The Symbolic Process and its Integration in Children

Chapter 5: The Behavior Content of Symbols

John Fordyce Markey

Table of Contents | Next | Previous

AT this point it will be well to summarize some further facts regarding the action content of symbols. The basis for such a consideration has already been discussed in Chapter III.

There is general agreement among sociological and psychological students upon the fact that symbols are acquired in social interaction. There is also a general consensus that the content of symbols is to a large degree action pertaining particularly to that for which the symbol is a substitute. But there is no agreement that the most complete explanation of symbols is action of one kind or another, or that symbols become what they are through action and are maintained by behaviour processes. A large number of persons wish to posit something over and above this action content, a "meaning " or " consciousness," or some other such factor in addition to behaviour. This problem of knowledge and meaning and the question as to how far the behaviour of organisms will account for meaning will be discussed later. For the present it seems desirable to consider the action content without going directly into this particular problem. However, it might be well to suggest, in passing, something of what is meant by the proposition that meaning exists through action on our part. Take the object " window-glass" for example. Because we can see through it, it is clear ; by touching, feeling and breaking, it is known to be hard ; by acts of rubbing, it is known to be smooth, and so on. Of course, these things are apparently not " known " in a reflective sense until symbols have arisen in social behaviour, thus making it possible for absent and past parts of behaviour

( 49) and experience to be brought into the present behaviour in an indirect manner by means of these social agencies, i.e., by symbolic action.

The present inquiry will naturally have implications concerning meaning and will also show at least some kind of "meaning "content. It maybe added that by the content of a symbol is meant the factors or processes embodied in it which go to characterize it as a symbol and make it more than a mere puff of wind through the vocal apparatus.

The behaviour content has been indicated quite definitely by various social psychologists and psychologists, particularly the behaviourists, who have studied the acquirement of speech in children ; for example, Watson, Allport, Mead, Block, and others. The role of action is clearly shown in the child's mastery of the use of symbols in his adjustment to his social surroundings.

In view of its evident importance one might expect that those studies dealing particularly with vocabulary acquirement in children would give a considerable amount of attention to the behaviour content, but he would be generally disappointed. These studies are often given over to a mere tabulation of words and to grammatical considerations rather than to psychological study. By thus ignoring the psychological factors much of this work remains of little worth, even from a grammatical standpoint, in showing the true character of child speech. Most of the studies on the learning of symbols by children classify words into the traditional adult parts of speech. Such a classification may mean very little regarding the significance of these words in the behaviour of the child. It is rather remarkable that during the considerable period in which this phase of child study has been in vogue, so much time has been put upon such comparatively uninformative facts, and little or no attention given to more significant factors. However, sufficient materials have been collected to show some of the importance of action and which yield, by further analysis, valuable information concerning the action content of the child's symbolic behaviour.

( 50)

The young child has few, if any, words of an abstract nature. Its early symbols are for simple concrete designations of the more obvious acts and action-objects [1] of its surroundings. This has been clearly demonstrated by such investigators as Nice (1917), Boyd (1914), Bohn (1914) and Drevers (1915, III), to mention a few. In addition to this, numerous investigators such as Chamberlain, Tracy, Dewey, Binet, and others have shown that the child's symbols are action words, i.e., their content is action. There is also practically universal agreement upon the fact that the first symbols of the child are in reality word-sentences designating action and object or subject, or all three at once. Thus, for instance, Koffka (1924, p. 301) says a " mother " is not only something which " is so," but more exactly something which " does this," " assists thus," or " punishes so." The gesture language of the child is also a potent testimony of the action content of symbols.

Binet (1890) made a study of his two daughters in an attempt to discover children's interests by asking them the meaning of 33 nouns that he had selected.[2] The selection is somewhat arbitrary, and the criticism might be brought that they lend themselves to the results. However, as the list really appears fairly representative, it is more than probable that any other list would give similar results. Still, if the above criticisms were true, it would not diminish to any appreciable extent the comparative value obtained by using the same set of words with other children and at different ages. This has been done by Barnes (1896-7) ; also by Shaw (1896).

It should also be pointed out that these studies are not studies in introspection ; the analysis is of the responses themselves as such. It might be objected that we are dealing with mere speech reactions and these children may have " meant " something different from what they said.

( 51) This objection is not valid. It is the speech reactionthe naive response of the child-in which we are interested and from which significant conclusions can be drawn. It is assumed that the child is not outside our universe of discourse ; that when he responds, " A hat is to put on your head," this speech reaction is related to the stimulus word and consequently constitutes one valid indication of its stimulus character for the child.

It is quite probable, too, that the child means his response. Furthermore, a certain" correct understanding " of the stimulus word is not essential ; all that is necessary is some distinctive response to it as a symbol.

