Language
Language
I. The Psychology of LanguageCharles N. Cofer
II. Language DevelopmentSusan M. Ervin-Tripp
III. Speech PathologyJack Matthews
IV. Language and CultureWilliam Bright
I THE PSYCHOLOGY OF LANGUAGE
The psychologist tends to look at language in the first instance as he would look at any other problem area. Unique or special problems are confronted as they arise, but there is no unequivocal case for or against a special psychology—one created for language alone. The best, or, perhaps, the only way to discuss the psychology of language at this time is to describe how psychologists have been looking at it. Much of what can be said receives consideration in other articles in these volumes. The present one can serve as an introduction to the other articles, and it will be concerned, in part at least, with an attempt to set the framework which marks psychological studies of language. Certain more specific problems will also be treated—problems which represent some of the writer’s special interests and which are unlikely to receive extended discussion elsewhere.
Early interests of psychologists in language
Studies of language—or, as psychologists prefer to call it, verbal behavior, which applies to both spoken and written forms—began very early in the postphilosophical period in psychology. Several themes may be identified (see Carroll 1953).
The “word-association experiment.”
First, Wilhelm Wundt, often called the first experimental psychologist, was interested in language—but more from a naturalistic than from an experimental point of view. However, workers in his laboratory early took an interest in the “word-association experiment,” the origin of which is usually attributed to Sir Francis Galton. Galton’s pioneering work was carried out with himself as subject: he wrote down stimuli on slips of paper and later looked at each slip (on more than one occasion) and recorded the thoughts that were thereby elicited. He also timed these reactions and noted the tendencies for the same or different thoughts to occur on the several occasions on which he looked at each slip (Warren 1921).
The “word-association experiment,” despite its early entrance into the psychological laboratory (some of the results obtained are still cited today; see e.g. Woodworth [1938] 1960), soon figured more prominently in clinical diagnostic work than in the experimental laboratory. While the associations made by pathological subjects and by normal subjects were often found to differ, on the average, the association method never reached a dominant position in the armamentarium of clinical techniques, in the detection of “complexes,” or, except in classroom demonstrations, in the detection of “guilty knowledge,” although it was applied to all of these problems. It is safe to say that until about 1950 it had fallen into relative neglect. In the years since 1950, however, interest in association methods has reached a very high level because of the demonstrated value of measured associations in predicting other verbal behavior [see alsoMarshall & Cofer 1963; Noble 1963; see alsoAnalytical Psychologyand the biography ofJung].
Verbal learning
The study of verbal learning was begun in the laboratory almost as early as the word association method. Hermann Ebbinghaus reported the first experiments in a classic monograph in 1885. Ebbinghaus, however, was not really concerned with general questions of language or verbal behavior; he was interested instead in the formation of associations, and he attempted to prevent such factors as meaning, meaningfulness, and connectedness from influencing his results by inventing and using the nonsense syllable. A vast outpouring of work on verbal learning followed Ebbinghaus’ pioneer efforts (see e.g. McGeoch [1942] 1952; Conference …1961; Conference…1963), but current opinion is that verbal learning cannot be divorced from features of the subject’s usual language. [SeeLearning, article onVerbal Learning; and the biography ofEbbinghaus.]
Mediational processes
A third concern in psychology’s study of verbal behavior is the role verbal processes may have in mediating between stimuli and responses. Quite early (in the 1890s) there was an interest in “mediate association,” the notion that while two terms may not be related directly to one another they may be linked by a third term. Thus, justice and war may not be directly associated, but they may be related if, when one thinks justice, he also thinks peace. The latter may lead to the word “war,” thus serving to mediate between justice and war.
Mediational processes, especially verbal ones, have provided a mechanism whereby objective and behavioristic psychologists can account for forms of thought, stimulus and response equivalence, and other phenomena which are otherwise refractory to a simple analysis in terms of stimulus and response (Goss 1961a). Much of the impetus to this way of thinking came from studies of semantic conditioning, both in Russia and in the United States (Osgood 1953; Jenkins 1963), in which verbal behavior is given an important role in the control of other behavior (Luriia 1961; Bollard & Miller 1950). There has also been great interest in the possibilities for altering verbal behavior itself (and thus perhaps its control over other behaviors) by means of reinforcement and nonreinforcement (Krasner 1958) or by other means (Cofer 1957; 1960).
Psycholinguistics
The fourth, and final, major trend appeared in the early 1950s and was a convergence of psychology (especially learning theory), descriptive linguistics, and information theory. This convergence was facilitated by the Social Science Research Council and the Carnegie Corporation (Carroll 1953). Essentially, it amounted to the bringing together of linguists and psychologists (see Osgood 1963; Osgood & Sebeok 1954) so that each group could become familiarized with the techniques and concepts of the other. As a result, an interdisciplinary “field”—psycholinguistics—has sometimes been identified, although its methods and concepts perhaps tend to be more aggregations than mergers of the concepts and methods of the disciplines. “Psycholinguisticians” tend to retain their primary identifications as linguists or as psychologists. The development of the notion of a” generative grammar” (Chomsky 1957) in linguistics may well alter the character of psycholinguistics. [SeeLinguistics.]
It must be pointed out that what has been said implies very little concerning language as a process of interindividual communication. It is obvious that communication is a major function of language, but from the present vantage point it seems that matters other than communication per se have engaged the interests of the major psychological students of language or verbal behavior, although studies of communication processes do go on. Also neglected in the foregoing account is speech perception and speech pathology, the study of which can yield important and useful information on a variety of problems. [SeeLanguage, article onSpeech Pathology; Perception, article OnSpeech Perception.]
Viewpoints concerning mediational events
A major theoretical concern among psychologists interested in verbal behavior has been the implicit events which may accompany overt speech production or writing. Cofer and Foley (1942) suggested that a mediating response underlies cases of stimulus equivalence (or transfer or generalization) among stimuli which are physically dissimilar. Thus, if a response is learned to a word like “fashion” and it transfers, without further training, to the word “style” or “mode,” the transfer cannot be explained on the basis of the visual appearance or the sound of the words. The argument advanced, in essence, was that each of these words elicits, as a result of prior experience, a common reaction. If this reaction occurs when fashion is presented, it will be associated with the new response being learned to fashion. Since style and mode also elicit the common reaction, its occurrence when they are presented would also result in the appearance of the response newly learned to fashion. The common reaction mediates the transfer. Cofer and Foley stressed the relation of synonymity among words as an indication of the existence of common mediating reactions among them, but they also spoke of interword associations in such a way as to suggest that associations, also, might serve as mediators.
Osgood’s views—representational mediation
Osgood (1953; 1961) has made the mediating reaction the basis of a theory of meaning. Two words mean the same thing to the extent that they share the capacity to arouse the same representational mediator. The conception is that, with respect to an object, there is behavior, Rt (for total behavior). A sign, e.g., a word, may be acquired in relation to the object; but to have meaning with reference to the object the sign must arouse some representative part of the Rt made to the object. The representational part would ordinarily be those features of, Rt which can readily be detached from the total and occur implicitly without interfering with other behavior.
This formulation led Osgood to make certain distinctions among word relations with respect to meaning. Thus, contrasting or opposite words (antonyms) cannot share common mediators and hence cannot have common meanings. This is because Rt with respect to objects or events must be very different in the case of antonyms. The representational mediators, in turn, would then be very different for antonyms. Thus, Osgood is forced to explain the fact that antonyms are often highly associated in word association tests (e.g., black–white, up–down) on the basis of a rote verbal habit rather than on the basis of common representational mediators. Similarly, he took the position that the existence of interword associations does not demonstrate, generally, the presence of common meanings. He also seemed to hold that transfer and generalization should ordinarily be predicated on common representational mediators rather than on common associations.
The semantic differential. As an index of a representational mediator, Osgood (see Osgood et al. 1957) emphasized the semantic differential. This is a technique by which a given word or concept (e.g., baby) is rated by a subject on each of a number of seven-point rating scales, the extremes of which are defined by polar adjectives. Thus, scales appear whose extreme points are designated by such adjective pairs as hot–cold, bad–good, active–passive, tense–relaxed. Factor analysis of the intercorrelations among such scales has usually yielded three more basic dimensions—evaluative (e.g., bad–good), activity (e.g., active–passive), potency (e.g., strong–weak). It is possible to describe a given word in terms of where its ratings fall in a semantic space defined by these three (and perhaps other) dimensions. As these dimensions are believed to characterize the representational mediator, they are an index of its meaning. Osgood has referred to this as connotative (or emotional) meaning in contrast to denotative meaning.
Among psychological investigators of verbal behavior, Osgood has perhaps been the leader in emphasizing that meaning is a critical factor in verbal behavior and that a satisfactory account of language must deal with it. Others have not been convinced. Skinner (1957) has written a psychology of language in which the concept of meaning does not enter and which treats verbal behavior as a case of the operant. Skinner is concerned with the analysis of this operant (and the classes into which it may be divided) in terms of the variables (reinforcement, stimulus control, deprivation, aversive conditions) which control its strength. This is essentially a descriptive or positivistic procedure, representing an extrapolation of notions developed in the animal laboratory.
Association in mediation
More closely associated with the mainstream of the psychological study of verbal behavior, however, have been the investigators who have emphasized interword associations as the important problem. They have insisted that the test of what is important is not whether a conception explains meaning. Rather, they have argued that the extent to which predictions can be made of the behavior of words in a variety of situations is critical. In other words, they have held forth not the criterion that a conception of verbal behavior explains (or ignores) meaning but the criterion that a theory of verbal behavior must accept measures and procedures which correlate well with other phenomena in the verbal realm.
To give this approach (it can hardly be called a theory) some concreteness, we may describe an experiment made a number of years ago at the University of Minnesota (see Jenkins 1963 for references and discussion). This experiment (and many others as well) was designed, in part, to determine whether meaning (in Osgood’s sense) or association is a more effective variable in predicting transfer in certain situations. Before discussing the experiment, we may turn first to a discussion of contemporary procedures of association tests.
Association tests. The word association test, as it has been typically used, consists of a list of words to which a subject reacts, one at a time, by saying or writing down the first other single word that comes to mind when he sees or hears a list (stimulus) word. This process is called free association since no restriction is imposed on what response the subject can give, save that his response must be a single word and that repetitions of the stimulus word are prohibited. Despite its name, the procedure is not to be confused with the free association procedure as used in psychoanalysis: the two techniques differ widely.
The word association test is usually given to a number of subjects (from fifty to one hundred are considered a minimum for the establishment of norms), and their responses are tabulated for each stimulus. When such tabulations are made (and the stimuli used are relatively common words) the typical result is this: one word is given by a substantial number of subjects, another by somewhat fewer subjects, a third by still fewer subjects, and so on until words are found that are given by only one subject. An illustration will clarify this situation. Suppose the stimulus word is “length,” and we obtain responses from 56 subjects. In such a sample the most frequent, or primary, response is width, occurring 18 times (32 per cent), the next is long, occurring 5 times (9 per cent), the next are height and measure, tied at 4 occurrences each, and there are 6 responses, each occurring twice— short, depth, foot, line, measurement, distance. In addition, there are 13 responses each occurring but a single time.
