:.NRLF DEC 29 tZDtie Wini\itvsiitp of Chicago STUDIES IN THE PSYCHOLOGY OF LANGUAGE I. THE CONSCIOUSNESS OF SAMENESS OF SPEECH SOUNDS II. THE FUNCTION OF IMITATION IN SPEAKING A DISSERTATION SUBMITTED TO THE FACULTY ;4lV£^^ 1/ OF THE GRADUATE SCHOOL OF ARTS AND LITERATURE IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF COMPARATIVE PHILOLOGY BY Karl Friedrich Muenzinger Private Edition, Distributed by THE UNIVERSITY OF CHICAGO LIBRARIES CHICAGO, ILLINOIS 1918 C{)e Wini\ittsiitp of Cfitcaso STUDIES IN THE PSYCHOLOGY OF LANGUAGE I. THE CONSCIOUSNESS OF SAMENESS OF SPEECH SOUNDS II. THE FUNCTION OF IMITATION IN SPEAKING A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF ARTS AND LITERATURE IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF COMPARATIVE PHILOLOGY BY Karl Friedrich Muenzinger Private Edition, Distributed by THE UNIVERSITY OF CHICAGO LIBRARIES CHICAGO, ILLINOIS 1918 Note : I wish to express my gratitude to my teachers, Dean J. R. Angell, Professor G. H. Mead, Professor E. Prokosch, Professor F. A. Wood, and especially to Professor C. D. Buck for their assistance and guidance in the preparation of these studies. To the latter I owe above all a scientific introduction to the complexity of the problems of sound change, two of which are approached here from a psychological point of view. K. F. M. June, 1918. A. THE CONSCIOUSNESS OF SAMENESS OF SPEECH SOUNDS The articulation of a speaking person produces certain acoustic phenomena, a 'speech continuum,' of which we may get an objective record on a phonograph or grammophone. If we examine an enlarged transcription of such a record^ we are surprised to find how the smallest elements vary continually, how their characteristics change by a gradual shading into each other, and how impossible it seems at first to mark off larger units representing what we call sounds and words. And yet a person hearing a speech continuum will have no difficulty in distinguishing sounds and words, provided the speaker belongs to the same speech community. We must suppose that the stimulation produced at the auditory receptor, the ear, corresponds closely to the speech continuum. How is it then that what the hearer perceives is not a sound continuum, but sounds and words? A 'perception' we are ususally informed is a compound of an im- mediate stimulation plus some past experience. Such a definition, however, does not do justice to the 'meaning' of a perception, or in other words to its function as a part of a larger psychic process. What is necessary to admit in the definition of perception beside the reference to the past is a reference to the future. This is con- tained in Professor Mead's conclusion that "percepts .... are compounds of the experience of immediate stimulation and the im- agery of response to which the stimulation will lead. The object can be properly stated in terms of conduct."^ That is, to illustrate, the pencil here by my side is never a mere pencil; it is usually a pencil to write with, but it may also be an object to keep a certain place in a book, or various other instruments which the situation calls for. This functioning of the "imagery of the response to which the stimu- lation will lead" is determined, so it seems to me, by the present conscious attitude. Not any random past experience may merge * Scripture, E. W., Elements of Experimental Phonetics, Plate I; cp. also Muen- zinger, K. F., A New Apparatus for the Enlargement of Phonographic Records, Bull. Univ. Texas, 1915, No. 24, Plates II and III. * Mead, G. H., Jour. Phil. Psych, and Sc. Meth., X, 401. 1 2.\ : : .; •*; .S.xudje&in t-he Psychology of Language with the sensation, but whatever is selected must fit somehow into the present situation and its bearing upon the future. In order to bring out this relation to the future more clearly, one ought to speak of the direction of an attitude. The problem becomes immensely compUcated if the percept is an element of a larger process, where it is partly merged with the sur- rounding elements, and where the process is perceived as a unit and in return leads to a response of similar complexity, as is the case in the conversation of two persons. And yet it seems to me that Professor Mead's interpretation must and can still be applied to the elements of the process. We must conceive of the total response to a social situation as prepared by and composed of a number of minor re- sponses that are all merged in that total response which alone stands out with a distinct unity. Let us suppose that in a conversation one sentence uttered by one person calls out a response by another also in the form of a sentence. The process going on in this second person must represent something like this. The attitude created by perceiving the first sentence is what we usually call the meaning or the contents of the sentence. This attitude is modified by other processes going on in the listener's consciousness, imagery of past experience, present stimulation other than those speech sounds, and finally the social situation thus created leads to the response of uttering another sentence. What is commonly termed an "expression of thinking" is in terms of conduct a response to a social situation. Just before the utterance takes place the attitude represents what Wundt calls the Gesamtvorstellung, whose elements are to appear in the following analytic-synthetic process of speaking. The perception of the words and sounds while hearing the sentence is a process secondary in importance to the process of perceiving the meaning. This is sometimes expressed by stating that our attention is primarily directed towards the meaning of a sentence. The hearer listening as the speaker produces the speech continuum by his articulation could at any moment reproduce most of the words he had heard so far, although the sentence stimulates him as a unit. And yet the perception of the words is not the primary object of the process of listening; it is the meaning of the sentence which is to be received. This meaning is synthesized through whatever the words impart. Studies in the Psychology of Language 3 The fact that the speech continuum is split up, so to speak, into words is on the one hand due to that past experience which had a similar sound sequence as a unit of stimulation followed by, or rather determine by, a unit of response. This is exemplified by listening to utterances in a foreign tongue in which one has only a limited vocabulary. It is often impossible then to say how many words one has heard, on account of the unfamiliarity of certain sound combinations. On the other hand the analysis is also due to the direction of the attitude, or the 'context' here, which selects with reference to the future. Our punsters make use of the excep- tional situations where the attitude may have two possible directions. The uncertainty that sometimes exists as to whether we shall write an expression as German zustandekommen in one, two {zustande kommen), or three words {zu stande kommen) does not disprove the existence of separate words, but rather adds some value to our principle.^ The entire expression zustandekommen answers to one dis- tinct meaning, i.e., calls out a single response. Since the parts stande, zu, and kommen have also appeared time and again as separate and distinct bearers of meaning, they have some independence which expresses itself in the uncertainty of the convention of writing. The units of responses next to the sentence are the phrases, and these again contain the words as smaller units. The degree of independence of the smaller units is not a fixed quantity, but varies according to the context and sentence accent. The adjective blue is more indepen- dent in the phrases a blue dress, a blue sky than in a bliiebird, or a blue print, whether convention makes us write the latter in three words or in two, a blueprint. Considered as a sound experience the word appears to be the same under different circumstances, i.e., at different times, in different sentences, or if spoken by different persons. This means, expressed in term of conduct, that our reaction to it as far as its physical nature is concerned, its phonetic material, is the same, since it is the reaction which determines the perception. Compared objectively, the parts of the speech continuum which we mark oflf as representing the 'same' words are sometimes quite varied in appearance, but in our consciousness they are the same inasmuch as they have a func- tional identity. The problems which arise in regard to these phenom- ' Bloomfield, L., Trans. Am. Phil. Ass., XLV, 66. 