It is John Ellis’s contention that notwithstanding the enormous amount of energy that has gone into language studies not very much progress has been achieved and our theory of language remains vastly fragmented and contradictory. Language, Thought, and Logic is presented as an effort to overcome this impasse. Many ideas are discussed, but the fundamental themes are the following four: the priority of categorization over communication, the functional nature of most categories, the attack on the severance of grammar from meaning which characterizes generative grammarians, and the dismissal of the debate on the relationship between language and thought as a meaningless issue.
In chapters two and three, Ellis—after having convincingly argued for the logical priority of (language induced) categorization over communication—proceeds to illustrate the difference between “real” and “functional” categories. He states that a few categories, such as chemical classes, have their base in “reality.” Most categories, however, (Ellis uses the example of poison) are not based on intrinsic characteristics of its members (poisons differ widely in chemical structure) but on the collective purposes of the speakers (identifying dangerous substances). The former are “real” categories, the latter functional ones, and on the basis of this opposition Ellis attacks “naive realists” who claim that most linguistic categories are grounded in the extra-linguistic world. Ellis’s dichotomy however is not fully convincing: reality, of course, includes human beings and their interactions with the world. Once we concede that the chemical structure of a given substance is a real fact, then we have to admit that the interaction of a given substance with the human body is equally “real,” and [End Page 142] that a category based on a given type of interaction is grounded in “reality” just as one based on a given type of chemical structure.
Chapter five instead is devoted to the traditional question of the relationship between thought and language. Ellis’s contention is that this discussion is in itself logically absurd, since language is an essential part of thought. This “initial logical misstep” inevitably leads thinkers such as John Lyons, Max Black, Eric Lenneberg, and no doubt countless others, to “astonishing” and “bewildering” statements, pseudoproblems,” and “comic” errors (pp. 56–58). In this case, too, Ellis’s position is not particularly convincing, and one wishes that he had devoted some of his unforgiving examination of the logical integrity of other people’s positions to his own arguments. For readers are bound to ask themselves why on earth does the fact that language is part of thought make it “not possible” to speak of the relationship between the two. Is it logically “not possible” to speak of the relationship of the part to the whole? Is it “not possible,” for example, to speak of the relationship of the heart to the rest of the human body?
In general, by denying the possibility of discussing the relationship of thought and language, Ellis ends up completely ignoring the tension between the more standardized and collective aspects of language (grammar, and denotation, in particular) and the various ways it becomes articulated within individuals, through its inevitable enmeshment in their different experiences. This is a pity, because in many ways Ellis is trying to carve a commendable path between the Scylla of “deep structures” and the Charybdis of “linguistic determinism,” between the severance of language from meaning, and their absolute identification. It is perhaps inevitable that in shying away from the latter peril he ends up contradicting himself. After having declared the logical impossibility of speaking of the influence of the structure of language on thought, he subsequently states that it is certainly not true that everything in our mental life is completely determined by the structure of our language” (p. 61).
Ultimately one is left with the impression that what would have best been offered as a refinement and integration of other positions has been polemically presented as a radical alternative, with the risk of helping to perpetuate the fragmentation of linguistic studies it was meant to overcome. Thus, as far as its ambitious goal of...