstructions based on etyma without ND cognates, though K believes (p.c.) that it is generally possible to weed out loanwords between CD and SD. However, in a number of cases K has reconstructed PD items on the basis of cognates found only in SDI and/or SDII—that is, in only one of the main branches—on the grounds that data for the nonliterary...