4 Studies in the Psychology or Language ena are similar in the case of speech sounds. They will be more properly treated in that connection. The splitting up of a part of the speech continuum proceeds, of course, without any distinct presence of past imagery of a similar sound combination. This is fused with the sensation and does not appear separately, at least not in the act of hearing. Besides the word as a sound experience we have also its function as the bearer of a meaning. This meaning is determined by past responses to the word, which have been of a varied nature. The particular response needed here is selected by the present situation, i.e., the part of the sentence already elapsed with its synthesized atti- tude pointing in a certain direction. The partially evolved meaning of a part of a sentence determines the meaning of the word, and this in turn modifies the direction of the attitude. Duerr, in speaking of the word, calls it a formula to be filled in by the sentence. This attitude in the process of synthesis during the listening to a sentence is of an active type which is always ahead of the modifying elements of the moment. In a general way the contents of the sen- tence is vaguely constructed before the listener has heard the end. The attitude runs along the well known schema of the sentence and produces a predisposition (Einstellung) for each word or phrase to follow. It seems to happen very frequently also that words or parts of words are not presented as actual sensations, but are supplied from past experience entirely. At any rate the auditory sensation does not pass through the center of consioucness as a continual sound exper- ience of constant broadness, but it supplies only a number of 'cues' that are woven into a whole. The space between cues may then be entirely lacking or be present only in a dim way. Again, all this is implied by saying that our attention is directed towards the mean- ing. Under ordinary circumstances the word functions in the way just described. But it may itself also become the object of attention. This brings it outside the process of speech and makes it a part of the physical world. The synthesis of the attitude which finally appears as the meaning of the sentence is dependent entirely on habits. The quick succession of the words, their fleeting character as perceptions, would hardly allow them to function as they do if they were not parts of a chain of highly mechanized acts. The ability to think stands in direct Studies in the Psychology of Language 5 relation to the amount of mechanized language processes one pos- sesses. Up to a certain point the problem of the perception of speech sounds is similar to that of the word. Within the speech continuum which gives rise to a word smaller elements are again split off. This depends here as there on the past experience in which a sound as such has been a unit of stimulation, i.e., where it has been responded to as such. Inasmuch as the perception of a speech sound depends upon the response to the stimulation it is not surprising to find that the acoustic phenomena which at various occasions lead to the perception of the 'same' sound are in fact never quite alike. Let us go back one more step and look at the articulations producing the sound waves. They belong to the most deeply ingrained habits of the human being. Not only have we the habitual movements of one particular set of muscles in order to produce a certain sound, but we also have the correlation of the action of several muscular systems, such as those of the vocal organs proper (lips, jaws, tongue, velum, larynx), of the chest, and of a large region of the trunk. Furthermore the transition from one articulation to another requires some special movements, as any one can see for himself if he tries to pronounce a sound sequence that does not occur elsewhere in the language he speaks.'' This com- plicated mechanism does not work machine-like; it does not produce the same results under all circumstances. Habitual movements under similar circumstances attain only a certain degree of exactness. They are hardly ever identical from act to act. In a certain sense the articulation is the resonance of the tonic condition of the body and of the emotional character of consciousness. More than that, the surrounding articulations leave their trace or show their effect before they are actually performed, e.g., a yfeis articulated at different places of the roof of the mouth accordingly as it is followed by an a, e, or o. Still another reason for variation in articulation is the difference of the speech apparatus between persons; compare only the extreme differences between a child and an adult, a man and a woman. In speaking of these fluctuations as far as they pertain to one speak- ing individual the term 'zone of variations' has sometimes been used.^ * Sievers, E., Grimdzuge der Phonetik.^ * Sievers, E., op. cit., 272; cp. also Passy, P. E., fyude surles changements phonitiques, 40; Paul, H., Prinzipien der Sprachgeschichte,* 54; Wundt, W., Volker psychologic,^ I, 387; Oertel, H., Lectures on the Study of Language, 104. 6 Studies in the Psychology of Language It would be better, it seems to me, to extend the term so as to include also the articulations of individuals of the same community, because it is not any objective condition of the physical world that creates the zone of variations, but a psychic activity that holds together sounds of various quahties and recognizes them as 'belonging to the same type' even if spoken by different persons. The social situation where one sentence as a stimulus calls out another sentence as response, the 'expression of thinking,' must be supposed to exist in a secondary way already in the act of listening. That is, the understanding of a spoken sentence implies already an organized set of responses. It must be borne in mind that the atten- tion is turned towards the attitude which will finally produce the answer, so that the motor responses play a less prominent part in consciousness. These do not stand out as separate processes, but are merged in the synthetic process leading up to the Gesamtvorstellung. But the fact that they are not analyzed distinctly in the stream of consciousness does not mean that they are not present. They must be present, because their effect is felt in shaping the attitude of under- standing. Thus the analysis of the field of stimulation represented by the speech continuum into words and sounds does not exist as so many separate acts, and yet it is there, because the change of one sound (Cp. English hat: had; hat: hot) or the addition of a sound like the inflectional 5 changes the 'meaning.' This reservation must not be forgotten in a discussion of this kind, lest one commit the well known 'stimulus error.' A certain sound experience will be accepted by consiousness as the 'same' whether that sound appears in different words, or is pronounced with more or less force, or in a higher or lower pitch, or even by differ- ent persons, although, as I have pointed out, it will objectively be different in all these cases. The sound a as in father is the 'same' whether pronounced by my sister or brother, whether spoken with a loud or soft voice, in joy or in sadness. The 'consciousness of same- ness' as regards such a sound is then distinctly a psychic creation in which certain kinds of objective differences are ignored. In terms of behavior we have an identity of reaction in each case. If we were to say that the sound is recognized as 'belonging to a certain type,' we should not state the facts quite accurately. For it is primarily not the past experience present in some typified imagery that creates the 'consciousness of sameness,' but the response itself, or the Studies in the Psychology of Language 7 attitude of response, depending on the function of the sound. Wher- ever slightly different phonetic elements have a functional identity there is the tendency to produce identical responses. If we turn back again in introspection upon such experience, we find a 'con- sciousness of sameness,' which does not necessarily have to be pre- sent at the time of the original hearing as a clearly distinct process. It is a well known fact that out of the great variety of speech sounds which are possible and are observed to exist in one language or an- other, only a limited number are employed in any one language. What is a common sound in one language may be wholly lacking in another. Furthermore, the number of distinct sounds recognized in a given language by accurate phonetic analysis is greater than the number distinguished by the discrimination of the phonetically untrained speaker, whether literate (in which case he may be strongly biased by the visual image of the written form) or illiterate. This selective simplicity of the phonetic system of a language is also in accord with two general considerations. On the one hand there is the general reason for the creation of types in our conscious processes. In terms of conduct the type is a stimulation which leads to uniform responses. A number of physical objects are all tables to us on account of the similar response or attitude to sit down before them or to put things upon them. This 'type,' or unified response, is alone active in consciousness if I hear of 'a table.' The type plus some specific response will be called out by 'the table.' In the first case individual traits are of no account, i.e., they have no meaning, they are not needed. Thus it is with the consciousness of sameness of speech sounds. All that is needed is a number of unified responses or attitudes which go to make up the perception of the word and finally of the sentence. A word will therefore be the 'same' where and however it occurs, and the sounds that compose it will be the 'same' in other words. On the other hand there is an historical reason for the assumption of the simplicity of phonetic systems, namely, the fact of sound change. We know definitely that a sound change often comprises the entire zone of variations, that the whole zone changes in a certain direction, and that the individual sounds do not all fall a prey to the surrounding influences of other sounds, but follow one direction. And wherever a special condition is strong enough to cause a split in the zone of variations, the new sound usually falls together with one represented by another zone. 8 Studies in the Psychology of Language From the point of view of an associational psychology it is the association of slightly different sounds that holds together the zone of variations, so that the individual sounds do not fall out of line through assimulating influences. Wundt says: ^^Wirken die Laut- kontakte differenzierend, . ... so wirken umgekehrt die Lautassozia- tionen uniformierend, indem sie solchen Lautgruppen und Einzel- lauten, die durch hdufigen Gebrauch in hoherem Mass eingeiibt sind, mehr und mehr das Ubergewicht verschaffen iiber andere, seltenere.''