This set of associations, and others like it, is often termed an association hierarchy, because the responses differ and may be ordered in accordance with their frequencies of occurrence to the stimulus word. Such hierarchies may differ in a number of properties. Thus, the primary frequency may vary considerably (e.g., table elicits chair as a primary from about 75 per cent of a group of U.S. college students, in contrast to the 32 per cent indicated above), and the difference between the primary and other responses themselves may vary. Likewise, the number of different words given to a stimulus varies; usually the higher the primary frequency the fewer other words there are, a fact dictated by arithmetic if we are dealing with a constant sample size.
These differences in associative frequency have been taken as a measure of association or habit strength, not only for the group, but, by inference, for the individual. Though it is by no means a certain inference, there is evidence which justifies it (cf. Russell 1961). At any rate, a number of experiments have been predicated on this inference.
The dominant association to length, as we have seen, is width. The dictionary records a number of meanings for the word “length,” but width is not among them; measure, measurement, and distance are, however, approximations to the dictionary statement of some of the meanings of length. We may say that width is a prominent associative meaning of length, if we wish to, but it certainly does not qualify in a dictionary or denotative sense as a meaning of length. This, of course, illustrates Osgood’s argument that associations are not the same as meaning.
Predictive value of associations. The question, however, may still be raised whether the associations are effective in mediating transfer (and other phenomena) and, if they are, how they compare with synonyms in doing so. To answer this question, the following experiment was performed.
Two lists were constructed, each consisting of a series of word pairs. The subject was required to learn the two lists in succession, and the interest lay in the transfer effects of learning the first list upon learning the second list. Response members of the pairs in the two lists were related in two ways. Some responses were antonyms, but highly associated; since meaning has been ruled out by Osgood as a factor in the interrelation of antonyms, these pairs test the role of association. Other response members of the pairs were synonyms which, according to the word association test, were not associated. These test the capacity of meaning to mediate transfer without association. There were also control pairs, in which the response members of the two lists were neither synonyms nor associates. Table 1 shows the general experimental plan. When compared to control pairs in List 2, the pairs having associative and synonym relations with pairs in List 1 showed the effects of positive transfer in List 2. The associated pairs tended to be learned in a somewhat superior fashion than the synonymous pairs. In this experiment, then, both the postulated factors were effective. One can say that common meaning may effect transfer but that transfer can occur without it if suitable associations are present.
| Table 1 – Experimental pla | ||
|---|---|---|
| CONDITION | LIST 1 | LIST2 |
| Associated | eagle–sickness | eagle–health |
| Synonyms | mutton–long | mutton–tall |
| Controls | needle–king | needle–close |
Many other experiments have been carried out which have explored and demonstrated the role of associations in a variety of situations. Among the situations studied have been free recall, list learning and recall, semantic generalization, recognition of words (in the context of other words) after very brief or inadequate presentations, verbal discrimination, transfer, and concept formation (for references see Deese 1961; Jenkins 1963; Gofer 1957; 1960; and Marshall & Gofer 1963). The role of meaning has also been studied in some of these situations, and some conflict has arisen over whether meaning or association is the better concept or whether one can be reduced to the other (see Bousfield 1961; Osgood 1961). Since the role of mediation is critical to this work, whether interpreted in terms of meaning or association, a good deal of effort has been devoted to basic mediational processes and mechanisms (Jenkins 1963).
Associations in other tasks. Two further areas of inquiry may be discussed to demonstrate the interest of psychologists in the role of associations in such tasks as concept identification and problem solving.
Underwood and Richardson (see Marshall & Gofer 1963) selected a large number of words, to each of which they obtained “sensory” associations from a large group of college subjects. The subjects were carefully instructed as to the kind of association they were to give; thus it was pointed out that to such a stimulus word as “apple” responses like red, sour, or round would be appropriate; responses like tree, fruit, or seed would not meet the criteria for a sensory association. To each stimulus, sensory associations were arranged in a hierarchy according to frequencies, called “dominance levels” by the investigators.
It is possible then to present several words to a new group of subjects, each word having a sensory associate in common with the other words of the set. The subject is asked what these words have in common. For example, one might present the words “chalk,” “milk,” “paste,” and “shirt” and ask what they have in common or in what way they are similar. The instructions here would not mention or suggest sensory attributes. An answer for the set just mentioned might be that they are or can all be white. The subjects are more successful in finding the expected solution for sets with high dominance levels than for sets with low dominance levels.
This finding suggests that in tasks of this kind solution will be more likely if the instances suggest the answer than if they do not. More generally, it is consistent with the idea that problem solution often derives from responses to the materials available. If the materials suggest available and appropriate responses, solution will be quick; if they do not, or if they suggest inappropriate or incorrect responses, solution may be delayed, impeded, or prevented altogether.
A somewhat similar analysis may be made of an experiment by Judson and Gofer (see Gofer 1957). These investigators developed a number of four-word items; in each item, the subject was to exclude the word that did not go with or belong with the others. The critical items had two possible solutions. Thus, the item
skyscraper temple cathedral prayer
can be solved by excluding prayer, as it is not a building, or by excluding skyscraper, as it has nothing to do with religion. In the investigation carried out, an important factor found to determine solution was the word order of the item. As the item is shown above, prayer is excluded more often than skyscraper. If, however, the positions of prayer and skyscraper are interchanged in the item as presented, then skyscraper is excluded more often than prayer. Evidently, associations are initiated by the first word of the item and reinforced by the next two words (which are ambiguous in that they refer both to buildings and to religion) so that the last item is typically seen as not belonging. With highly religious persons, however (as measured by church attendance), the nonreligion item tends to be excluded no matter what its position is. Thus, while word order itself is important to the dominance of the concept to be identified (and thus to the word to be excluded), dominance can also arise from other, perhaps more personal, sources.
Associations should influence behavior over a period of time if their importance is great. A demonstration that is relevant was made by Judson and Gofer (see Gofer 1957). The procedure was divided into two parts, separated by a six-week interval. In part 1, each subject was asked to give ten free associations to each of ten stimulus words. Only one of these sets of associations was used later, however. Suppose one stimulus word was music and the first four associates given by one subject were tune, song, instrument, and melody. In part 2, for this subject, tune was presented on a card along with three other words (each on a card) that were new to the experiment. The subject was told to pick one of the four words. If he was “correct” (i.e., he chose tune) he was told that he was right. If he was “wrong” (i.e., he picked one of the other words), he continued to pick until he chose tune. After performing satisfactorily on this set of cards, he was then shown four more cards containing song and three new words. This time, he simply made a choice, with no information being given as to its correctness. Then he was shown four more cards, containingnstrument and three new words and made a choice (without information); then melody with new words and so on until his sequence of ten associates was used up. The question was, would the subject choose more often the words that he had associated with music after the first response (tune in the example) was said to be correct than he would choose the new words which were not part of the associative chain? As compared to the control group, the subjects answered this question affirmatively, indicating that the chain established in part 1 was still intact and that subsequent choices were influenced by the designation of the first word of the chain as correct.
Another way of investigating the role of associations in problem-solving behavior was used by Gofer, Judson and Gelfand (see Gofer 1957). In this experiment, the subjects were asked to solve certain problems. Prior to attempting the problems, the subject was taught several short lists of words. One of the lists contained a sequence of words, which, if active in the subject’s mind at the time of problem solving, could influence the kind of solutions used in solving the problems. The results were suggestive, if not definitive. That is, for male subjects at any rate, the frequency with which one particular solution appeared for each task was augmented (as compared to controls), and this solution was the one to which the word list was related.
These experiments on concept formation and problem solving do not cover the entire range of these phenomena but, so far as they go, they do suggest that verbal associative processes can have an important effect in these tasks. And experiments such as these illustrate very pointedly what is meant by the role of verbal responses in mediating and controlling other behaviors. [SeeConcept FormationandProblem Solving.]
Work of the kind we have just described illustrates the influence of association tests, learning theory, and the interest in mediation processes in the control of behavior. Our description should be clear on one point: much of what we have said has been concerned essentially with single words and their relations to one another. Language involves more complex arrangements of words than this, and in psychology some attention has been paid to these complexities. Both linguistics and information theory contribute knowledge and techniques to considerations of this problem of complexity.
Contributions of information theory
Information or communication theory has contributed, in addition to a model of the communication process and a means of measuring the amount of information transmitted, the important concept of redundancy. In successive segments of a sequence of words taken from normal English, there are or may be dependency on what has gone before. In such cases, the listener (or reader) can often predict accurately what is to come, and when it comes its occurrence does not contribute any information (i.e., resolve any uncertainty) over and above what the listener or reader already has. [SeeInformation THEORY.] Some of this redundancy arises from structural features of the language; other aspects of it may arise from semantic (associative?) features. In the sentence “Tom went to the ——” there are many items (such as movie, play, concert, game, exhibit) which can be inserted in the blank with equal appropriateness. However, all of them are nouns; we cannot insert pronouns, adverbs, adjectives, verbs, conjunctions, and prepositions if only one word is to be placed in the blank and the sentence is to end with that word. This is an example of structural redundancy: our choice is limited to nouns in filling the blank. We cannot use just any word in the language to do so. As the noun class contains a very large number of items, we still have many choices but we are, nonetheless, restricted by the grammatical situation (here a prepositional phrase which contains an article before the blank) to selection among those words which have the privilege of occurrence in the situation.
In many situations, it is easier to fill in blanks when they stand for items that have chiefly structural rather than semantic significance. Thus, it is more likely that the blank in the following “Tom —— going to the play” will be replaced with the word “is” or “was” than it is that the blank in the first version would be replaced by any of the noun alternatives listed. There is, we may say, more redundancy in the case of is or was than there is in the blank in the noun phrase, although there is redundancy there also. With content words (nouns, adjectives, verbs, or adverbs), the semantic or associative context often severely limits the possible choices. Thus, in the sentence “Tom went to read the part in the ——” few choices (play, drama, for example) remain to the reader or hearer; redundancy here is high.
One of the advantages for learning that connected discourse has over unconnected words is conferred by its redundancy, both structural and semantic (Deese 1961). The influence of structural redundancy is perhaps confined to strings of words in which some features, at least, of the language are preserved. Strings of unrelated words, however, may be associatively or semantically related; thus there may be some redundancy in such terms.
It is clear that redundancy is an important feature of language; it is probably essential to accurate communication (especially in noisy channels), and, for memory, it is very helpful that large amounts of material can be coded (thus reducing the load on memory) in terms of redundant features (Miller 1956). That language is often used to code or categorize (Brown 1958) the environment is perhaps a feature of the greatest importance to intellectual functioning and to the role of language in the control of behavior.