Binet's conclusions were that children are impressed to a very small extent by visible aspects of things. Their ideas possess only slight abstract characteristics. Their greatest interest is to be found in the use of things.

This last conclusion is the one which bears most directly upon the present problem. If true, it means that the content of the symbolic process as indicated by the child's defining responses in its early years at least has to do primarily with action, the content of symbols being behaviour or action of one kind or another. Barnes took the same list of words and obtained responses from more than 2000 children in Monterey County, California. Out of these 200o he took the responses of 5o boys and 50girls for each age from 6 to 15 years, making zoo for each age and 1,00o in all. The selection was made by taking the first reports which were sent in until each 5o had been selected. It seems that such a selection of responses would be a fairly random one, as far as the action content of responses is concerned. Apparently, such factors as promptness on the part of the teacher had little constant relation if any to the action content of the responses of the children. Barnes classified the replies into 9 categories. Due to the fact that some answers fell under more than one category, the total statements collated were 37,136. An example of a statement falling under several heads is ; A knife is a tool (larger term) made of iron (substance)

( 52) having a blade and handle (structure), and is used to cut bread (use).[3] The question asked was, " What is a . . . ?" The results of this study are unmistakable regarding the behaviour content of these words. Two of his categories obviously express behaviour, i.e., use and action. The percentages are given in Table IV. The obvious action content is very marked in the earlier ages and decreases with age, at 6 years 82 % , at 15 years 33 %, almost 50 % difference.

As a partial check upon this evidence of action predominance, I classified the responses which Billet records for his children to the same list of words with the exception that with the younger child these words were omitted clock, doll, omnibus, armchair, bottle, finger. The following words were submitted: papa, spoon, bed, chair, lobster, eye. Responses to several other lists of nouris were obtainable from different sources. These were classified in a similar manner. The results are given in Table V and corroborate the study of Barnes.

In addition to the above, three other sets of data may be given. Chamberlain (1909) gives a list of 1186 words, the meaning of which was asked of a child during its 47th and 48th months. These were taken from the different parts of speech. Checking over these, I found as a fairly conservative figure approximately 70% of the responses expressing obvious action. A stricter consideration of the words should raise the percentage. Wolff (1897) states that in the Boy's Dictionary (the boy was 7 years old) 75% of the less abstract words (not nouns) express action. In the sample facsimile given by her, of 173 words at least 60 % indicate action. Shaw (1896) by simple association, no attempt being made otherwise to get the content of the word itself, found for 600 city children between 8 and 13 years inclusive, that 34% of the statements express

( 53) obvious action (use or action). He used Binet's list of words.

Table IV Responses of Children to Noun Stimulus Words (original figures from which percentages were obtained taken from Barnes 1896-97)
Age Use and action Larger Term Substance Structure Form Place Quality Colour Unclassified Total
6 82.0 3.1 1.8 1.1 .3 .3 1.8 .6 9.0 100.0
7 70.8 7.5 2.4 1.4 1.1 1.4 2.7 1.5 11.1 99.9
8 73.4 6.7 4.3 1.9 1.0 2.8 3.8 .5 5.6 100.0
9 67.8 11.8 6.2 2.6 1.0 2.7 3.0 .8 4.1 100.0
10 61.6 12.1 7.3 3.2 .9 4.3 2.0 .3 8.3 100.0
11 48.3 19.6 10.1 3.9 1.9 5.3 6.3 .5 4.0 99.9
12 46.7 16.6 10.9 5.6 1.5 5.7 4.2 .2 8.6 100.0
13 37.2 22.8 11.6 7.0 3.1 5.5 6.4 .3 6.1 100.0
14 41.2 25.5 9.7 7.6 2.6 4.8 4.6 .7 3.3 100.0
15 32.7 38.0 11.0 6.4 3.1 3.6 3.0 .5 1.8 100.1
Ave. 56.2 16.4 7.5 4.1 1.65 3.6 3.8 .6 6.2 99.9

Table V. Use and Action Responses
Author Child Age No. Of Stimulus Words % of Use and Action Responses
Binet 1890 Girl younger 2.5 —3.25 33 87
Nice 1917 Girl E 3 14 87
Nice  1917 Girl D 4 6 [1] 100
Pelsma 1910 Girl E 4 66 78
Binet 1890 Girl older 4.25—5 33 84
[1] One word, bathe, not a noun

In view of the apparent representative character of Barnes' cases, although they were from the country, they are corroborated by Shaw's cases from city children [4] as well as by the other incidental data cited above, it is probably that they depict a general situation. The per cent. of obvious action in the definition at 15 years is practically the same as that given by Webster for these words.


In so far as these data are concerned Binet's conclusions are substantiated, that young children are impressed very little with abstract characteristics, and are mainly concerned with use. The term action should be included because the use of a thing includes action. The small per cent. of responses in those categories which approach abstract characteristics is shown in the Table (IV). Binet's conclusion regarding the small impression by visual aspects is probably only superficially true and then only for certain aspects, for the visual function seems quite important in use and action discrimination.