^ The criticism as regards such a view is a general one: The principle of association does not explain the process as yet; the principle itself must be accounted for.'' There is another fact which shows that the sound possesses some independence: The possibility of such 'mistakes' which in philology are called metathesis, dissimilation, and distant assimilation of sounds. Compare Old English waxan from wascan, Latin celebrum for cerebrum, cocodrillus for crocodillus, Italian grolioso for glorioso. In contrast to the consciousness of sameness just described, there also exists a consciousness of difference of speech sounds. We hear the 'same' sound to be sure under different conditions, but we are also aware of its being different, or at least of its different import. Referring to the instance given above, we are aware by whom the sound a in a given word is pronounced and also in what mood it is uttered. Actual variations as regards pitch, force, and quantity may be partly analyzed as such in the field of stimulation, although usually they are not, but variations in quality very rarely appear as separable or distinguishable experiences. Looked at from the view- point of import, those variations that play a part in forming the understanding of the sentence are the ones that can be separated in the field of stimulation. They are those characteristics which represent the dynamic and chromatic accent. But the differences which impart the mood of the speaker, his tonic bodily condition, his personality itself, can as a rule not be analysed, There is danger then of taking the term 'consciousness of difference' just as the term 'consciousness of sameness,' too literally, and against this we must guard. Just as we found that in terms of conduct the perception of similar sounds really means a functional identity and consequently a similar response, so the difference in the perception of similar * Wundt, W., op. cit., 532; cp. Karsten, G., Phonetische Studien, III, 8. "> Mead, G. H., Psych. Bull, I, 380. Studies in the Psychology of Language 9 sounds means in fact a functional difference and consequently a difference in response. Two persons may pronounce a sound or a sound combination, a word, in precisely the same manner, so that they are seemingly 'the same' in both experiences of the hearer, and yet each single experience also contains the reference to the speaker. I can tell without analysis or reflection who it is that speaks in the next room, my brother or my sister. Besides the various attitudes called out by the stimulation of the utterance as already indicated there is also present an attitude directed towards the speaker implying his personality. Such an attitude can attain any degree of intensity, from the actual perception of the speaker as a self to the very dim awareness of his presence. Similarly it is with the other imports of the consciousness of difference. Objectively considered, the variations caused by the varying emotional states of the speaker's consciousness may exist, but it depends on our response whether and to what degree we shall be aware of them. II The existence of a unified response to a particular group of similar speech sounds for which I have used the term 'consciousness of same- ness, is of considerable importance in regard to the problem of uniform sound change.* This dogma of uniform change is now about 40 years old. Long before the seventies of the last centruy the fact of great regularity in many sound changes had been recognized. Thus Grimm had found that Indo-European surd stops are replaced in Germanic languages by spirants (compare Latin pater: English /a^/fer, Latin tres: English three, Latin cornii: English horn), and many other sweeping changes had been recognized and tabulated. More and more "phonetic laws ' were discovered. But even those scholors who emphasized most forcibly the importance of such laws did not hesitate to assume casual exceptions; and it was not until 1876 that Leskien stated that "language is not a fit subject for scientific investigation" if we allowed the "exceptions" to stand beside the "regularity of sound laws." If we find that a sound change is not what it ought to be according to the "sound law" then we must account for its occurrence ^ For a history of the dogma see Wechssler, E., debt es Lautgesetze? , 77 ff. with full bibliography; also Oertel, H., op. cit., 258. 10 Studies in the Psychology of Language instead of regarding it as a matter of chance.^ The following ten years saw one of the most violent battles ever fought in any science. The philologists were separated into two clearly distinct camps. Along with Leskien, the "Junggrammatiker," among whom were Brugmann, Osthoff, Paul, Delbrueck, and Sievers, defended the dogma that in a given dialect and period all phonetic laws work without exceptions, unless they are counteracted by other forces such as analogy. Against these stood another group of eminent linguists, among whom were Curtius, Whitney, Easton, Schuchardt, Bezzen- berger, Collitz, Tobler, and Jespersen, who maintained that an abso- lute uniformity was inconceivable, because speech as a psychical process could not be expected to be subject to laws similar to those of the physical world, since the conditions were all too varied and unanalyzable. Present linguistic science recognizes a measure of truth and value in both positions and tends towards less dogmatism in theory but greater strictness in practice. On the one hand the uncompromising assertion of uniformity served to discredit the loose assumption of casual exceptions, and led in countless specific cases to a satisfactory explanation of apparent exceptions. Futhermore, it is universally admitted that a great proportion of phonetic changes exhibit uni- formity under like conditions which are capable of formulation; that is, not merely under absolutely identical conditions the insistence in which would make any formulation impractical, but under ap- proximately like conditions in which minor differences may be ignored like the remoter decimals in a mathematical calculation. Again the principle of causality finally led to attempts not only to account for the 'exceptions,' but also for the 'laws' themselves. On the other hand the insistence on the fact that variations do exist in great num- bers as shown in our everyday speech, has opened the eyes of the linguist to the complexity of conditions that accompany sound production and sound change. Besides the relatively simple factors which the supporters of the dogma had in mind at first as determining like and unlike conditions (surrounding sounds, accent, etc.) there are many others which, while often practically negligible, must also be reckoned with. It has also paved the way for a clear distinction between variations of sounds in any direction due to momentary * Leskien, A., Die Deklination im Slavisch-Lilauischen und Germanischen, p. xxviii. Studies in the Psychology of Language 11 conditions, as emotional or bodily states, and changes of sounds arising out of persistent variations in one direction due to phonetic tendencies. As a result of adverse criticism we now find such carefully worded statements as that of Paul : ''Wenn wir daher von konsequenter Wirkung der Lautgesetze reden, so kann das nur heissen, dass bei dem Laut- wandel inner halt desselbe^i Dialektes alle einzelnen Fdlle, in denen die gleichen lautlichen Bedingungen vorliegen, gleichmdssig behandelt werden Man muss dabei natiirlich sdmtliche Momente der Laut- erzeugung inhetracht Ziehen. Namentlich muss man auch das Wort nickt isoliert, sondern nach seiner Stellung innerhalb des Satzgefuges betrachten. Erst dann ist es moglich die Konsequenz der Lautgesetze zu erkennen.^'^^ In theory there seems to be no objection to such a principle, but the study of any living dialect shows that there are very few cases of 'like conditions.' A careful observer like Weigand for instance denies them entirely.^^ The lack of uniformity in a given dialect has also been shown by Gauchat^^ who has given us the most interesting dialect study we have and whose results are in striking contrast to those of Rousselot.^^ And even if we should agree with Paul's assertion (arrived at by a logical process carried too far) that "we must distinguish as many languages as there are individuals"^* there would still be the different social situations which cause the individual to speak differently at times as well as his own changing bodily and mental states. Our 'consciousness of difference' is a sufficient indication of ever changing conditions in speaking. Imitation has been recognized as the most potent factor that works towards uniformity in the group though one particular type of imitation may disturb uniformity. Its mechansim will be analyzed in the second of these studies. That consciousness of sameness functions in a similar way in regard to the speech of one individual will be shown in the following pages. From the point of view of con- " Paul, H., op. cit., 69. " Weigand, G., Linguistischer Atlas des dako-rumanischen Sprachgebietes, 19. ^^ Gauchat, L., L'unite phonetique dans la patois d'une commune in Festschrift fiir Morf, 175. *^ Rousselot, J., Les modifications phonetiqites du langage — etudiees dans la patois d'une famille de Cellefrouin. Ret. des patois Gallo-Rom., IV, 65; V, 412 ff. " Paul, H., op. cit., 37. 12 Studies in the Psychology of Language duct both processes are somewhat alike in so far as they rest upon identity of responses. Ill In order to prove the possibihty of uniform change of a sound in all words of the language of an individual, Brugmann and Paul argue that it must be the motor image of a single sound that changes and not that of a word: ^^Das Bewegungsgefiihl bildet sich ja nicht fiir jedes einzelne Wort besonders, sondern iiberall, wo in der Rede die gleichen Elemente wiederkehren, wird ihre Erzeugung durch das gleiche Bewegungsgefiihl geregeli."