Contributions of linguistics
It is probably fair to say that linguistics has had its greatest impact on psychology because of its knowledge of the structure of language and the rules that govern the structure. Phonemic and morphemic analysis has been instructive, but morphology and syntax are perhaps more provocative.
There are at least two fundamental issues that morphology and syntax raise for a psychology of language. First, it is clear that no theory of verbal behavior which confines itself to semantic and associative relations among words alone is complete. Mechanisms must be developed to cope with semantic and associative changes that undoubtedly occur in the context of other words, and syntactic restraints must be an important aspect of this context. Furthermore, no psychological theory of morphology or syntax has been presented as yet. Analysis so far has largely been confined to specific cues in speech or writing which might elicit word inflections or specific syntactic forms (Goss 1961 b).
The second issue arises from the fact that at a descriptive level it is clear that speech regularities appear very early in the development of the child (Brown & Fraser 1963). Furthermore, many rule-abiding features of speech are known to occur widely, even though the speakers are unaware of the rules or of the fact that they are following rules.
Young children display the operation of a general rule when they pluralize goose incorrectly as gooses or form the past tense of the verb “take” as “taked.” Perhaps more subtle examples are found in the speech of adults. We all recognize that there is something wrong with the expression “the sheer two silk stockings” and would prefer to say “the two sheer silk stockings.” We would have difficulty in formulating the rule that makes one expression acceptable, the other not. Despite its semantic nonsense, we can recognize the grammatical correctness of Chomsky’s (1957) famous sentence, “colorless green ideas sleep furiously,” or make the appropriate substitutions of nouns, verbs, and modifiers in such strings as “the glibs duxed the neglan gojeys.” Semantic factors are not essential to the identification of grammaticalness or to the identification of parts of speech.
The facts are, then, that morphological and syntactical principles govern much speech and writing behavior and that many of them cannot be verbalized by a large number of adults and appear in the speech behavior of young children without explicit instruction. These features suggest that a psychological theory or account of morphology and syntax may have to include concepts of a higher order than those of stimulus and response and that explicit coding responses cannot be invoked to mediate these phenomena of grammar.
Structural linguistics has provided descriptions of these phenomena and has described the regularities that prevail in language. However, it has not presented a theory of the behavior involved. Recently, Chomsky (1957) and Miller (1962) have argued for a new conception.
This conception holds that, in addition to a phrase-structure grammar, there is also a set of transformational rules which, when used, modifies a statement from one type to another including the necessary morphophonemic changes. Thus, a sentence model exists which appears to be the basic form (the kernel) of adult utterances, but it is transformed systematically, as needed, by the use of transformational rules like the interrogative, the passive, the negative, and so on. More than one rule may be applied, as when a simple declarative sentence becomes a passive interrogative negative. For example, “Tom hits the girl” would become “Is not the girl being hit by Tom?” by the application of the appropriate transformations.
Chomsky, a linguist, has been explicit in asserting that a simple stimulus–response formation and habit utilization theory cannot, in his opinion, cope with this conception of grammar. Miller, a psychologist, seems to agree, and while he thinks work on association is important, he believes that the mechanisms underlying the transformational grammar can hardly be of the associational kind. A psychological theory of the development and use of transformational grammar has not been formulated, but if these authors are correct it would presumably involve high-order processes not readily reducible to association or meaning.
We have surveyed some of the major interests and problems characterizing psychologists’ work on language. Beginning with the historical trends in the psychological study of language, we have concentrated on mediational processes, conceived from both a meaning and a word-association point of view. While some progress on problems related to mediational processes has been made, many questions and further problems remain. Psychological understanding of language in the sense of connected words has not progressed very far, although concepts and methods of information theory and linguistics have been found useful. Whether the rule using that is evident in the verbal behavior of both children and adults, conceived either in terms of conventional morphology and syntax or in terms of transformational grammar, can be reduced to the concepts of contemporary learning theory is a question to be decided in the near future.
Charles N. Cofer
[Directly related are the entries Communication; Linguistics; Literature. Other relevant material may be found in Information THEORY; Perception, article on SPEECH PERCEPTION.]
BIBLIOGRAPHY
Bousfield, W. A. 1961 The Problem of Meaning in Verbal Learning. Pages 81–91 in Conference on Verbal Learning and Verbal Behavior, New York University, 1959, Verbal Learning and Verbal Behavior: Proceedings. Edited by Charles N. Cofer. New York: McGraw-Hill.
Brown, Roger 1958 Words and Things. Glencoe, 111.: Free Press.
Brown, Roger; and Fraser, Colin 1963 The Acquisition of Syntax. Pages 158–197 in Conference on Verbal Learning and Verbal Behavior, Second, Ardsley-on-Hudson, N. Y., 1961, Verbal Behavior and Learning, Problems and Processes: Proceedings. Edited by Charles Ne. Cofer and Barbara S. Musgrave. New York: McGraw-Hill.
Carroll, John B. (1953) 1961 The Study of Language: A Survey of Linguistics and Related Disciplines in America. Cambridge, Mass.: Harvard Univ. Press.
Chomsky, Noam 1957 Syntactic Structures. The Hague: Mouton.
Cofer, Charles N. 1957 Reasoning as an Associativ Process: III. The Role of Verbal Responses in Problem Solving. Journal of General Psychology 57:55–68.
Cofer, Charles N. 1960 Experimental Studies of the Role of Verbal Processes in Concept Formation and Problem Solving. New York Academy of Sciences, Annals 91:94–107.
Cofer, Charles N.; and Foley, John P. JR. 1942 Mediated Generalization and the Interpretation of Verbal Behavior: I. Prolegomena. Psychological Review 49:513–540.
Conference on Verbal Learning and Verbal Behavior, NEW YORK UNIVERSITY, 1959 1961 Verbal Learning and Verbal Behavior: Proceedings. Edited by Charles N. Cofer. New York: McGraw-Hill.
Conference on Verbal Learning and Verbal Behavior, SECOND, ARDSLEY-ON-HUDSON, N. Y., 1961 1963 Verbal Behavior and Learning, Problems and Processes: Proceedings. Edited by Charles N. Cofer and Barbara S. Musgrave. New York: McGraw-Hill.
Deese, James 1961 From the Isolated Verbal Unit to Connected Discourse. Pages 11–31 in Conference on Verbal Learning and Verbal Behavior, New York University, 1959, Verbal Learning and Verbal Behavior: Proceedings. Edited by Charles N. Cofer. New York: McGraw-Hill.
Dollard, John; and Miller, Neal E. 1950 Personality and Psychotherapy: An Analysis in Terms of Learning, Thinking and Culture. New York: McGraw-Hill. ? A paperback edition was published in 1965.
Goss, Albert E. 1961 a Early Behaviorism and Verbal Mediating Responses. American Psychologist 16:285–298.
Goss, Albert E. 1961 b Acquisition and Use of Conceptual Schemes. Pages 42–69 in Conference on Verbal Learning and Verbal Behavior, New York University, 1959, Verbal Learning and Verbal Behavior: Proceedings. Edited by Charles N. Cofer. New York: McGraw-Hill.
Jenkins, James J. 1963 Mediated Associations: Paradigms and Situations. Pages 210–245 in Conference on Verbal Learning and Verbal Behavior, Second, Ardsley-on-Hudson, N. Y., 1961, Verbal Behavior and Learning: Problems and Processes: Proceedings. Edited by Charles N. Cofer and Barbara S. Musgrave. New York: McGraw-Hill.
Krasner, Leonard 1958 Studies of the Conditioning of Verbal Behavior. Psychological Bulletin 55:148–170.
Luriia, Aleksandr R. 1961 The Role of Speech in the Regulation of Normal and Abnormal Behavior. New York: Liveright.
McGeoch, John A. (1942) 1952 The Psychology of Human Learning. 2d ed., revised by Arthur L. Irion. New York: Longmans.
Marshall, George R.; and Cofer, Charles N. 1963 Associative Indices as Measures of Word Relatedness: A Summary and Comparison of Ten Methods. Journal of Verbal Learning and Verbal Behavior 1:408–421.
Miller, George A. 1956 The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information. Psychological Review 63:81–97.
Miller, George A. 1962 Some Psychological Studies of Grammar. American Psychologist 17:748–762.
Noble, Clyde E. 1963 Meaningfulness and Familiarity Pages 76–119 in Conference on Verbal Learning and Verbal Behavior, Second, Ardsley-on-Hudson, N. Y., 1961, Verbal Behavior and Learning, Problems and Processes: Proceedings. Edited by Charles N. Cofer and Barbara S. Musgrave. New York: McGraw-Hill.
Osgood, Charles E. (1953) 1959 Method and Theory in Experimental Psychology. New York: Oxford Univ. Press.
Osgood, Charles E. 1961 Comments on Professor Bousfield’s Paper. Pages 91–106 in Conference on Verbal Learning and Verbal Behavior, New York University, 1959, Verbal Learning and Verbal Behavior: Proceedings. Edited by Charles N. Cofer. New York: McGraw-Hill.
Osgood, Charles E. 1963 Psycholinguistics. Pages 244–316 in Sigmund Koch (editor), Psychology: A Study of a Science. Volume 6: Investigations of Man as Socius: Their Place in Psychology and the Social Sciences. New York: McGraw-Hill.
Osgood, Charles E.; and Sebeok, Thomas A. (editors) (1954) 1965 Psycholinguistics: A Survey of Theory and Research Problems. Bloomington: Indiana Univ. Press.
Osgood, Charles E.; Suci, G. J.; and Tannenbaum, P. H. (1957) 1961 The Measurement of Meaning. Urbana: Univ. of Illinois Press.
Russell, Wallace A. 1961 Assessment Versus Experimental Acquisition of Verbal Habits. Pages 110–123 in Conference on Verbal Learning and Verbal Behavior, New York University, 1959, Verbal Learning and Verbal Behavior: Proceedings. Edited by Charles N. Cofer. New York: McGraw-Hill.
Skinner, B. F. 1957 Verbal Behavior. New York: Appleton.
Warren, Howard C. 1921 A History of the Association Psychology. New York: Scribner.
Woodworth, Robert S. (1938) 1960 Experimental Psychology. Rev. ed. by Robert S. Woodworth and Harold Schlosberg. New York: Holt.
II LANGUAGE DEVELOPMENT
Language development refers to the child’s acquisition of his first language, usually under informal natural conditions. By the end of the first four years of life, children have mastered the essentials of this most distinctively human attribute. Given normal hearing and a normal brain, variations in rates of mastery are small. Indeed, so gifted in language learning are they that many children become skilled in more than one language in their early years. The evolution of a child’s knowledge of his language is basic to his intellectual and social development, yet explanations of this process make demands on theories of human learning which have not yet been met.