There is another significant conclusion which may be tentatively drawn. It is that with increasing age there is a decrease in obvious action content and an increase in refined action content, as shown by such responses as larger term, structure, substance, and the like (Graph I). It is undoubtedly true that this trend is not so much a function of age as it is a function of social behaviour and experience.

What is the significance of these facts regarding the action content as indicated by responses to noun stimulus words, for the content of the whole vocabulary of the child? If they can be taken to represent only the action content of nouns in the child's vocabulary this is important, for the percentage of nouns sometimes goes well over 8o

(see Nice, Boyd, Tracy, and others). Kirkpatrick (1891) estimates from 55% to 85`% nouns. Nouns thus form a considerable portion of the total vocabulary.

However, the classification of a child's vocabulary upon the basis of the adult parts of speech is a highly arbitrary and fictitious process, particularly with reference to nouns. These first so-called nouns, as Dewey (1894) and others have pointed out, are in reality verbal-adjectivalnominal or nominal-adjectival-verbal symbols and the like. In reality they are action words and word-sentences, as we have seen, and are often accompanied by the appropriate action on the part of the child. Due to all these facts it would seem as appropriate to classify the child's nouns on the basis of usage, as action words or verbs.


Graph 1


Assuming that these facts are sufficient to indicate the behaviouristic character of nominal symbols, it remains to consider the other parts of speech. Here the task is relatively simpler. Verbs are action words by definition. Adjectives and adverbs, which are of later development, are modifiers, consequently serve to indicate characteristics of other action words. The connectives, forming only a small per cent., serve as aids to other parts of speech and to sentence differentiation. Interjections, also a small per cent., serve as emotional expression. The pronouns stand for nouns, or persons, which are acting organisms. The behaviouristic content of these. remaining parts of speech should also be evident.

In this connection the evidence from Anthropological study is very instructive. Among pre-literate [5] people there is the same absence of abstract general terms (Jespersen, 1923, p. 429 ff). For instance, the Tasmanians have no word for " tree," but have words for " gum-tree," " wattle-tree," etc. ; the Zulus have no word for " cow," but have words for "red-cow," "white-cow," etc. The Cherokees, instead of one word for " washing," have different words according to what is washed. Brinton tells of the poor missionary to an Oregon tribe, who, to convey the idea of " soul," found no word nearer than one which meant '` the lower gut."

Of course it is evident that there is a certain amount of abstraction in any symbol, still there is considerable difference between talking about trees in general and this, a particular tree.

On the other hand, words mean action and use, something to be done, as shown so well by Malinowski (1923) ; the name of a thing means its proper use, and similarly verbs receive their significance in active participation in action. The speech of early humans consisted of " irregular conglomerations; " they expressed whole sentences in a word which might contain a half-dozen different ideas (Jespersen, 1923, p. 421). Brinton (1890, p.403 f) from his analysis principally of American languages also concludes

( 57) regarding these sentence-words that they partake of the nature of verbs rather than nouns. There is practical agreement upon this word-sentence character of preliterate speech. Philologists have also pointed out that roots indicate action, they are verbal in nature (Whitney, 1910, p. 260 ff ; Miller, 1887 ; Romanes, 1889 ; et al.). It must be kept in mind, however, that roots are not limited to primitive speech, but belong to all periods. But it is of some significance that all of the words of the voluminous English language can be etymologically reduced to a few roots which are verbal, i.e., express action, rather than nominal in character. The tracing of roots is a very devious process and would not mean much if only one authority came to such conclusions, but there is a wide agreement even among such antagonists as Whitney and Miller, for example.

The facts regarding gesture language among pre-literate peoples also give added data concerning the concrete nature and obvious action content of symbols (see Wundt, 1916, pp. 53-75).

This action content is not without considerable significance for the imitation theory. If words are acquired merely by imitation, why this preponderance of action content, particularly among pre-literate peoples and the lower age groups among children ?

The facts and tendencies shown by the material from Barnes are sufficiently clear on their face for the use made of them. However, from a methodological standpoint, it might be worth while to analyse them somewhat more thoroughly.

The marked decrease with age of responses indicating the obvious action content of the symbols has already been commented upon. It was also pointed out that this cannot necessarily be taken as evidence of the decrease of the action content itself. When in the beginning the content of the child's symbols is action, a decrease of obvious action may be accounted for either by the introduction of a greater and greater amount of unexplained " meaning " or by a greater refinement of action content

( 58) resulting from a more abstract analysis, but still carried on by behaviour processes. As the problem of meaning will be discussed later, and as the present discussion has already furnished evidence of the action content of symbols, the second alternative, viz., that of an increase in more refined behaviour analysis of symbols, which is in line with the behaviouristic hypothesis, may be tentatively accepted.