^^ This opinion is supported by the fact that whenever a child is able to pronounce a new sound it will use it in all the words where its own small community uses it. Meringer ventures a flat contradiction: "Ich halte das einfach fur falsch. Die Aussprache wird in der Tat fiir jedes einzelne Wort besonders gelernt, was man gleich sehen wird, wenn man beobachtet, wie vorsichtig tastend wir oft ungewohnliche Worter sprechen. Diese werden wirklich buch- stabierend hervorgebracht, aber geldufiges Reden ware wohl ohne die Worts prechbilder unmoglich."'^^ This argument refers only to such 'unusual words' which contain unfamiliar sound combinations. It overlooks the fact that not only the articulation of a sound is based on habits, but also its joining to the preceding and following sounds. We can pronounce without difficulty or hesitation new words, e.g., proper names, as long as they contain besides famihar sounds also familiar sound sequences. It is for the absence of these latter that foreign names are often difficult and that we proceed hesitatingly in trying to pronounce them. A necessary sequence to the theory that the word changes as a unit would be that a sound change must be transferred from one word to another in order to become uniform. This is Weigand's position who bases his statement "Die Lautverdnderungen gehen von einzelnen Wdrtern aus^'' upon a multitude of observations in modern Roumanian dialects. Gauchat^* observes the change of a certain sound in some words and not in others in the case of the same individual, and in more words in the speech of some individuals (usually the younger) « Ibid., 69. " Meringer, R., und Mayer, K., Versprechen und Verlesen, 7; cp. Wechssler, E., op. cit., 22. 17 Weigand, G., op. cit., 37. " Gauchat, L., op. cit., 205. Studies in the Psychology of Language 13 than in others. TarbelP^ and Whitney^" cite the New England shortening of the long o in whole, stone, etc., which occurs only in certain words, varying in different communities. Wheeler referring to this change-^ expresses the opinion that all "ultimate uniformity" is brought about in this way. According to him the difference be- tween a new and an old pronunciation of one sound in one word is carried over to the same sound in other words; this is "the compelling force".! confess I cannot see a "compelling force" in the mere difference between two pronunciations even if one were felt as new and the other as old. There may be an uncertainty in regard to the choice of either pronunciation, but then an additional cause would be needed to fix one or the other of the alternatives. It seems to me that this issue of sound or word change is not a real one. Judging from different sets of facts each side is correct in part, but neither could prove its point entirely. For it is not at all a question of either-or, whether we speak on the basis of motor images of sounds or of words. The fact is, we have and use both. Furthermore, the articulation of a single sound consists in the co- ordination of a number of muscular systems. If Paul's reasoning is correct, it ought to apply to the elements of such a coordination as well, which indeed it seems to do when we find that in a language all voiced stops irrespective of their place of articulation are shifted to voiceless stops. The function of a motor image of a sound is not so very different from that of a word. Both contain highly complex coordinations, the word more so than the sound, so that we may regard both from the point of view of the execution of habitual movements as similar actions, but of a different order. Opinions of this kind are sometimes coupled with a rather too- mechanical interpretation of the relation of habitual movements, to their corresponding images. It is no doubt correctly supposed that a certain neural disposition, a diathesis as Oertal calls it," lies at the bottom of such a movement and it is further implied that a stim- ulation of the diathesis always creates the same movement as a mould can impress but one form. A movement as well as a sensation is modified by past imagery and the present 'situation.' For this " Tarbell, F. B., Trans. Am. Phil. Assoc, 1886, p. 5. " Whitney, W. D., Indogerm. Forsch., IV, 32. " Wheeler, B. J., Trans. Am. Phil. Assoc, 1901, 14. ^ Oertel, H., op. ciL, 102. 14 Studies in the Psychology of Language reason it is ever varying. Furthermore it exhibits a constant variation according to specific surrounding movements, as is the case in sound articulations of a fixed sequence. How it is that in spite of these modifying influences similar sounds are kept together so to speak to develop all along a definite line or to remain stationary will be shown in the following pages as being due to the consciousness of sameness (or functional identity) of speech sounds. For our purposes all gradual phonetic changes may be grouped into three classes: First, we have the case of a uniform sound shift that affects a certain sound in all words and under nearly all conditions as for instance the shift of Indo-European p to Germanic/. These we may call generic changes. Cases of uniform shifts under all possible con- ditions are extremely rare. Secondly, a sound under special conditions may be diverted from the type to which it belongs, a process which we may call a singular change. If this type happens to be changing, the sound may show either a retardation or an acceleration of the change in a number of words where a special condition exists, as the Indo-European p remained a /> if a spirant preceded, or it was shifted not only to /, but also to V if it happened to be the the final of an unaccented syllable. If the type happens to be stationary, the diversion from type can of course happen in only one way, namely, that a sound is changed under a special condition while similar sounds under other conditions remain. These two classes of changes have usually been called 'spontaneous' and 'conditioned,' terms which reflect the argumentation as regards the invariability of phonetic laws. Generally one of the 'conditions' of a change of the second class was easily recognized, as in the case of the stationary sp the preceding spirant, but did this mean that a change of the first class, as that of p to /, was unconditioned? Delbriick remarks truly: "Es ist anzunehmen, dass bei fortschreitender Erkenntnis sich fiir uns das Gehiet des bedingten Lautwandels gegeniiber dem unbedingten immer mehr erweitern wird."^^ Thirdly, a sound may change only in one word or in a number of words of a particular class and remain stationary elsewhere even where similar phonetic conditions seem to exist. This has often been termed a sporadic change, though at one time owing to the too free ^ Delbriick, B., Einleitung in das Studiutn der indogermanischen Sprachen, 150. Studies in the Psychology of Language 15 assumption of "sporadic changes" the term fell into discredit. Most (not all) examples of metathesis, distance association and dissociation are of this character, likewise changes in words of special use, as in greetings or auxiliaries. I have stated before that the historical material which gives us numerous cases of changes, namely generic changes of sounds or rather zones of variations under nearly all conditions, is one of the reasons for the assumption of a consciousness of sameness. It now remains to show the way in which this consciousness of sameness must be supposed to function in regard to phonetic change. What is presented in consciousness as a single sound includes in fact a number of different phonetic elements. A sound then repre- sents a zone and not a point, a sound change a band and not a line. Generally speaking the usually unanalyzed differences giving rise to responses termed the 'consciousness of difference,' (see above, p. 8) do not take part in a change; they are contained alike in the two zones at the beginning and end of a change, where beginning and end very often mean in a continuous process of shifting only two stages which happened to become reflected in the written documents of the periods. It is exactly that group of variations in quality which is of 'no account' and is therefore responded to by a unified reaction that takes part in the shift. At the same time the identity of function remains stable; we must suppose it to be at the final stage of the shift what it was at the initial, for the same sound occupies in both instances the same positions in the different words. (Cp. English'/Mn' : German 'Jiinn'; English 'fa/Aom': German 'Fac?en.') In other words, the consciousness of sameness is stationary while its actual content is being shifted. This fact is referred to by linguists when they state that gradual phonetic change goes on 'unnoticed' by the speakers. The explanation of this phenomenon is supposed to lie in the direction of attention. This is not entirely correct, if attention be taken with its usual conno- tation, for I can Hsten intently to every sound of the speaker and yet fail to notice any differences. It is the direction of the response called out by the speech sounds that creates the consciousness of sameness. The listener is not concerned with the speech sounds as physical phenomena or objects, but about their meaning, the attitude which the utterance of the speaker is about to create. Attention must be identified with the direction of the response; the analogy of the field 16 Studies in the Psychology of Language and its focus is misleading, or at least, while appropriate for 'content' of consciousness, it is insufl6cient in regard to its processes. The zone