Historical aspects
Studies of child language development in modern times fall into three phases. The first studies in the nineteenth century consisted primarily of parental diaries. Authors trained in linguistics (Gvozdev 1961; Leopold 1939-1949) have continued to use case studies. More than any other form of behavior, language reveals obvious internal patterning, and linguists have been loath to lose sight of these patterns by pooling quantitative measures of output. In the second phase of research, carried out by psychologists, standardized measuring methods of large samples were emphasized (McCarthy 1954). For the most part these studies have been atheoretical and have been concerned with variations in performance with age, sex, race, social class, and so on.
A landmark in modern research was Velten’s (1943) application of Jakobson’s (1941) theory of phonemic development. This theory proposed that changes in each child’s linguistic system followed an orderly sequence of increasing differentiation of significant features; thus it related language development to perceptual development and provided a theoretical framework which made comparison of children in different linguistic environments possible.
Subsequent research, particularly on phonology and grammar, is a product of the mating of psychology and linguistics. The differences from the preceding phase lie in the later work’s sensitivity to linguistic theories, respect for the systematic nature of the child’s linguistic knowledge, and emphasis on inferences about the structures or rules which underlie the observed situational changes in verbal behavior.
Prelinguistic phase of infancy
By the end of the first year of life, the average child understands some adult speech and can execute some directions. At first, he merely responds to adult sounds by attention and then by making sounds himself. Later he comes to recognize gross differences between voices and may distinguish intonational contrasts. His gradual discrimination of the critical features in the speech of adults is probably the most important aspect of language learning, yet almost nothing is known about the growth of comprehension.
The infant at first makes sounds which are closely tied to feeding and breathing. His vocalizations become increasingly triggered by vocal stimulation from others. In the first three months cooing comes to be highly associated with vocal stimulation and can also be increased in frequency if followed by other forms of adult responsiveness. On the other hand, the actual sounds uttered seem to be autogenous, since the sounds of deaf children and of children who hear are indistinguishable before six months.
Cooing gradually gives way to the consonant-vowel sequences and repetitive patterns which sound to adults like the syllables of their own language. During the babbling period, there is a rich variety of vocal play by the normal child. Sometimes he persists in making sounds which are idiosyncratic and not present in the speech around him. Further, he readily produces sounds and sequences in the adult repertoire which he will be unable to produce again for years after he begins to use language.
The role of babbling in language development is controversial. A child who lives in a linguistically stimulating environment may babble a great deal, and he may be more fluent than average when language develops. Yet, contrary to popular belief, there is as yet no evidence that babbling is in any specific way practice for language development. Indeed, the formal properties of babbling and of speech are so different that they suggest somewhat distinct central processes in their production.
Semantic system
Early sentences
An average child uses meaningful utterances by the middle of the second year, and by the middle of the third year babbling virtually disappears in play except as a stylistic device.
At first these utterances vary considerably in both sound and meaning. They may be quite idiosyncratic. Because of the simplicity of the child’s phonological system, there may be numerous homonyms if his vocabulary grows rapidly. Meanings are generalized readily, so that a given utterance might on various occasions mean “my coat,” “my hat,” “my baby carriage,” “let’s go for a walk,” “bye-bye,” and so on. Frequently there is a close integration with gesture.
If the first utterances are based on adult words, they may sometimes be very inclusive in reference (“bird”) or very narrow (“Bobby”). Adult beliefs about the semantic contrasts needed by children may influence the words they use in talking with children. From the standpoint of adult meanings, the words used by children tend to refer to animal and human movements and to concrete items with characteristic shapes and sizes (Brown 1958). When the child’s range of reference of a term is actually tested, it appears to depend on the variety of verbal contexts in which a word is heard and, to a lesser degree, on the variety of physical referents.
Conceptual changes
As the size of vocabulary increases, inevitably there is a change in the conceptual system. The number of semantic contrasts marked by vocabulary grows, and the referential range for each word must therefore narrow. A child might begin by calling all adults “papa,” next use the word for his own father, and then later learn the word “adult,” which has a similar referential range to the primitive “papa” (Leopold 1939-1949).
In this example, there has been more than a change in range. By the time the word “adult” is present, the child has a hierarchical system of superordinates. The criteria for each class have increased in specificity and changed in character, becoming less visual. Changes in the semantic system have in fact been only sketchily studied. They may include increases in the specificity of terms, increases in knowledge or in concept range as experience grows, a shift from sensorimotor to relational bases for concepts, and shifts in the verbal structure, so that antonyms, synonyms, superordinates, and other structural relations in the vocabulary reflect the critical contrasts employed in the language.
Studies of the deaf and other experimental research have suggested that conceptual development may have important nonverbal roots. Yet words give society a sure way of imposing its conceptual system. Societies differ considerably in the semantic contrasts marked by their vocabulary of kin, color, quantity, shape, time, space, and so on. The sanctions against deviant denotative use of these terms bring the child’s comprehension, speech, and presumably his conceptual system into line with his language community.
Verbal relations
The basis of the child’s system of relating words shifts by mid-primary school from sounds to meanings and grammatical features. The sound of a word is salient to the preschool child, who readily produces rhymes and alliterative sequences in word play. Words which sound alike are easily confused, leading to contaminations of meaning. On the other hand, children’s greater interest in the sounds of words may facilitate the learning of a spelling system or the sound system of a second language. In contrast, adults, who tend to respond less to the sound and who are concerned more with meaning, are known to have greater difficulty with the phonology of a second language, although they may acquire vocabulary far more quickly.
As children mature, their vocabulary increasingly becomes organized in terms of grammatical and semantic replacement classes, and they become more able to isolate words from their situational and sentence contexts. Each word more readily elicits other words. Interverbal associations are more rapid, more fluent, more specific, and more predictable from the semantic and grammatical system. This high rate of mutual evocation of words may be a help in intelligence tests and in school tasks requiring verbal retention of information. Children vary considerably in their readiness to convert their experience—including nonverbal experience—ipto spontaneous verbal thought. This skill seems to be a consequence of the verbal milieu in which the child is reared, and hence is responsive to training (see Ervin-Tripp in Hoffman & Hoffman 1966).
Phonology
As soon as a child knows several meaningful utterances it is possible to study his phonological system. The units of this system are phonemes, which are constructs that represent the smallest distinguishable unit of speech and which account for the significant formal contrasts between utterances. A very primitive system is illustrated when a child has two words, such as “baba” and “tata.” At this point there are three phonemes—one vowel and two consonants. Extrapolating from a small number of diary studies, we might expect the following changes to take place: (1) We might next find a system containing four consonant phonemes in which there is a contrast between the stop consonants and continuants. For example, the child might add /m/ and /n/, or /f/ and /s/ to the primitive system. At this point the consonant system would have two intersecting features—place of articulation (front versus back) and type of articulation (stop versus continuant). (2) In the vowels, probably a low versus high contrast will appear first, so the next vowel might be a single higher vowel, followed by a front-back contrast, as in /i/ versus /u/. (3) Typically, a contrast of position is followed by a voicing contrast. When voicing becomes significant, we might find both “dada” and “tata” with different meanings. (4) The consonant system will usually be more elaborate at the beginning of words, so that the number of contrasts may be greater there during the course of development. (5) Syllable repetitions, as in the example, usually decrease in the second year, and the length and variety of word-formation patterns increase. (6) Some sequential arrangements, such as consonant clusters (ts, kr, pi), are absent for a long time, even when the component phonemes are present (see Ervin-Tripp in Hoffman & Hoffman 1966).
This has been a description of the child’s phonology in its own terms, but most observers are more likely to notice the mapping of the adult’s system onto the child’s when the child makes substitutions. Thus, in the above system with three phonemes and a CVCV syllable repetition pattern, any word acquired by the child must become either “tata” or “baba”—an extreme case to be sure. The most obvious substitution patterns are simple replacements, such as the mapping of both stops (t, d) and affricates (ch, j) onto stops when there is no stop-affricate contrast. The word formation pattern of the child may require omission of whole syllables or members of clusters. The conversion of “father” into “papa” is predictable for a child who has not yet developed a stop-fricative contrast (between /p/ and /f/), who has a CVCV syllable repetition rule, and who selects the adult stressed syllable.
The child with a simplified phonemic system sometimes has a preferred phonetic realization of a given phoneme and sometimes has random or free variation. In the first case, he might always say [t] and never use [d] in articulation of “tata.” In the latter case, especially common for vowels, he may oscillate unpredictably. A third possibility is that he may use each predictably in a given context. For instance, he may use [d] at the beginning of words and [t] in the middle to realize a single phoneme (Velten 1943).
Although the number of phonemes in children’s systems usually is less than in adults’, their substitution rules are not always simple and may baffle parents or teachers unfamiliar with phonemic analysis. For instance, some children have a rule making an initial consonant nasal if there is a nasal consonant anywhere in the adult model word, producing imitations like /nil for “green.” There may be phonemic contrasts in the child’s system which are absent in the adult system. Assimilation is very common in children’s word-formation patterns, so that neighboring phonemes or successive syllables may influence each other. Yet all of these are orderly rules. The factors producing these idiosyncratic patterns are as yet unknown.
What makes a child’s phonemic system change? Perhaps the presence of numerous homonyms encourages change; yet some children tolerate extensive homonymy when they have a large vocabulary combined with a very simple phonemic system and restricted word-formation rules. Understanding or hearing a phonemic distinction is not a sufficient condition for producing it, although each process evidently facilitates the other. Thus, in adults, acoustic discrimination is much sharper when the listeners have learned a corresponding discrete articulation. The teaching of reading and spelling and the alteration of phonology when children learn to speak second dialects and second languages might be much easier if teachers knew more about phonological development [see Perception,article on Speech Perception].
Grammar
Languages differ considerably in their organization of basic grammatical devices, so it will be necessary to study children’s language development in many types of languages to establish well-founded generalizations. For example, in English, the constituent order (e.g., subject-verb-object) has a high degree of regularity; deviations are significant; and the selection of entries in the subject and object position is semantically important. In Russian, on the other hand, there may be much more variability in constituent order, since inflectional suffixes mark the relations indicated by constituent order in English. Children learning both languages show order regularities before they use markers such as inflection or function words consistently, but children learning highly inflected languages have been reported to learn inflections much earlier (Gvozdev 1961).
Simple syntax
From the very beginning of multiword sentences, children usually reveal order regularities in speech. Certain frequent words may occupy fixed positions. For instance, “where,” “there,” and “this” may always be in antecedent position, and “up,” “on,” and “off” in last position. Other words and phrases, such as “broken,” “blue,” “truck,” “the truck,” “the broken,” then could occupy either position in complement to the fixed items. On the basis of position alone, one can identify primitive word classes and usually a nominal phrase also, which can expand one of the complement classes. From the examples it can be seen that these classes are idiosyncratic and do not correspond exactly to adult classes (Bellugi & Brown 1964).