This decrease of obvious action is most closely associated with age. A Pearsonian correlation gives an r of - .975, showing the tendency for the responses to take another form with advancing age. There are only ten items correlated. Yule says there should be at least 15, in order

to balance extraneous or chance deviations. Thus in the above correlation there is an opportunity for a large deviation to have undue weight in producing the coefficient. However, allowing for this factor would probably not reduce the r by a significant amount. Furthermore, these percentages are taken from over 37,000 statements and are probably fairly representative. The deviations at 13 and 14 years are probably due to school selection. The deviation at 7 years is apparently due to the fact that the teacher wrote out the responses for the 6-year-old children, thus giving them freer expression, while the task of writing was still difficult for the 7-year-olds who wrote out their own responses. This undoubtedly accounts to some extent for the larger per cent in unclassified and the smaller proportion of obvious action responses in the 7-year-old group in comparison with the 6 and 8-year groups. If correction could be made for these deviations, it would probably result in a more exact relationship. Taking all these factors into consideration along with the nature of the data and the manner of its selection, it is probable that the above coefficient is somewhere near the true relationship. However, on account of the small number of items and as a check upon the Pearsonian productmoment r, a Spearman rank order coefficient [6] (p) was computed. It gave p = (-) .986. Translating this into an r

( 59) by Pearson's correction gives r = ( -) .99, a slightly higher coefficient than that obtained by the first method. It is interesting to add the two cases of Binet to the lower end of the series and note the further increase of obvious action with the lower ages (Table VI). While the validity of comparing two cases with i,ooo can be questioned, the results are as we should expect from a random sample, judging upon the basis of the r already obtained.

Source Table IV and V
Age Use and Action   Age Use and Action
3 87   10 62
5 84   11 48
6 82   12 47
7 71   13 37
8 73   14 41
9 68   15 33

The results obtained by Shaw on the study similar to that of Barnes from G00 city children ranging from 8 to 13 years show an increase up to ii years in both action and use responses. Shaw thinks this contradicts the results of Barnes. However, the results are not comparable in this respect. Shaw's instructions were given in an attempt to get an association response rather than a response showing the content of the stimulus symbol. He thus gets a larger number of categories for classification. At least two, if not more, of these categories would tend to decrease the action and use percentage in the earlier ages. Sentence making takes up over 14% at 8 years, but only 2.5% at 13 years, both an absolute and a relative decrease. This category is bound to be in part a subtraction from use and action. The case is similar, although not so marked, with the category of possession responses. It remains fairly constant for 8, 9, 10 years, then has a marked decrease. Further, Shaw's classification is different from that of Barnes. His definition of use is much more abstract. He classifies under action statements which Barnes classifies under use. In short,

(60) Shaw's data are not contradictory to those of Barnes. If the same use and action category could be segregated, it seems quite probable that it would be similar to that of Barnes up to ii years as it is after 11 years.

Coefficients, computed from Barnes' data, showing the association of the decrease in obvious action responses with the increase of the tentatively assumed more abstract or refined action responses are given in Table VII. Pearsonian is were computed. Then, as a check, on account of the small number of items, the Spearman rank order p's were also computed, then translated by Pearson's correction into r coefficients. The results of these two methods corroborate each other.

Correlation of Obvious Actions with More Refined Action Content
  r. r from p.
Action and Age -.975 -.99
Action and Substance -.949 -.95
Action and Structure -.946 -.95
Action and Form -.943 -.90
Action and Larger Term -.935 -.98
Action and Place -.804 -.76
Action and Quality -.61 -.61
Action and Colour +.33 +.44

The inverse association is quite high with all except colour, which is positive, but too low to be of much significance. If the latter were really significant it would be evidence counter to Binet's conclusion regarding visual responses.

The main conclusion remains that in young children the content of symbols is action. With increasing age the obvious action content tends to decrease, while at the same time, apparently, the refined action content increases.


  1. By an action-object it is meant to indicate that our Objects exist for us due to behaviour regarding them.
  2. The list is as follows : knife, bread, doll, water, armchair, hat, garden, piece of sugar, thread, horse, table, mamma, potatoes, bottle, flower, snail, mouth, lamp, bird, dog, carriage, pencil, earthworm, shoes, finger, clock, house, wolf, omnibus, balloon, village, box, handkerchief.
  3. Examples of other categories:
    Action-a clock goes tick-tock. Colour-a clock is yellow.
    Quality-a clock is pretty. Form--a clock is round.
    Place-a clock is on the wall.
  4. See page 59 ff for further analysis of Shaw's data.
  5. Faris (1929) proposes this term for so-called primitive peoples.
  6. See Rugg (1917, pp. 284-92) and Kelley (1923, p. 193).

Valid HTML 4.01 Strict Valid CSS2