The point of departure is in fact in the use of the term language behaviour rather than `language' or `knowl- edge of the language'. A child is said to have learnt language when "he can use it to interact with others in the language modality...,"(p.9). To claim that "the knowledge exists in the child as a set of rules" could be validated "through the perform- ance of another behavioural system which given these rules... is able to exhibit (similar) language behaviour...". The theoreti- cal distinction between competence and performance, and between knowledge and behaviour is untenable.

With the above view point, behavioural pragmatics becomes the object of study. A detailed field study is described using the Vanitha corpus, which goes to show the different stages of child language acquisition. The learning process begins to fall in place when one uses notions from pragmatics rather than "syntax". The corpus shows among other things the language schemata used by adults while interacting with Vanitha. Narasim- han argues that the child is exposed to highly structural lan- guage data ("motherese" schemas) rather than degenerate one. And this happens in situations in which the child has information derived from sensory modalities. Thus, the child is able to relate language expressions with sensory inputs under highly structured situations.

But language goes beyond this. The sensory-motor behaviour system deals with `now' and `what is out there'. With language capability, one is able to refer to and mediate between the sensorily apprehended aspects of the world. In Pavlovian terms, it is a signal of signals. It is this capability that allows it to be deployed for strategic purposes (e.g., planning, problem-solving, social manoeuvring) as opposed to mere instrumental usage (e.g., describing, communicating). One sees ample evidence of it in the corpus, where the child is not just relating language expressions to other sensory inputs, but is also relating to her needs of exploring and manipulating as well. Narasimhan analyzes the data from the Kanzi Chimpanzee experiments, to show that Kanzi uses his language capability at most in the instrumental mode. In contrast, Alia (the human child also involved in the experiment) deploys her language capability to strategic ends.

Imitation and rehearsal have been observed in any learning, including that of language behaviour. Lately their importance has been denied as a result of the influence of the Chomskyan paradigm. In their place, studies have been conducted on the exposure of the child to syntactic structures graded by complexity of syntactic structure. Results of such studies have belied the expectations of the paradigm. Narasimhan points to several such studies in the literature and for some of them shows how the "anomalous" data is not anamalous any more if one views it from pragmatics. The impulse to initiate, urge to tell one's experience to others, and capacity for role playing (imitation, rehearsal and analogizing) form important behavioural processes for learning. The language behaviour is learnt through the same processes as used in learning other skills and abilities.

Narasimhan also considers the issue of the theory of mind. He considers the work of Premack and Woodruff (1978), Olsen (1988), Savage-Rumbaugh et al. (1993) etc. and relates to his overall thesis. I would recommend readers to look at Table 8.1 for a summary of the increasingly complex levels of language modality. An implication is that what distinguishes human from animals is not tool use but language use; that too, not in mere instrumental usage but in the ability to model others' minds and deploy such a model for strategic purposes.

The book contains a wealth of information besides the above. Author also discusses language acquisition studies related to the blind and the deaf and what these tell us about language acquisition process. He also dwells upon literacy and orality work and how the heavy duty theories of syntax might apply to the literate language and not to the oral language (which is naturally learnt). He considers the controversy regarding the evolution of language and what implications it has for language acquisition.

What does the book imply for natural language processing (NLP) and machine translation (MT)? It says that while acquiring language, humans learn the pragmatics of its use. The learning involves learning myriads of facts and fragments, rather than a few rules of syntax. Learning of discrete syntactic rules, if any, is tied to the rest of the system of their use. This is close to the modern trend of corpus-based processing in NLP and MT. Memory-based or example-based approaches are closer to the thesis. Rule-based approaches such as those used in case-based theories and Paninian grammar (Bharati et al., 1995) would be useful adjuncts as they relate to semantics and pragmatics. The current need for man-machine interaction in the language modality is already bringing in pragmatics into NLP in a big way.

This book with its panoramic view of diverse researches relating to language: orality-literacy, child and chimpanzee language studies, linguistics and psychology - must be read by every serious researcher. It presents a framework in which real questions relating to language use can begin to be posed and answered. Thus, it paves the way for a fresh new approach to language related research which can serve as a foundation for viewing and combining the different pieces of the languge puzzle.

REFERENCES:
Astington, J.W., P.L. Harris, and D.R. Olsen (eds.), Developing
Theories of Mind, Cambridge Univ. Press, UK, 1988.
Bharati, Akshar, Vineet Chaitanya, and Rajeev Sangal.  Natural 
Language Processing: A Paninian Perspective, Prentice-Hall of 
India, New Delhi, 1995.
Olsen, D.R., On the Origins of Beliefs and other Intentional
States in Children, in Astington et al. (eds.) (1988), pp. 414-426.
Premack, D., and G. Woodruff, Does a Chimpanzee have a Theory of
Mind?, Brain and Behavioural Sciences, 4, 1978, pp. 515-526.
Savage-Rumbaugh, E.S., J. Murphy, R.A. Sevcik, K.E. Brakke, S.L.
Williams  and D.M. Rumbaugh, Language Comprehension in Ape and
Child, Monograph of the Society for Research in Child Development,
58, Nos. 3 & 4, 1993.