How do primitive classes develop? They cannot be based simply on imitation, since at this stage children’s imitations are at least as simple as their own spontaneous speech, and classes appear even if the adult language has variable order (Gvozdev 1961). They reflect order regularities when they exist in adult language, and in part they reflect semantic features of adult classes (Brown 1958). In their idiosyncratic features they clearly reveal creative and analogizing activity by the child and cannot be based on rote learning. The study of the child’s interpretation of what he hears may provide the key to this enigma. But as yet there have been no true experiments on grammatical acquisition in the very young. There is disagreement as to the character of “rules” in children’s early grammar and as to the relation between the comprehension system and speech.
Inflections and function words
During the period of simple syntax, English-speaking children employ function words like “the” at random, and as their frequency increases such anomalous sentences may appear as “I see the Mary.” The order of mastery of both function words and inflections is influenced by their semantic obviousness.
In English, inflectional suffixes such as plurals are randomly present at first, but in highly inflected languages, such as Russian, often a single form is used before the period of random variation. In many languages there are some irregular inflections, such as “go-went,” which are preserved because of their high individual frequency. Children do not usually learn these first, as tense contrasts at least, even though they are frequent. Among English nouns and verbs, the largest variety of verbs have an inflectional suffix, such as “stop-stopped.” It is these forms which are productive or generalized to new words by children. These analogies can be highly resistant to adult influence; typically, a form like “foots” can remain in a child’s speech for months after he has learned the plural form.
Sometimes alternative inflections depend upon the preceding phoneme; for example, a plural is formed by “-s” in most cases, but by “-es” in some, as in “mats” versus “matches.” Children use the most common form first, then both at random, before discovering the phonemic conditioning rule. During the period of random variation one finds forms such as “handses” and “toasteded” (Ervin-Tripp in Hoffman & Hoffman 1966).
Complex syntax
It is a common observation that children’s sentences grow longer as they get older. This is a superficial effect of deeper changes. As we have seen, this increase is due in part to the systematic use of function words and suffixes. Some of the constituents which were optional in the simple grammars—such as verbs—may become obligatory in the more complex phrase structure. Potentially infinite expansions of coordinate and subordinate structures appear. These are all changes in the child’s syntax itself; in addition, there may be increases in the amount of information a child can include in a given utterance. One might say that programming capacity increases with age.
In the course of these changes, it is not easy to establish criteria for mastery of the adult syntactic rules. Sometimes a given sentence in a child’s speech has the appearance of an adult sentence, but it derives from a different syntax. For example, during the early stages in the development of negation, children commonly use “no” or “don’t” as simple negativizing words, much like the plural. Later, as inflection of verbs appears, one may find “he doesn’t goes.” Eventually the child will arrive at the adult rule, which in the negative marks tense and person only in the auxiliary. The intermediate stages of the evolution of the negation rule can be inferred if one has appropriate tests or a sufficiently large text from a child and is willing to make such inferences.
At the age of six, there remain a number of constructions which are still absent in the speech of many children, such as nominalizations of simple verbs (Menyuk 1964). Some of these constructions may be rare in the colloquial speech used to chil dren. Perhaps, on the other hand, they involve inherently more complex rules requiring conceptual maturation.
Children’s syntactical rules are always inferred from behavior, but the study of comprehension, imitations, spontaneous speech, or other measures may give different results. Comparison of these varying measures may help in the analysis of such inferred processes. Each makes different demands on knowledge, much as recognition and recall make different demands on memory. Imitation, for example, requires phases of perceptions, storage, retrieval, and speech. At different ages imitations are structurally quite dissimilar. At first they appear to be markedly simplified by the child’s speech production rules. Later, restrictions occur in the other phases of the process, but eventually the child can perceive and store even the mistakes which he hears (Ervin-Tripp in Hoffman & Hoffman 1966).
Language functions
Private functions
Vocalizing begins as an autogenous activity, and throughout life there exist varieties of speech in which listeners do not actively mediate the speaker’s satisfactions. In babies these may include cooing, babbling, and at first, crying. Vocal play with sounds continues into the phase of organized speech. Babies vocalize when they see objects and when they engage in action, but around four and five such speech appears to have an implicit audience. Such monologues, for instance, decrease in isolation, are inhibited by strangers, and increase when difficulties are encountered (Vygotskii 1934). Eventually speech not only is a product of play but aids in its organization, perhaps being a precursor, as Vygotskii believed, of some aspects of covert thought.
Social functions
Social play—for example, in gestural games—may be at least as significant in the evolution of early language as demands for goods and services. Such play receives social rewards, and even a three-month-old infant may vocalize more if so reinforced. It thus becomes difficult to mark any point when one can say that vocalizing becomes intentional. The rhetorical question, ritual naming of objects, and many other vocal performances in the second year appear to be socially motivated and probably vary markedly according to milieu. True offering of information, addressed to lacunae in the listener’s knowledge, demands mature social development, and difficulties can still be found in this function during the school years. Because of the heavy dependence of these different functions on social milieu, it is particularly difficult to establish any generalizations in the absence of experimental research.
Children’s speech changes according to the audience, the function, and the setting. The most extreme adaptation occurs in multilingual children, who by three can usually distinguish their languages appropriately. Yet even within a language there may be formal alternations, as in the suffixes “-in” and “-ing” in English, which mark subtle style shifts in the speech of American children.
In many languages there is a special form of speech used in addressing infants, which may be called baby talk, and involves changes in phonology, vocabulary, and syntax. Children may employ baby talk as a stylistic device depending on the listener or the role during play. In languages such as Japanese there may be many changes in syntax or vocabulary according to the age or status of the listener, and these are acquired by children only gradually during the school years. Optional features of syntax, such as use of qualifiers and subordination, may be affected by audience or function. Conversations with adults or expository speech in the schoolroom may draw on syntactic skills not usually employed in conversations with other children.
Schooling
Schooling may indeed mark the most radical change in language functions encountered by children. In the home and among friends the function and form of speech may be quite different—briefer, situationally imbedded, narrative, repetitive, topically limited, aimed at solidarity or aggression. In school, there is a demand for precision, independence from the speech situation, exposition, novelty, breadth of topics, and transmission of information. The latter functions demand more abstract vocabulary and greater use of grammatical structures that signal analytic differentiations. Written language, since it is private, with neither audience nor situational support, is at the farthest remove from the colloquial practice children bring to school (Vygotskii 1934).
Knowledge of language increases, of course, after early childhood. The everyday vernacular of the peer group, not parents, in the preadolescent years remains as our spontaneous adult speech. We acquire new vocabulary from travel, reading, and the acquisition of skills. Later, as our social relationships become more complex we may come to control a greater range of the structural facilities of the language and its stylistic alternatives for such varied purposes as persuasive, expository, and aesthetic discourse. But these changes usually do not occur in children who lack any secondary education. While it is true that all normal children master the essentials of language before they enter school, variations in the language functions provided by the family and school milieu affect markedly the extent to which children go beyond these essentials.
Susan M. Ervin-Tripp
[Directly related are the entries Learning,article on VERBAL LEARNING; LINGUISTICS.Other relevant material may be found in Developmental PSYCHOLOGY; Hearing.]
BIBLIOGRAPHY
Bellugi, Ursula; and BROWN, ROGER W. (editors) 1964 The Acquisition of Language. Monographs of the Society for Research in Child Development, Vol. 29, No. 92. Univ. of Chicago Press.
Berko, Jean; and Brown, Roger W. 1960 Psycholinguistic Research Methods. Pages 517–557 in Paul Mussen (editor),Handbook of Research Methods in Child Development. New York: Wiley.
Brown, Roger W. 1958 Words and Things. Glencoe, 111.: Free Press.
Carroll, John B. (1957)1960 Language Development. Pages 744–752 in Encyclopedia of Educational Research. 3d ed. New York: Macmillan.
Gvozdev, A. N. 1961 Voprosy izucheniia detskoi rechi (Problems in the Language Development of the Child). Moscow: Akademiia Pedagogicheskikh Nauk RSFSR.
Hoffman, Martin L.; and HOFFMAN, Lois W. (editors) 1966 Review of Child Development Research. Volume 2. New York: Russell Sage Foundation. ? See especially the article by Susan Ervin-Tripp on “Language Development.”
Jakobson, Roman 1941 Kindersprache, Aphasie, und allgemeine Lautgesetze. Uppsala (Sweden): Almqvist & Wiksell.
Leopold, Werner F. 1939-1949 Speech Development of a Bilingual Child: A Linguist’s Record. Studies in the Humanities, Nos. 6, 11, 18, 19. Evanston, 111.: Northwestern Univ. Press.
Leopold, Werner F. 1952 Bibliography of Child Language. Evanston, 111.: Northwestern Univ. Press.
Lewis, Morris M. (1936) 1952 Infant Speech: A Study of the Beginnings of Language. 2d ed., rev. New York: Humanities.
Lewis, Morris M. (1963)1964 Language, Thought and Personality in Infancy and Childhood. New York: Basic Books.
McCarthy, Dorothea 1954 Language Development in Children. Pages 492–630 in Leonard Carmichael (editor),Manual of Child Psychology. 2d ed. New York: Wiley.
Menyuk, Paula 1964 Syntactic Rules Used by Children From Preschool Through First Grade.Child Development 35:533-546.
Templin, Mildred 1957 Certain Language Skills in Children: Their Development and Interrelationships.Institute of Child Welfare, Monograph No. 26. Minneapolis: Univ. of Minnesota Press.
Velten, H. V. 1943 The Growth of Phonemic and Lexical Patterns in Infant Language.Language 19:281-292.
Vygotskii, Lev S. (1934) 1962 Thought and Language. Cambridge, Mass.: M.I.T. Press. ? First published as Myshlenie i rech’.
III SPEECH PATHOLOGY
Speech pathology is a branch of communications dealing with disorders of speech and language. Speech pathologists concern themselves with diagnosis, treatment, and prevention of speech and language disorders. Although speech pathology in Europe was once a branch of medicine, today, throughout most of the world, the field has developed as an independent speciality, in somewhat the same fashion that clinical psychology has emerged as an area separate from psychiatry. As is the case with clinical psychology, most training programs in speech pathology in the United States are a part of the graduate program offered by colleges of arts and science. A limited number of training programs are affiliated with colleges of education or colleges of medicine.
Unlike the general field of speech, which is concerned with improving the normal and slightly subnormal, the area of speech pathology focuses on defects of speech and language. Speech is considered to be defective when any one or combination of the following conditions exists: speech is lacking in intelligibility; speech differs so much from normal that it calls undue attention to itself, often with the result that listeners pay more attention to the speech deviations than to what the speaker is saying; speech behavior leads to the development of negative attitudes by the speaker toward his own speech, which in turn often interfere with his over-all adjustment.