of variations of a speech sound, as far as only its quality is concerned, is the result of two groups of influences, the momentary conditions such as emotional and physical states, and the phonetic tendencies. Of these there are in the first place those particular ten- dencies which are characteristic of a language and which produce generic changes. In the second place there are the universal tendencies whose effects can be seen in many widely scattered languages, and they tend to produce those singular changes in a limited number of words, of which the assimilations are the most characteristic ones. Usually one is satisfied to 'explain' an assimilation by referring to the special condition which existed in the word before the change had taken place. Perhaps it is not the duty of philology to go further. But a linguistic psychology must ask the question: Why does a certain assimilation occur in one language and not in another, al- though apparently the same phonetic conditions exist in both? For instance, the sequence pt becomes in the Western Romance languages U (Latin septem becomes in Italian sette, Spanish siette, Old French set) while it remains pt in the Eastern languages (Roumanian sapte). It generally remains unchanged in the Germanic languages (Cp. English kept, German schnappte) except in isolated cases (The German auxiliary hatte from Old High German hapta). The fact is that not all the universal singular tendencies, although existing in most languages, are able to divert the specially condi- tioned sounds from their type or from the direction of the particular dialectic shift, but that the generic tendencies are stronger so to speak than the singular tendencies. Otherwise the historical material would present a picture of similar changes in all dialects wherein those corresponding to the generic changes would appear as 'except- tions.' As it is, the generic changes are the most easily recognizable traits of dialectic developments with the singular changes appearing as exceptions to such a degree that they have been and still are some- times so called. The question asked above should from this point of view be formulated thus: Why does a generic change affect a sound under all or most conditions; why are so few stragglers left on the road or others to take a different path? What is the process by which the zone of variations, as regards quality, is kept within certain Studies in the Psychology of Language 17 limits; why is it not more frequently split up into several zones, as indeed we ought to expect it on the basis of the many existing singular tendencies? Consciousness of sameness implies in a negative way the absence of awareness of any differences. It is sometimes quite remarkable how sounds are perceived as the same which are in fact widely apart. This appears especially in adapting one's self to a foreign language. A German learning English will almost invariably use the same initial combination / plus sh in 'germ' as well as in 'chamber'. In his language only this combination exists; he does therefore neither 'hear' the g (d plus z as in 'azure'), nor is he able to pronounce it until he is made to distinguish between the voiced and voiceless character of the two, i.e., until his field of stimulation is analyzed to a greater degree. Under such circumstances it may very well be that the generic and the singular tendencies, which I suppose to be active during the adult life of the individual, actually cause a broadening of the zone of variations in regard to quality. The results of the universal singular tendencies which work under special conditions in a limited number of words may actually be productive without the speaker's noticing it. Sounds affected in this way still belong to the zone of variations with which he once started his life's career as a speaking being. The identity of function for these sounds remains the same in his speech and consequently the consciousness of sameness in regard to them. For instance, if a generic tendency shifts the vowel e (as in 'bet') towards a closer pronunciation in the direction of i (as in 'bit'), it may well happen that a singular tendency works towards retaining the e before an r on account of the hollow shape which the tongue assumes in articulating this consonant. This actually happened in Gothic, where the e was generally shifted to i (cp. Gothic hilpan: help), while it remained an e before r (and Ji). At the beginning of this shift we may assume that two qualitites of e, one closer than the other, existed side by side unnoticed by the speakers. The only possible way it seems to me where the consciousness of sameness for a group of sounds might thus be broken up is in the learning process, but this may also cause a giving up of differentia- tions and consequently cause a narrowing of the zone of variations. 18 Studies in the Psychology of Language It is a well observed fact that children do not learn all the sounds of their mother tongue at once. Before they have attained complete mastery in speaking they substitute for an unacquired sound the one from among the acquired sounds which is nearest to it in acoustic effect or place of articulation. Thus for / and th they may use only /"; or instead of t and k they may use only t {free for three; tome for come). As the process of adaptation is going on, thechild will substitute each new sound as soon as it has learned it wherever it belongs, i.e., wherever its function demands it according to the language of the environment. A functional identity of / is thus eventually broken up into two; / and th. As soon as a new member is added to the sound system of the learner, a number of unfamiliar sequences are created which have to be mastered as well as the new sounds. This is a stage where the singular tendencies produce their quickest results. Child language abounds in assimilations. They seldom become strongly fixed by habits, but appear rather sporadically. It is only in words of a special use that they may be retained. In most words, usually in all, the assimilations are by and by replaced by well mastered sound sequences, since the identity of function insists on the putting in place of the new sound, according to the model of the speech of the group. It is important to note that such sequences are 'new' in the sense that they have been acquired through a mechanical process by the child himself and need not represent the variations of adult speech produced by the singular tendencies after the learning process had been finished. It seems possible then that singular changes occur in a minute way constantly in the speech of adult individuals without leading anywhere, because they are lost with them, while their children start on a fresh course; in such a way the zone of variations of a speech sound is narrowed by every new generation, because the articulation of a new speech sound is introduced into the phonetic material already existing as a type movement fresh from the mould, so to speak, in every instance and is joined to the neighboring sounds by movements that arise in a mechanical way. It is to be supposed that the subsequent development of the child's language may arrive at two different ends. On the one hand, his number of speech sounds, i.e., zones of variations, may coincide with that of the speech community of the adults about him and may remain so. This means that the child has definitely accepted the Studies in the Psychology of Language 19 phonetic material as handed down to him at the stage to which the tendencies had shifted it without allowing for the beginning of a new movement in the direction of a singular change (Diagram I, column 2). This must be the more universal case, since otherwise we should experience singular changes more frequently. On the other hand, the number of the zones of variations which his consciousness of sameness distinguishes may not coincide with that of his environment. That is, either two zones may have fallen together, which, although closely touching each other, must still have been separate in the speech of the older generations. (Note: Cp. the falling together of the Indo-European hh and the specially conditioned p at the end of an unaccented syllable in Germanic v) Or a zone of variations may be broken up in two so that a sound is diverted from the type to which it belonged and may now move into a direction prescribed by special conditions. In such a case the ob- jectively noticeable differences may be about the same in the speech of the adult as in that of the child, but where there is one zone of variations as regards the consciousness of sameness in the one, there are two zones in the other (Diagram I, column 4). For instance, a mother may pronounce the sound a with slight variations according to what other sound follows. In the diagram the small index figures are to indicate such variations. The child may learn for all the different a's just one sound a. But it will soon also develop the variations, because the singular tendencies remain the same. But its child in turn, the third generation, may split up the zone of consciousness in two so that it has substituted two sounds for one. DIAGRAM I First Generation Second Generation Third Generation 1 2 3 4 5 Variations Vowel heard Variations Vowels heard Variations produced and attempted produced and attempted produced Op a flp a dp a. a a. a a. ah a ah a ah ai a ai d di Or a Or a dr a„ a a«, a d„ 20 Studies in the Psychology of Language Such a difference of sharper distinction may also exist in the case of two persons of the same generation. At any rate, as soon as the distinction is a general one in the majority of the group there is the chance of the rapid development apart of the two sound qualities. If the position which I have