336 Language and cognition
what kinds of words may be interchanged in sentence frames without changing the grammaticality of the sentence . For further details see Fodor and Katz ( 1963 )
Categories may also be related to one another such that the relationship itself is labeled by morphemes or words ( for example , by , -ing , for , is , etc.). The semantics of these relational elements perhaps illustrates best how words do not refer to real things but to cognitive processes . It is in connection with the cognitive process of relating abstract concepts（name categories or structural sentence-types）that the intimate relationship between semantics and syntax is most clearly revealed .
(4) Preliminary Conclusions
The basic cognitive mechanisms underlying semantics appear to be similar to those of syntax, namely processes of categorization, differentiation, and interrelation (transformation). Indeed, the latter two are merely aspects of the categorization process itself. This argument may also be extended to include the cognitive processes underlying phonology. Although categorization is a universal phenomenon in the animal kingdom, the categorizations peculiar to language operate through the application of highly species-specific principles.
Naming is a process, not a catalogue of rigid conventions. There are two types of constraints that determine naming-behavior: the biological constraint upon the physiological processes that determine the species’cognitive capacities, particularly the conditions under which similarities are recognized; and the constraint necessitated by the communicative function subserved by naming. Notice, however, that the naming process may go on in the absence of communication. Neologisms are created and assigned meanings by the schizophrenic patient or the genius creating words for his highly idiosyncratic concepts，regardless of whether they will actually transfer information from one individual to another.
It is matter of social dynamics, roles, values, or group-mechanisms whether anybody will take the trouble to discover what conceptualization process was being tagged by such neologism . Communication is a social phenomenon , whereas naming is an intrapersonal one ; the intrapersonal process may become a social one by virtue of enormous similarities between the cognitive functioning of all individuals and an apparent specific motivation in human to interact socially . Again, group cohesiveness is a widespread phenomenon in the animal kingdom, but the mechanisms vary greatly from species to species. In man, the prime vehicle of this interaction is the mutual adjustment of concept formation, tagged by words.
III.Empirical study of naming : the language of experience
III. THE EMPIRICAL STUDY OF NAMING : THE LANGUAGE OF EXPERIENCE
In section II I have proposed that words tag conceptualization processes, not things directly . But since the input to at least some conceptualization processes is the sensory perception of things, there is , of course, also a relationship between words and object. This relationship may be studied empirically, under laboratory conditions. For this purpose, it is desirable to choose words that have simple referents, in the sense that they can be exhaustively described by objective measurement of all of their physical properties. This condition is fulfilled in only a small class of words, namely those that describe the sensation of certain physical properties themselves. Let us refer to these words collectively as the language of experience; examples are words for temperature, taste, hearing,or vision
(1) Description of the Referents
The language of experience is particular well-suited for research because its referents have four advantages over the referents of most other type of words：first. they may be ordered by object，logical criteria (for example, the centigrade scale for temperature：the frequency scale of pure tones of equal intensity , etc), whereas furniture , relatives, or most other concrete referents have so much logical and unique orders.) Second, the referents have continuity in nature (for example, within certain limits any degree of temperature may be encountered or any sound frequency heard. Whereas the domain of chairs do not grade into the domain of say, benches, nor do uncles grade into aunts. Third. words in the language of experience refer to closed class；our sensory thresholds set limits to perception and thus there is a bound to the range of phenomena that may be called hot. loud, green, etc. Fourth, the referents are simple in the sense that each instant may be completely specified by a fixed and very small number of measurements. Temperature is specified by just one measurement; pure tones by two－intensiy and frequency：colors by three－for example, the Munsell scale of hue, brightness, and saturation.
338 Language and cognition
(2) Referent Spaces
If it is possible to describe phenomena completely and accurately in terms of certain dimensions, it is also possible to construct coordinate systems in which every point corresponds to one specific phenomenon. Every point on a thermometric scale corresponds to one specific temperature；every point within a pitch-intensity coordinate system corresponds to one specific pure tone. Sometimes such coordinate systems are called spaces. The coordinate systems mentioned so far are unidimensional and two-dimensional spaces；for the specification of color we need a three-dimensional spaces, and since these spaces are merel abstract, mathematical concepts, one can also have four-, five-, and n-dimensional spaces. The three-dimensional color is most easily visualized. Figure 8.1 shows the arrangement of the coordinates, and Fig. 8.2 shows an actual model of a color space. The choice of a polar coordinate system is arbitrary. It offers some conveniences, but the three dimensions might have been arranged differently.
With the introduction of the notion of a color space we might review once more why we wish to arrange physical phenomena in this way. Our aim is to study how words relate to object. But we cannot study the behavior of words unless we keep some control over the object