Speech and language problems can be described in terms of the acoustic end product perceived by listeners or in terms of the etiology of the problem.
Problems of the acoustic end product
From the standpoint of the acoustic end product, speech and language problems can be considered under five general headings: delayed speech and language, defects of articulation, defects of voice, defects of rhythm, defects of symbolic formulation.
Delayed speech and language
Most normal children are able to speak their first word at approximately age one. Language comprehension and production increase along with articulation ability, By age eight the majority of children have learned to articulate all sounds correctly. Speech pathologists use the term delayed speech to describe the speech of children who either do not talk or who per sever ate habits and patterns of infantile speech. If there is vocabulary deficiency, inadequate formulation of ideas, or retarded sentence structure, the term delayed language may be employed (Matthews 1957 a, pp. 394–395). A well-accepted listing of causes is provided by Van Riper (1963) and includes mental retardation, hearing loss, faulty coordination resulting from disease or paralysis, pro-longed illness during infancy, lack of motivation for speech, improper teaching methods employed by parents, confused hand preference, necessity for learning two or more languages simultaneously, shock during the speaking act, emotional conflicts, and aphasia.
The treatment of delayed speech depends upon the significant etiological factors. The child with a hearing loss may need a hearing aid. Other causative factors should be removed or lessened. Van Riper describes therapy procedures for delayed speech. In general these procedures attempt to provide a stimulating environment in which the child is “bombarded” by speech and is taught to associate speech sounds with meaningful people and objects.
Defects of articulation
Defects of articulation in general can be divided into four types: substitution, distortion, omission, and addition. Substitution errors consist of the replacement of one sound with another. A common example is the substitution of w for r, as in “wed” instead of “red,” “twuck” for “truck,” and “dwink” for “drink.” Other common substitution errors are w for /, as in “weave” for “leave” th for s, as in “thick” for “sick”; and d for th, as in “dis” for “this.”
The second type of articulation error involves sound distortion. This can be observed in the s sound when it is faultily produced with a whistling component, nasal air escape, or slushiness. Distortions in the pronounciation of the English r are often produced by speakers who have learned English as a second language.
The omission type of articulation error is illustrated by the individual who omits the r sound and says “ed abbit” instead of “red rabbit.”
The addition type of articulation error is seen in the inclusion of an r sound where it is not called for, as in “idear” for “idea.”
Articulatory disorders are sometimes classified as functional or organic in terms of possible causal factors. Matthews (1957b) has summarized the literature relating mental retardation and articulation disorders. A frequent cause of articulation disorders is faulty training resulting from inadequate speech environment. If a child’s speech model is defective, he is likely to learn defective speech. (A detailed discussion of etiological factors in functional articulation disorders can be found in Van Riper 1963; Johnson et al. 1948; and Berry 0026 Eisenson 1956.)
Although the majority of articulatory disorders are of functional origin, there are organic factors that can adversely affect articulation. West and his associates (1937) discuss in detail articulation disorders resulting from lesions of the central nervous system. Cleft palate, dental anomalies, structural disorders of the peripheral speech organs, and hearing loss can cause articulation disorders.
Where a structural abnormality contributes to an articulation disorder, the organic defect may need correction prior to speech therapy. Sometimes speech therapy may proceed in conjunction with the correction of the organic disorder. Often, complete correction of the structural anomaly is not possible but speech therapy may prove to be beneficial.
The correction of articulatory errors requires teaching the production and habitual use of sounds. As a rule the sound is taught first in isolation, then in combination with other sounds, and finally in words, sentences, and conversation. Milisen describes a variety of approaches to articulation therapy (Milisen et al. 1954).
Defects of voice
Defects of voice can include complete absence of voice as well as problems of pitch, volume, and quality. Voice may be lacking because of total or partial paralysis or removal of the vocal folds, or the vocal structures may be normal and the absence of voice may be a reflection of a hysterical condition.
Pitch may be inappropriate for the age and sex of the speaker. This would be illustrated by a male talking in an extremely highpitched voice or a female talking with low pitch. Monotone consists of a sameness of pitch. The failure to vary pitch can not only result in a speech pattern that is dull and uninteresting to listen to; it can also interfere with speech intelligibility, inasmuch as certain aspects of meaning are conveyed by means of variations in pitch.
Just as pitch may be inappropriate, it is possible that volume, also, may be inappropriate. Volume may be so weak that it is difficult for listeners to hear what is being said, or it may be so loud that listeners experience a certain amount of discomfort. Normal speakers do not use exactly the same volume from beginning to end of a message. Just as there are normal variations in pitch, so there are normal variations in volume. The absence of these variations in volume would be considered a type of voice disorder.
The third general type of voice disorder consists of defects in quality. These defects are difficult to describe and should be heard to be fully appreciated. Defects of voice quality sometimes are labeled with such terms as “nasal,” “hoarse,” “husky,” “breathy,” “harsh,” etc. The quality is considered defective when it is not only unpleasant to listen to but seriously detracts a listener’s attention from the message of the speaker.
As is the case with articulatory disorders, there are both functional and organic causes of voice disorder. Imitation of poor speech models, psychological maladjustments, adolescent voice change, poor breathing habits, as well as laryngeal pathology, paralysis, and adenoidal obstructions, appear as causes of voice disorders.
In the case of any type of laryngeal pathology, voice therapy should not be attempted until a medical evaluation and laryngeal examination have been completed. Voice training often includes auditory training, to make the client aware of the differences, in both sound and “feel,” between proper and improper use of the voice. Where the voice problem is related to emotional disturbances of any kind, psychological help may be necessary. In those instances where the vocal mechanism has been removed, the client may use an electrical or mechanical vibrator as a substitute for vocal-fold vibration or the client may learn esophageal speech, a technique that involves using air expelled from the esophageal tract as a source of sound vibration.
Defects of rhythm
Defects of rhythm can be described as repetitions, prolongations, or hesitations. Individual sounds, words, or phrases may be involved. Frequently the disruption in rhythm may be accompanied by facial grimaces, tics, or other bodily movements, which in time can become an intimate part of the disruptions of rhythm. Very frequently the disruptions in rhythm and the accompanying grimaces carry with them considerable fear and many negative attitudes on the part of the speaker. The speaker’s fear that he will have problems with his speech often contributes to additional difficulties with speech rhythm.
Stuttering. More has been written on the topic of stuttering than on any other single disorder of speech. Matthews (1957a), in a brief summary of theories of stuttering, cites representatives of three broad theoretical viewpoints: (1) dysphemic theories, which suggest that stuttering is related to constitutional abnormality; (2) personality theories, which hold that stuttering is related to psychological maladjustment; and (3) developmental theories, which suggest that stuttering develops largely as the result of environmental conditions.
Therapy approaches are influenced by the therapist’s theory concerning the etiology of stuttering. Those who view stuttering in terms of a general neurosis will employ some form of psychotherapy. A therapist who sees stuttering as the result of lack of cerebral dominance will seek ways of establishing a dominant gradient of excitation in the central nervous system. Environmental modification is the goal of many therapists, regardless of the theory of causation of stuttering they subscribe to. Van Riper (1963) not only summarizes much of the literature on etiology and treatment of stuttering but also outlines in detail a therapy approach widely employed in the United States. His therapy seeks to help the stutterer stop reinforcing his stuttering. An important aspect of this therapy consists of eliminating the stutterer’s avoidance of feared situations and words.
Defects of symbolic formulation
The defects of symbolic formulation may be of both an expressive and a receptive type. Although defects of symbolic formulation are often categorized as either expressive aphasia or receptive aphasia, in actuality most patients who have difficulty with symbolic formulation have some difficulties in both the expressive and the receptive realm. Individuals with receptive aphasia may be able to hear a speaker say the word “chair” but will not be able to translate the sounds in this word into the concept of a piece of furniture on which a person may sit; they may be able to see and to recognize each of the five letters used in writing the word “chair” but be unable to translate the written letters into the concept of chair. The individual with an expressive type of aphasia may know what a chair is and be able to pronounce all of the sounds in the word “chair” but be unable to put these sounds together so they become a recognizable symbol of the concept “chair.”
The treatment of aphasia involves reeducation and retraining, which utilizes the past speech background of the patient as much as possible. A detailed discussion of therapy for aphasics can be found in Wepman (1951).
Negative attitudes and frustrations
Often the individual with a speech problem is handicapped by more than just the acoustic end product of his speech. Frequently the speaker’s attitude toward his speech constitutes one of his most serious problems. In some instances the speaker may actually talk with a fairly high degree of intelligibility. However, if the speaker’s attitude toward his own speech is apprehensive and fearful, he may experience a handicap far beyond what would normally result from the faulty articulation or faulty voice quality alone.
A speech problem not only can interfere with defective communication in social development but also can lead to the frustration of not being able to make oneself understood and of constantly being made to feel different. Such feelings of frustration, difference, and inferiority can lead to personality problems and attitudes which may be more serious than the original speech problem itself. For this reason, speech pathologists must be interested in the total adjustment of the person who has a speech problem.
Etiology and therapy
If speech problems are examined from the standpoint of etiology, they can be categorized as organic or functional. In actuality it is extremely difficult to fit all speech problems neatly into one or the other of these two broad classifications. The two classifications are presented because they frequently appear in the literature.
Organic and functional factors
Speech problems arising from primarily organic causes would include speech associated with cleft palate, cerebral palsy, dental abnormalities, brain damage, hearing loss, or any other type of anatomical or neurological involvement affecting any of the mechanisms used in speech production.
Etiological factors of a nonorganic type would include lack of stimulation to speak, withdrawal tendencies associated with emotional disturbance, failure to have available good models of speech for imitation, mental retardation, faulty learning, etc. The organic and functional aspects frequently are intertwined. This can be seen in the individual who is aphasic because of brain damage. Because of his aphasia, he experiences considerable frustration. This frustration in turn leads to withdrawal behavior. Frustration and withdrawal behavior contribute to further deterioration of speech. Frequently it becomes difficult to separate the components of the speech problem that are of organic causation from those that represent some sort of functional overlay.
Therapy. The speech pathologist initiates therapy after a thorough diagnosis has been performed. This diagnosis often must include a medical examination, psychological assessment, and determination of hearing acuity. Dental and social-work information are often necessary before an adequate diagnosis can be made. Wherever possible, causative factors are removed or minimized. In the case of cleft palate, for example, surgical or dental procedures may be necessary to provide a mechanism adequate for speech production. In some instances it may be impossible to provide a mechanism that is completely normal. Under these circumstances methods must be found to compensate for the structural inadequacies. In the case of a severe hearing loss, it may be possible to provide sound amplification in the form of a hearing aid.