here taken is correct, then one part of our question, namely, why is it that so few singular sound changes occur despite the constant presence of singular tendencies, has been answered. But the other part, why it is that singular changes occur in one dialect and not in another, is still left without an answer. And I do not see how the ultimate reason for this fact can be found anywhere except in the racial characteristics of the peoples speaking those dialects. My analysis has carried the problem only one step farther back. I have found that where a singular change starts, a zone of variations has been broken up. In place of a consciousness of sameness for one sound, we now have one for two. The analysis of the speech continuum has been more minute so that two types now share the place of one before. But whence does this finer sense for distinctions come? What racial trait does it indicate? Sporadic changes are those that occur only in one word and not in others although similar phonetic conditions may exist there. As a rule it is only our incomplete analysis or knowledge which causes us to suppose such a similarity. Often we overlook the additional conditions of increase or decrease in tempo or energy of articulation which are a constant accompaniment of such words and which do not appear in the written documents. But besides all possible variations of mechanical conditions, there is a wholly different factor involved in changes characteristic of certain classes of words. Greetings; exclamations of fear, wonder, and joy; curses and benedictions; proper names — all these appear to suffer more change than other words. Not infrequently it happens that these same words if used as a part of a sentence retain their regular form. As a greeting a speech community may use morn or moin, while it pronounces the word as morning when used as a part of a sentence. The case is somewhat different with words which are sometimes classed, rather vaguely, as particles, auxiliaries, etc. and which Meillet groups together under the expressive term "mots accessoires" .^* ^ Meillet, A., Introduction d I' etude comparative des langties Indo-Europeennes,* 13. Studies in the Psychology of Language 21 Of such a word there may occur two forms determined largely by the sentence accent, as Old High German nio-wiht and nicht which have merged into Modern German nicht. Common to both cases is that the phonetic type of a word has been split in consequence of a separation of a particular attitude, called out by a particular situation, from the group of all possible attitudes. A word, as I have stated above, out of the range of mean- ings determined by previous experience receives its particular meaning through the context. Our response towards it is in part determined by the responses to the surrounding words. Now if it happens that it frequently stands alone or in a phrase as a single response to a social situation (as a greeting), or if it appears frequently in the same function in a sentence (as an auxiliary), the particular reaction called out by it will become fixed as a definite experience and separated as a different experience from the reactions which may ordinarily be called out by it. The word then is no longer one, but two. Additional conditions, such as accent, may then produce a result in one, and not in the other. A necessary preliminary to changes of this kind is a specialization of meaning to such a degree that all or most other possible meanings are excluded. To be sure, words do assume all possible shades of special meanings in different situations, but those are processes of momentary significance only. Such special shades of meanings are accompanied by particular sound variations, which will not lead to a change, however, because they counteract each other. Only if one shade of meaning is singled out so that it becomes a distinct habitual response, the sound variation peculiar to this meaning may lead to a permanent change, because there are then no other variations to counteract it. It is not incorrect then to speak in such instances of the change of a word in contrast to sound change, since the word as a whole has broken away from the 'regular' development and follows its own particular path. Strictly speaking, the sound changes it suffers are just as 'regular' as those of other words; only they occur under very singular conditions. In connection with these sporadic changes it would not be uninter- esting to mention an older theory which tried to explain them as caused by a too frequent use of the words. Friedrich Schlegel said that a word might be smoothed off by much use and circulation 22 Studies in the Psychology of Language just like a coin. Schuchardt holds the same view: "Sehr selten gebrauchte W drier hleiben zurilck, sehr hdufig gebrauchte eilen voran."^'^ Thus howdy would resemble the smooth nickel, as compared with how do you do, the coin that has preserved its lettering nicely. Wechss- ler's reply: "Wenn wir die Sachlage vom Standpunkte des Psychologen beirachten, dann miissen im Gegenteil oft gebrauchte Worter ah Erinner- ungsvorstellungen oder Geddchtnisresiduen umso fester geworden sein^'^^^ is not quite to the point, because we might also say that much re- peated articulations afford the singular tendencies a better chance to differentiate the sounds. Frequency is a factor in these changes, but not the only one, nor the important one. Frequency of a particular response may separate it from other similar responses, i.e., one particular shade of meaning from the rest, whereupon we may have a 'sporadic' sound change. Mere frequency of the word in all its shades of meaning could not produce such a change on account of the counteracting influence of the varying conditions. One particular fact in regard to these words must not be lost sight of. Greetings, curses, and other expressions or words that are used with a restriction to one meaning are the objects of the creative powers of the speaker more often than other words. Thus a form like darned may be due to a special dissimilation of the nasals in damned, but it may also have been created at one time on the spur of the mo- ment. Summary: 1. The word as well as the speech sound has a distinct psychological existence, although the (minor) responses called out by them are fused in the (total) response called out by hearing the entire sentence. 2. The identity of function of a given speech sound under different conditions gives rise to like responses and consequently to a conscious- ness of sameness of that sound. Or in other words, consciousness of sameness of a speech sound is a unified response called out by a group of similar sounds that possess an identity of function. 3. The responses called out by the variations of a sound under different conditions produce the consciousness of difference, which need not be an actual awareness of differences. 2* Schuchardt, H., Uber die Lautgesetze, 24; cp. Schlegel, Fr., Sprache und Weisheii der Indier, 15; Thurneysen, R., Die Etymologie, 20, ^^ Wechssler, E., op. cit., 134. Studies in the Psychology of Language 23 4. A sound change means that the consciousness of sameness of a sound (identity of function) has remained stationary, while its actual contents have been shifted. 5. A generic (spontaneous) sound change is conditioned psycholo- gically through the consciousness of sameness working towards uniformity, especially in the stage of language acquisition of children. 6. A singular (conditioned) sound change is conditioned by the ever present possibility that the consciousness of sameness of a single sound (in the older generation) may appear as a discrimination of tivo sounds during the stage of lanuage acquisition (in the younger generation). 7. The position of a word in which it occurs very frequently with a definite shade of meaning may be singled out through the specificity of function from the other positions in which it may assume a variety of meanings. The counteraction of consciousness of sameness is then suspended and that word may suffer a sporadic change when functioning in that particular shade of meaning. B. THE FUNCTION OF IMITATION IN SOUND CHANGE I The dogma of the uniformity of sound change (see p. 9) naturally was the starting point for the discussion of the question whether the whole group took part in a sound change or whether it originates in one or a few individuals and then spreads to the rest by imitation. This issue did not produce such a clear division among philologists as the dogma itself. On the whole the more common opinion is that imitation not unlike fashion causes one member of the group to talk like the rest and to follow the phonetic development of its dialect. (For the following compare especially E. Wechssler, debt es Laiitgesetze? , 176ff ; and H. Oextel, Lectures on The Study of Language, 136.) Whitney says: "Not that the speakers of a language act in unison and simultaneously to produce a given change. This must begin in an individual, or more or less accordantly in a limited number of indivi- duals, and spread from such example through the community. .... What he (the individual) suggests by his example must be approved by the imitation of his fellows, in order to become lang- uage."^® Even more definitely Collitz states: "Dze lautliche Um- wandlung .... iihertrdgt sick allmdhlich auf eine grosser e Anzahl von Individuen. Sie gefdlU denen, welchen sie aufgefallen ist, sie ivird Mode."^'' This view has had strong supporters in modern times. Meringer expressed himself in the following drastic way: "Ein Lautgesetz ist um nichts merkwiirdiger als das Gesetz des roten Schirmes und schwarzen Kopftuches beim Bauernweib, der Krinoline, der Puf- drmel zu andern Zeiten.'