Concept formation is considered to be synonymous with categorization. A natural language was already said to tag certain types of conceptual processes by means of words. We may ask now whether the existence of such a word (say in the speaker’s language knowledge) is responsible for making a given concept particularly salient and easy to attain.

A survey of reports of such tests (Carroll, 1964) does, indeed, give that impression; in most instances concepts that may be named, or where the principle may be formulated easily in the native language of the subjects, are felt to be easier to attain in experiments than when this is not the case. However, these findings do not necessarily indicate that natural language is a biasing factor in the formation of concepts in general. The concepts tagged by the vocabulary of natural languages are not completely arbitrary as may be seen from the large degree of semantic correspondences between languages. It is true that translation always brings out some absence of correspondence between two languages. However, the experience of the physical environment finds expression in all languages. It is mostly the aspect or mode of reference and the metaphorical extensions that vary. Comparative studies in the language of experience indicate that those phenomena that have perceptual or cognitive salience in the environment (for our species) always are particularly amenable to reference, regardless of the natural language. Therefore, if namability tends to be coupled with cognitive salience, it is not certain whether results in concept formation experiments are due to subjects’ naming habits or whether both the naming and the concept attainment are due to a more basic factor such as biologically given cognitive organization.

Certain other considerations should make us doubtful about any strong claim upon the ‘ constraint ’ of words upon the speaker’s cognitive capacities. A wide range of human activities is based upon concept formation that must have taken place in the absence of naturally occurring words. Examples are the development of mathematics (where a language is simply created ad hoc as the concepts are developed), or of music, or of the visual arts or of science in general. Peoples in underdeveloped countries who are suddenly introduced to a new technology for which there is no terminology in their language can learn the new concepts by simply introducing foreign words into their vernacular or by making new use of old words. Once more, this may be explained by seeing naming as a creative process, not a rigid convention.

The most dramatic semantic difference between languages may be found in the realm of feelings and attitudes. Here, indeed, translation is often a total impossibility. Are we able to demonstrate that our awareness of personal feelings is selectively enhanced by words handed down to us through semantic traditions? Empirical demonstrations would be difficult and I am, frankly, dubious about their promise. Consider the difficulty to describe accurately the nature of feelings, for instance during a psychiatric interview; or our awareness of how coarse and nondescript some of the words for feelings and attitudes are, for example, honor, love, pride, etc. In many instances, our emotions appear to be more subtle than can be indicated by the use of these threadbare terms.