Articulatory disorders constitute the largest group of speech problems encountered in children. The majority of articulatory problems are not caused by anatomical, neurological, or other organic factors. Most articulatory problems can be traced to lack of stimulation, poor speech models, faulty learning, and other factors quite far removed from defects of the oral structures. An early step in treatment consists of describing adequately the nature of the speech deviation. The speech pathologist helps the individual with the speech problem to understand the detailed nature of the problem. In some cases this consists of the speech pathologist’s helping a client become more aware of his speech errors. A client must learn to distinguish between the correct and incorrect speech productions. A wide variety of procedures will be employed to help the client articulate a sound correctly, to eliminate an undesirable component in voice quality, to produce a pitch level more appropriate to the age and sex of the client, etc. Frequently a client learns to make correct speech productions in isolated sounds or words but finds it difficult to carry over these new patterns into everyday speaking situations. Frequently the newly acquired speech skills result in a pattern of speech that sounds strange to the client. The speech pathologist often spends considerable time helping the client understand his feelings toward his speech problem, as well as his feelings toward the new speech patterns which he acquires. Often the attitude of a parent toward a child and toward the child’s speech is a contributing factor. For this reason, speech pathologists frequently spend a good deal of their time in parent-counseling activities.
Because some voice disorders are associated with organic pathology, the speech pathologist may carry out some of his treatment procedures in cooperation with an otolaryngologist. Such collaborative therapy activities may be carried out with psychiatrists, clinical psychologists, plastic surgeons, and dentists, as well as classroom teachers. The speech pathologist cannot limit himself to a consideration of the mechanical aspects of sound production. His interest must be in the person with a speech problem. The modern practice of speech pathology makes comparatively little use of the tongue, lip, and jaw exercises that were prevalent in the field a decade ago.
Training in speech pathology. Standards for the training of speech pathologists have been established by the American Speech and Hearing Association. The association awards a certificate of clinical competence to individuals who successfully complete the requirements for a master’s degree in the field of speech pathology. This graduate program includes didactic classroom instruction, supervised clinic practice, and four years of successful practice, under supervision, in an environment where speech and language problems are diagnosed and treated. Graduate training programs in speech pathology in the United States are accredited by the American Board of Examiners in Speech Pathology and Audiology of the American Speech and Hearing Association. The association is a clearinghouse for information regarding clinical facilities, training institutions, and other data relative to the field of speech pathology.
Jack Matthews
[See alsoMental Disorders, article on Organic Aspects. Other relevant material may be found inHearing; Language, article onLanguage Development; Mental Retardation; Perception, article On Speech Perception.]
BIBLIOGRAPHY
Berry, Mildred F.; and Eisenson, Jon 1956 Speech Disorders. New York: Appleton.
Johnson, Wendell et al. (1948) 1956 Speech Handicapped School Children. Rev. ed. New York: Harper.
Matthews, Jack 1957a Speech Defects. Pages 391–424 in Chauncey M. Louttit (editor), Clinical Psychology of Exceptional Children. 3d ed. New York: Harper. ? First published in 1936.
Matthews, Jack 1957b Speech Problems of the Mentally Retarded. Pages 531–551 in Lee E. Travis (editor), Handbook of Speech Pathology. New York: Appleton.
Milisen, Robert et al. 1954 The Disorder of Articulation: A Systematic Clinical and Experimental Approach. Journal of Speech and Hearing Disorders Monograph Supplements, No. 4.
Van Riper, Charles 1963 Speech Correction Principles and Methods. 4th ed. Englewood Cliffs, N.J.: Prentice-Hall. ? First published in 1939.
Wepman, Joseph M. 1951 Recovery From Aphasia. New York: Ronald Press.
West, Robert W.; Ansberry, Merle; and Carr, Anna (1937) 1957 The Rehabilitation of Speech. 3d ed. New York: Harper.
IV LANGUAGE AND CULTURE
The key role of language in all human activities has made it perhaps inevitable that the field of linguistics should represent a mingling of several streams of interest. Modern linguistics has arisen from the philological tradition, concerned basically with the classical and modern written languages, and from the anthropological tradition, which has been concerned largely with preliterate peoples. The anthropologist has long recognized the importance of language, not only as a tool for more effective field work, but as a critical element of the cultural fabric which he studies. Thus we sometimes refer to “anthropological linguistics”, which may be defined as the study of previously unknown speech varieties in the context of their cultures; the term contrasts the anthropological approach to language with philological, psychological, or philosophical approaches. Alternatively, we may wish to speak of “linguistic anthropology,” focusing attention on language as one element of human culture; the term is analogous to “social anthropology,” “economic anthropology,” and the like. The older term “ethnolinguistics” may well be used to refer to the same area of interest.
All writers in this field have struggled with the expressions “language and culture” versus “language in culture,” both of which are in common use as titles for university courses, scholarly symposia, etc. “Language and culture” seeems to imply a dichotomy, which we must then reject in the light of our position that language is part of culture. But if we speak of “language in culture, “we then lack a separate name for all the other cultural areas whose relationship to language we wish to study. Perhaps the best solution is to give formal recognition to what has most often been done in practice and to use the word “culture” in two ways, on two different levels of a semantic hierarchy. Distinguishing these meanings by subscript numerals, culture1, on the higher level of generality, constitutes learned patterns of human habitual behavior. Language is included along with everything else that contrasts with instinctive behavior. Culture2, on a more specific level, is that part of “culture1” which is not verbal communication; in this sense, “culture” contrasts with “language.” In most cases, the context of discussion will make it clear whether we are referring to “culture1” or “culture2,” just as context normally eliminates confusion between “man1” (opposed to “animal”) and “man2” (opposed to “woman”).
Taking the view that language is part of culture, linguistic anthropologists have been concerned with these basic questions: In what respects does language fit into the general conception of cultural systems, and in what ways is it distinguished from other components? What similarities are there between the internal structures of language and of other branches of culture? What role does language play in the over-all functioning of culture? In what way do language and culture reflect each other’s structure at a given point in time or influence each other over the span of history? What techniques may we use to infer linguistic from nonlinguistic behavior, or vice versa, either in terms of predicting the future or of reconstructing the past? At this moment, most of these questions still lack definitive answers; the rapid growth of ethnolinguistics, however, suggests that the near future will bring, if not answers to all questions, at least a more unified framework for discussion.
The cultural nature of language . Language is assured a position as a branch of culture by its distinctively patterned nature, by its restriction to the human species, and above all because languages are learned, not transmitted genetically. In spite of the fact that race and language frequently have a historical connection—so that many people who share ancestors also share a common language—such connections are in no way necessary. The nongenetic transmission of languages is vividly demonstrated by the linguistic “melting pot” of the United States, in which people of the most diverse racial backgrounds share common standards of English usage. However, the fact that individual languages are transmitted culturally, not genetically, does not rule out the possibility that mankind has certain unique inborn capacities for linguistic behavior. For some purposes, we may distinguish between language, an inherited set of capabilities, and languages, particular structures which are built on those capabilities by culture.
The distinctiveness of language . Language obviously stands apart from other communication systems used by humans or animals because of the magnitude of its resources. It is especially impressive to consider that every normal child, by the age of four or five, is capable of using the language of his community to produce a literally infinite number of meaningful utterances. We are far frorh understanding all of the characteristics of language or of the human nervous system which make this possible. Two things, however, are clearly important—man’s ability to invent symbols and the duality of patterning in linguistic structure.
If we understand a sign to be anything from which the existence of something else may be inferred, then we may define a symbol as a special kind of sign—one with arbitarary, conventionally assigned meaning. Thus, black clouds are a sign of rain, the relationship being intrinsic; but a particular weather flag, as a conventional sign of rain, is a symbol. By the same token, the word “rain” is a symbol; our use of this particular word is conventional and subject to change. Other animals may learn to respond to many arbitrary signals, including words of human language, but it is uniquely human to have the ability to assign arbitary meaning to signs, i.e., to invent symbols.
But language goes beyond other symbolic systems, such as those of gestures, in one very specific feature—the duality of patterning. The meaningful symbols of language—such as words and meaningful parts of words, called morphemes— are not indivisible, like a flag or a gesture, but are themselves built up of smaller units. These smaller units are the phonemes or sound units of spoken language, and they are meaningless in themselves. Every language uses a small number of these meaningless units—usually less than fifty—to build up a huge number of meaningful units. It is this two-level structuring which gives language a degree of efficiency that is qualitatively superior to, not merely quantitatively different from, other communication systems.
Similarities between language and culture . The identification of such building blocks of language as the phoneme and the morpheme has given linguistics great prestige among the branches of anthropology; it is sometimes said that linguists are the only social scientists to have identified the basic units of their subject matter. The method used in this process of identification is one which moves from the level of observation to the level of structure. First, raw data are classified in terms of a universal taxonomic grid; in studying sound systems, this is the phonetic classification. Then the investigator finds that some phonetic differences, in particular languages, are not associated with contrastive meaning; e.g., the meaning of the Spanish día (“day”) is the same whether the initial d is pronounced as an occlusive (completely blocking the flow of air with the tongue and then releasing it), or as a fricative (letting air issue continuously between the tongue and the teeth). In every language, however, the linguist also finds that some phonetic differences are correlated with differences of meaning; e.g., the difference between occlusive and fricative, although nonsignificant in Spanish, is contrastive in English, serving to distinguish “day” from “they.” The result of such observations is the replacement of phonetic classifications by phonemic classifications, unique for each language. The phoneme is defined simultaneously by the range of noncontrastive sound differences which it subsumes and by the contrasts which it displays with the other phonemes of the system.
This method of qualitative contrast, first applied in phonological study, has been successfully extended to the identification of morphemes, i.e., grammatical units; and many scholars have speculated about their applicability to other areas of culture. The terms etic and emic have been coined (after “phonetic” and “phonemic”) to refer to the observational and structural levels, respectively, which might be distinguished in such areas as kinship, religion, music, art, and folklore. Such studies are still in their infancy, but they constitute one of the most interesting frontiers of anthropology, based as they are on the assumption that each branch of culture, or indeed culture as a whole, is, like language, an internally cohesive system.
The role of language in culture . Language is not merely one of several aspects of culture: it is, at the very least, prima inter pares, in that it makes possible the development, the elaboration, the transmission, and (particularly in its written form) the accumulation of culture as a whole. One can imagine handicrafts being taught by one generation to the next without the use of language; but social, legal, religious, political, or economic institutions are another matter. It is hard to imagine that a community of deaf-mutes (if they were deprived of such speech surrogates as writing) could carry on human social life.
But how, exactly, does language (or any other symbolic system) relate to experience? It is commonly said that symbols, like signs in general, “stand for” or “mean” something else. The definition of meaning itself clearly cannot be taken for granted. A variety of theoretical models for the concept of meaning, each one valuable for its own ends, has been proposed by philosophers, psychologists, and linguists of various persuasions. The model presented below is not intended to compete with others in defining the “real” nature of meaning, but it may be useful as a framework for ethnolinguistic discussion.