^"^^ An entirely opposite view is held by Rousselot: ^^La cause dS- terminante de revolution (phonetique) est d'ordre general; elle agit sur » Whitney, W. D., article Philology in Encycl. Brit. (1911), v. XXI, 424. *' Colhtz, H., Anzeigerf. deutsches Altertum, 1879, 321; cp. also Miiller, F., Techmer's Zeitsckrift, I, 213; Bezzenberger, A., Gottinger gel. Anzeiger, 1879, 651; Schuchardt, H., op. cit., 12. ** Meringer, K.,Aus dem Leben der Sprache, 239; cp. Meyer-Luebke, W., Grammatik der rumanischen Sprachen, I, 69; Bremer, O., Deutsche Phonetik, p. xiv; Sievers, E., op. cit., 267. 24 Studies in the Psychology of Language 25 la masse de la population. C'est une sorte d'epid^mie cL laquelle personne n'6chappe."^^ This evolution is prepared in the parents, but it 'breaks out' in the children: "Ce principe est dans V enfant. Ou bien c'est une tendance absolue et hereditaire qui le porle a modifier dans un sens determine le jeu des organes de la parole."^° This seems to be borne out by the facts observed by Rousselot himself and also by the re- markable case presented by Gauchat, who found that the differences between the speech of two generations in one Swiss village (Charmey) were exactly paralleled by those found in another village (Cerniat), although there was no intercourse between the two.^^ It is interesting also that Gauchat states emphatically: '^Nos matiriaux n'' of rent aucune trace d'influence personelle J^ai etudie, d'une fagon sommaire, environ 50 langues individuelles et je n'y ai rien trouvi dHndividuel.'^^^ However, he does not deny imitation absolutely; he notices it in cases of 'infiltration' from other dialects. But how shall we reconcile his statement which I have quoted with that of Weigand, another careful observer of dialects: "Die lautlichen Verdnderungen gehen vom Einzelnen aus"^^? Can it be possible that linguistic conditions in Roumania are so essentially different from those in French Switzerland that one could arrive at exactly opposite conclusions? Hardly so. It seems that Gauchat is speaking only for communities that lead a relatively secluded life, while Weigand devoted himself to a large group of dialects where he constantly met those 'infiltrations' which the former has also noticed. An intermediary position is taken by others. Oertel, leaning more towards one side than the other, says: "Only a small number of linguistic changes of whatever kind arise simultaneously in many individuals. The bulk originates in one individual and gains currency when imitated by others."^* Wundt's position is somewhat reversed: "Nun spielt zwar die Nachahmung im gesellschaftlichen Leben iiberall eine mitwirkende Rolle von nicht zu unterschatzender Bedeutung, aber bei den tiefer greifenden und allgemeineren Verdnderungen kommt ihr niemals die Eauptrolle zu. Vielmehr erweisen sich diese Verdnder- " Rousselot, J., Rev. des patois Gallo-Rom., V, 413; cp. Meillet, A., op. cit., 15, " Rousselot, J., Rev. des patois Gallo-Rom., V, 412. " Gauchat, L., op. cit., 227. ^UbU., 231. » Weigand, G., op. cit., 19. ^ Oertel, H., op. cit., 147; cp. Delbruck, B., op. cit., 98. 26 Studies in the Psychology of Language ungen, wo wir imstande sind ihren Bedingungen nachzugehen, regel- massig als solche, die nicht von einem Individuum und nicht einmal von einer bestimmt hegrenzten Zahl von Individuen ausgehen, sondern auf Einflussen beruhen, die entweder die sdmtlichen Mitglieder einer Gemeinschaft oder mindestens deren iiberwiegende Masse tre^ffen."^^ Wechssler makes the distinction that a common change is possible only in a small community whose speaking is controlled only by hearing, while in the written and spoken language of a larger district, the KoivT} or Kultursprache, we have a single starting point for a change.^® It seems then that the two kinds of changes are existing. Indeed, even the historical material, aside from all studies of modern con- ditions, forces us to account for both. I shall call imitative a change that originated in one or a few individuals and spread to others by imitation; and a change that takes place through a tendency in the majority of the group I shall call tendential. These terms are not a substitution for Oertel's 'primary' and 'secondary' or 'imitative' changes, since he is evidently using these only in regard to the language of the individual.^^ I shall use the terms 'tendential' and 'imitative' as regards the language of the group. This change of terminology is desirable since, as will be seen below, a 'tendential' change affecting the whole group is already a compound of primary and secondary changes as regards the individual. There are some important reasons why we must postulate a tendency which it would be difficult to point out as an active force by direct observation. (If Gauchat's observations (see p. 25) could be repeated in the same villages at intervals of perhaps ten years, we might then actually 'observe' a tendency.) If all sound changes started in individuals then the development of a dialect would be entirely haphazard. We should expect to find a zig-zag line, because different individuals would add different innovations, which are in turn imitated by the rest. The historical material which we possess shows in many cases a remarkably straight line, which points to a persistent sound change in one direction for many centuries. It is inconceivable that from generation to generation individuals should » Wundt, W., op. cit., 22. *• Wechssler, E., op. cit., 36. »' Oertel, H., op. cit., 136. Studies in the Psychology of Language 27 spring up who would all produce innovations in the same direction, or that the rest of the community should imitate only these individuals and no others who might have produced other deviations. It is therefore best to conceive of such changes as being produced by a majority of the group. Furthermore, since they are spread out over a large period of time, we must suppose a constant ultimate cause — in the last analysis perhaps some racial characteristic — underlying the whole movement, in other words a cause producing a tendency to shift certain sounds in one definite direction. Some members of the group will always be unproductive, just as we always find some that do not exhibit the peculiar traits of the group. The others probably will take a more or less active part in the tendential change. Despite these individual differences we always have a comparative uniformity in the dialect of a group. This process of levelling out is due to imitation. Thus every tendential change is is in the end a resultant of a tendency and imitation. My own position can be formulated thus: If we find a change that has persisted for some time in the same direction we must suppose that it is caused by a tendency which affects a majority of the group, individual differences between the members of the group being leveled out by imitation. Besides this partially tendential change, a purely imitative change seems possible, as is in fact attested by several historical instances, e.g., the spread of the uvular r in Germany. Such changes take their origin in one or a few individuals and may be due to various causes, to an individual tendency, to a defect in articulation, or even to mistakes. As regards the way in which they affect the community they are purely imitative. Again the ultimate cause is not imitation, but some cause that lies beyond it, as superiority of the leaders, class distinction, etc. It is due to these that among the many in- dividual innovations in a community only a few are selected and made universal. As to the character of both of these types, the tendential and the imitative, we can state with some definitness only in regard to the first that it is gradual in its movement. It may proceed faster or slower under different circumstances, but hardly in leaps and bounds. That is, if a tendency works at all, we must expect a gradual shifting, say from p to /, and not a substitution. On the other hand it is 28 Studies in the Psychology of Language probable that an imitative change is usually sudden, i.e., proceeds by substitution. In the majority of cases, e.g., pure substitution, dissimilation, some distant assimilations — this is fairly certain, while in others the character is more difficult to determine, as in umlaut. At any rate we can say that sudden change is always imitative, while gradual change may be imitative or tendential; or tendential change is always gradual, while imitative change is as a rule though not always, sudden. In case it works in connection with tendential change it is always gradual. For my purposes here I shall draw the line a little more roughly and speak of pure imitative change as if it were always sudden in contrast to secondary imitative change, the complement of tendential change, which is always gradual. DIAGRAM II Tendency ^_^_____— Imitation tendential imitative i .-—-'"1 gradual *---" sudden The real difficulty for philology lies in determining whether a sound change, of which only the beginning and end have been re- corded, has been a gradual one or whether it does not go back to a sound substitution by a single individual adopted by the rest. Some are clearly recognized as imitative changes, as the introduction of the uvular r in the dialects of many German cities. Others are doubtful, as the German umlaiil. Others are clearly recognized as tendential, as 'Grimm's Law.' II The function of imitation in speaking is twofold. L In communities in which the individuals show differences in the amount of variations, i.e., where some go further along a certain direction of an existing tendency than others, imitation has a levelling effect; it causes the members who are almost unaffected by the ten- dency to accept at times the pronunciation of those who are greatly affected and vice versa. 