Structural linguists have customarily been extremely cautious in semantic matters, sometimes attempting to exclude them from linguistics altogether. Until very recently, a strictly behaviorist conception of meaning was much in vogue:
We have defined the meaning of a linguistic form as the situation in which the speaker utters it and the response which it calls forth in the hearer. . . . The situations which prompt people to utter speech include every object and happening in their universe. In order to give a scientifically accurate definition of meaning for every form of a language, we should have to have a scientifically accurate knowledge of everything in the speaker’s world. . . . We can define the names of minerals, for example, in terms of chemistry and mineralogy, as when we say that the ordinary meaning of the English word salt is “sodium chloride (NaCl),” . . . but we have no precise way of defining words like love or hate, which concern situations that have not been accurately classified. . . . (Bloomfield [1933] 1951, p. 139)
These statements seem to imply a model of linguistic function with just two parts—on the one hand, the linguistic form, and on the other hand, the associated nonlinguistic events (and, presumably, contextual linguistic events as well). Thus the definition of the word “salt” would be, at least in part, the actual substance NaCl. But Bloomfield seems to ignore the essentially arbitrary association between the word “salt” and the substance NaCl, in that his model has no place for the human individuals or the human cultures which have chosen this particular linguistic form.
A more satisfactory model was provided some two thousand years ago by the Hindu philosopher Pata?njali: “Concentrate separately on the word, the meaning, and the object, which are mixed up in common usage”—which a modern commentator explicates with this example—“When we utter the word ‘elephant,’ we find that the word, the meaning and the object are mixed up; the word lives in air, the meaning lives in mind, the elephant lives by itself” (Pata?njali, Aphorism . . .). It is indeed true that the word “lives in the air,” in the sense that it is transmitted as vibrations of air molecules. It is equally true that the actual elephant “lives by itself,” i.e., exists independently of all human conventions of nomenclature. The only way that these two isolates are related, then, is through the human mind; and we may define meaning not as a “thing,” but rather as the relationship which associates word and object.
This three-part model is more adequate than Bloomfield’s but still does not clarify the relation of language and culture. In order to do so, we may expand the model still further. First, a division may be made between the observational, or etic, universe, to which “word” and “object” belong, and the structural, or emic, universe, within the human mind. Second, we may distinguish linguistic behavior from its subject matter or content (though
the subject matter may itself, as a special case, be linguistic behavior, as when linguists talk about language). The two dichotomies then intersect as shown in Figure 1.
In this figure, the arrows marked a, b, and c indicate relationships of importance to the ethno-linguist. Arrow a is the relationship which concerns him when he functions purely as a linguist: it may be thought of inductively, in terms of the process by which the investigator sets up a structure to account for his raw behavioral data, or deductively, as the process by which psychological patterns of linguistic competence give rise to observable linguistic performance. Arrow b is the analogous relationship that is investigated by the ethnographer: the actual objects and events which concern a particular human group are here linked, by induction or by deduction, to subjective patterns of organization. Finally, the set of arrows marked c represents the relationships to which we assign the term “meaning” this is conceived of not as a direct connection between the utterance “elephant” and the flesh-and-blood Elephas maximus, but rather as a connection mediated by “elephant” as an item of the English lexicon and by “the elephant” as an item in the cultural inventory of English speakers.
There are two types of structural units which are linked by the relationships of meaning. The relevant linguistic units are not phonemes or morphemes, but units of a higher level, which are called lexemes: these are the minimum units which participate in arbitrary relationships of meaning, Thus, single morphemes like “green” and “house” are lexemes, but so also is the two-morpheme combination “greenhouse” (as opposed to “green house”), since it arbitrarily designates a particular kind of structure. There is still little agreement about structural units of cultural behavior; insofar as they can be identified, they are often called sememes. To be sure, there is not always a one-to-one correspondence between lexemes and sememes; people sometimes show culturally determined differences in behavior where their language provides no lexemic differentiation. However, the general regularity of lexeme–sememe correspondences reflects the close integration between language and the rest of culture, and it is in this way that language may be regarded as a key to culture as a whole.
Ethnosemantics . The study of vocabulary as a guide to the way in which members of a culture divide up their universe has received increasing attention, and the relativity of cultural classifications is emphasized with every new empirical study. Thus, where English vocabulary reflects its users’ approach to spatial orientation with the fourway classification “north, south, east, west,” the Indian languages of northwestern California reflect the functionally similar but incommensurable division “upriver, downriver, toward the river, away from the river.” Anthropologists have begun to pay close attention to such lexemic systems, understanding that they reflect an emic view of the culture being studied, a view uncontaminated by the varying etic frameworks of outside observers. Various terms have been used to identify this study, such as “ethnoscience,” “folk taxonomy,” “structural semantics,” and “ethnosemantics.”
A further development in ethnographic semantics is generally known as “componential analysis.” This method tries to answer the question, Given a particular set of taxonomic terms used by members of a culture, what are the criteria for applying the individual terms? Taking an example from kinship terminology, if some male collateral kin in generations above ego’s are called “uncle” and some are called “cousin,” what does one need to know in order to label a particular kinsman correctly? Attempts to answer such questions have resulted in the idea that terms may be conceived of as bundles of simultaneously occurring semantic components. Thus the term “uncle” is applied when the features of “maleness,” “ascending generation,” and “colineality” are simultaneously present. (A “colineal” in this case is a nonlineal kinsman all of whose ancestors are included in the ancestors of ego.) The term “cousin” is applied in a larger number of cases, but they include those where the features of “maleness,” “ascending generation,” and “ablineality” are simultaneously present. (An “ablineal” is a consanguineal kinsman who is neither a lineal nor a colineal.) This type of analysis, as applied to kin terms, results in definitions which are both more concise and more exact than the extensional definitions given in traditional ethnographies. Application of componential analysis to areas other than kinship holds great promise.
Language and world view . In addition to correspondences between vocabulary and cultural inventory, a much more controversial type of correlation between language and culture has been proposed. This involves, on one side, whole grammatical systems or subsystems, and, on the other side, whole philosophies or ways of life held to be characteristic of particular cultures (though often not brought to the level of conscious formulation). The interest of anthropologists was drawn to such correlations by Edward Sapir, who not only recognized a linguistic relativity, covarying with cultural relativity, but also postulated a linguistic determinism operating on culture:
Human beings do not live in the objective world alone, nor alone in the world of social activity as ordinarily understood, but are very much at the mercy of the particular language which has become the medium of expression for their society. . . . The fact of the matter is that the “real world” is to a large extent unconsciously built up on the language habits of the group. No two languages are ever sufficiently similar to be considered as representing the same social reality. The worlds in which different societies live are distinct worlds, not merely the same world with different labels attached. (Sapir [1910–1944] 1949, p. 162)
Benjamin Lee Whorf, a student of Sapir, continued the exploration of the matter, although with less emphasis on the tyranny of language over culture. His position has become known as the “Whorfian hypothesis,” which holds that “language patterns [and] cultural norms . . . have grown up together, constantly influencing each other. But in this partnership the nature of the language is the factor that limits free plasticity and rigidifies channels of development in the more autocratic way” (Whorf [1927–1941] 1956, p. 156). The deterministic role of language is easy to understand when we consider how much of culture is transmitted through the linguistic medium. However, the Whorfian hypothesis is easier to accept intuitively than to prove in a rigorous way; in particular, no correlations can be traced between language and world view until specific world views are themselves defined in terms of observable behavior. Whorf shows that Hopi linguistic structure is compatible with a world view involving a peculiar relation between subjective and objective experience; but he tends to assume, rather than to demonstrate, that the Hopi actually hold such a view of the world. Pending the outcome of extensive, strictly controlled, cross-cultural testing of the Whorfian hypothesis, we may limit our acceptance to the following modified formulation: “Insofar as languages differ in the ways they encode objective experience, language users tend to sort out and distinguish experiences differently according to the categories provided by their respective languages. These cognitions will tend to have certain effects on behavior” (Carroll 1963, p. 12).
Language and society . While the studies mentioned above have regarded each language as a unified whole, another type of research has focused attention on the variation that exists within languages or within multilingual speech communities. Such variation, apart from that associated with geographical dialects or with the idiosyncrasies of individuals, is commonly found to be correlated with one or more socially defined factors, such as the social identity of the speaker, the addressee, or the person referred to, and the social context in which communication takes place. Study of the covariance between linguistic diversity and social structure thus constitutes the new field of sociolinguistics. The findings of this field are applicable, from the synchronic viewpoint, to the diagnosis and analysis of social encounters, and, from the diachronic viewpoint, to examination of the ways in which linguistic patterns and social systems each change under the influence of the other.
William Bright
[Directly related are the entries Cognitive Theory; Componential analysis; Linguistics; Semantics and semiotics;and the biographies of Bloomfield; Sapir; Saussure; Whorf.]
BIBLIOGRAPHY
The most valuable reference that can be given for linguistic anthropology is Hymes 1964, which contains not only a rich selection of papers in the field but also very extensive bibliographies.
Bloomfield, Leonard (1933) 1951 Language. Rev. ed. New York: Holt.
Carroll, John B. 1963 Linguistic Relativity, Contrastive Linguistics, and Language Learning. 1RAL: International Review of Applied Linguistics 1: 1–20.
Diebold, A. Richard Jr. 1964 [Review of] Sol Saporta (editor), Psycholinguistics. Language 40:197–260. ? An extensive review of the whole field of psycholinguistics, including many matters of interest to linguistic anthropology.
Hammel, Eugene A. (editor) 1965 Formal Semantic Analysis. American Anthropologist New Series 67, no. 5, part 2 (Special publication).
Hymes, Dell H. (editor) 1964 Language in Culture and Society: Reader in Linguistics and Anthropology. New York: Harper.
Nida, Eugene A. 1964 Toward a Science of Translating. Leiden (Netherlands): Brill. H? Chapter 5, “Referential and Emotive Meanings,” summarizes recent work in ethnosemantics.
Pata?njaliAphorisms of Yoga. Translated into English with a commentary by Shree Purohit Swami. London: Faber, 1938.
Romney, A. Kimball; and D’Andrade, Roy Goodwin (editors) 1964 Transcultural Studies in Cognition. American Anthropologist New Series 66, no. 3, part 2 (Special publication). ? Contains contributions by linguists, anthropologists, and psychologists to problems of ethnosemantics.
Sapir, Edward A. (1910–1944) 1949 Selected Writings in Language, Culture, and Personality. Edited by David G. Mandelbaum. Berkeley: Univ. of California Press.
Whorf, Benjamin L. (1927–1941) 1956 Language, Thought and Reality. Edited by John B. CarroU. Cambridge, Mass.: M.I.T. Press.