2. In a community where an individual or a number of individuals have produced an innovation in pronunciation, this may be accepted through imitation by the rest of the community. What causes the Studies in the Psychology of Language 29 other members to imitate is a different question. Imitation is here only the immediate cause, the instrument through which a more remote cause can be active. Uniformity is the result of imitation in both cases. In the first, however, it is a tendency producing the change, while in the second it is imitation, provided the change has started at some particular point. More clearness in regard to the functioning of imitation will be gained by analyzing the conscious processes involved in it. Most psychologists^^ regard imitation as an instinct. Others have rejected such a theory. Wundt, while rejecting the idea that imitation is an instinct or some other kind of special mental function, would reduce the processes that fall under this term to an association between af- fects and actions.^^ The defect of such a theory is that the crucial point of the question, namely, how the association, supposing its existence, has come about, is left unexplained if one does not choose to leave it entirely to accident. But there are many cases where we cannot assume a previous association between feeling and action, and where we can do this, we must wonder why such a general feeling, say of pleasantness, is all that causes one action in one instant and an entirely different one in another. There is one analysis of imitation which, besides covering a multi- tude of facts, though not all as I intend to demonstrate, has the advantage that it emphasizes the purely social character of imitation. It is based on the observation that a person who produces a social stimulation, especially a vocal gesture, is thereby stimulating not only another, but also himself, and in response thereto he will naturally fall into the ways of acting like some other self. In taking the roles of others as a response to self stimulation we get 'imitation,' this is Professor Mead's point of view. "The response to the social stimulation of the self may be in the role of another — we present his arguments in imagination and do it with his intonations and gestures Not that we assume the roles of others toward our- selves, because we are subject to a mere imitative instinct, but because in responding to ourselves we are in the nature of the case taking the attitude of another than the self that is directly acting, 38 James, W., Principles of Psychology, II, 408; Angell, J. R., Psychology, 360; McDougall, W., Social Psychology,^^ 105. 3» Wundt, W., op. cit., 39. 30 Studies in the Psychology of Language and into this reaction there naturally flow the memory images of those responses of others which were in answer to like actions."*" It is important to note that such an analysis does away with the distinction between conscious and unconscious imitation as two different types, for it is clear that the reaction may assume all kinds of shades of awareness, from the 'unconscious' to the intentionally produced. This taking of the roles of others is most dramatic in childhood. The child, if left to himself will create imaginary companions to take part in the events of his play. His consciousness is then a stage on which his own self is only one among other actors. These may repre- sent persons of his actual little community, or they may be entirely fanciful. But even the kings, the fairies, and the goblins bear a strong resemblance to those that move in his everyday life. During this period the child acquires his habits of articulation. This process is not quite as mechanical as Watson supposes.*^ A large part of it is really 'social,' i.e., depending on imitation in Meads' sense. The problems of the succeeding steps of adaptation and of the degree of exactness of adaptation of a child's language are such that can be treated by close observation of particular cases and by a statistical method. However, the problem to be treated here is in regard to the adapta- tion which is going on in the speaking process of an adult after he has mastered the language. The child's play has now been abandoned. But it is only the proscenium, the wings, and the setting of the stage that have been left behind. The acting has remained in its essential traits and is now called thinking. At any moment, of course, a stage can be re- built with all the colors and reality of childhood play, but it is seldom that we do so. Thinking is the social process of converse brought within one consciousness. We still have stimulus and response as between two selves, but these two selves are now the agents of but one consciousness; they are actors whose personalitites are quite faded out as compared with the child's imaginary playmates.*^ *" Mead, G. H., Jotir. Phil. Psych, and Sc. Meth., X, 377. " Watson, J. B., Behavior, 330. « Mead, G. H., Jour. Phil. Psych, and Sc. Meth., X, 377. Studies in the Psychology of Language 31 The well known observation that a child has a greater adaptability to his environment, that he learns a language easier than an adult has often been referred — and without doubt correctly — to the greater plasticity of his neural organism. But may it not be that the difference in adaptability is also due to the loss of realistic appearance of the agents of the 'inner forum'? If we now come to apply Professor Mead's fundamental analysis of imitation to the facts of sound change outlined above, we have to enlarge his conception. Imitation exists not only where one's self takes the role of another, but also, where he in taking his own role produces certain characteristics which were formerly not his own, but the other's. In the first and fundamental case the entire response is that of the other; in the second case the major response — the utter- ance of the sentence — is essentially his own and yet it contains minor responses belonging to the other self. He has borrowed so to speak for his own action some of the colors of the make-up from the other actor. In as far as such minor responses, one particular sound or word or accent, are concerned we might still speak of taking the other's role, but it seems better to consider these modified processes as a different type of imitation. This partial imitation — as we may call it in contrast to Professor Mead's total imitation — is the one with which we are primarily concerned in analyzing the mechanism of sound changes. It appears in two different types accordingly as the character of a sound change is gradual or sudden, tendential or imitative. Now let us examine a little more in detail how one individual imitates another's speech, which already contains the new pronun- ciation that is to become a sound change. In total imitation, i.e., in taking the other's role entirely, along with the familiar intonation and sound qualitites of that person a new accentuation of a word may be imitated, as address ior address*^ or a new vowel quality as a (as in father) for ce (as in had) in a word like past. To the imitating individual all the elements imitated are primarily those that form the basis of the consciousness of difference. I have pointed out elsewhere that consciousness of difference does not imply an actual awareness of the factors of differentiation, it only implies an attitude of awareness towards the personality of the ^' Bloomfield, L., The Study of Language, 152. 32 Studies in the Psychology of Language speaker and his emotional and physical status. To imitate totally then means to step into the center of that self towards whose speaking one has responded frequently with a consciousness of difference and to act from that center so that the own self would react again in the same way with the same consciousness of difference. It means to create with one's own action a response from the self as if another well known self were stimulating the own self. When the individual again takes his own role, he of course relaxes into his own usual ways of speaking. But now, two things can happen. He may use, while always acting in his own role, a word like past with a new sound a substituted for the old (b, and he is usually aware of such cases. Or he may pronounce a certain sound throughout with the slightly different quality characteristic of the pronunciation of the other speaker, e.g., an initial t as an asporated t; of such cases he is usually not aware. These two cases of 'partial' imitation, which are typical of the two sound changes, the imitative and the tendential, are, so it seems to me, sufficiently different from Mr. Mead's case of 'total' imitation to be treated as separate types, and they are also sufficiently distinct from each other to require separate analyses. They may to some extent depend on previous 'total' imitation, but not necessarily so. For in the first place, the taking of the roles of other individuals is in ordinary thinking a rather colorless process in adult life as compared with the dramatic realism of a child's thinking or playing. And in the second place we often use peculiarities of another's speech without having had a chance to imitate him in inner speech or even in 'thinking aloud.' Ill The first case of partial imitation is prevalent in the kind of sound change which I have called imitative, i.e., where a certain inno- vation of pronunciation by one or a few individuals is accepted by the rest of the community. Clear instances of this kind are those of substitution. Suppose that in a certain community the older pronunciation oifast with cb (as in had) is being replaced by a pronunciation with a. An individual who is still using the old pronunciation but hears occasionally the new, has in his vocabulary two doublets, i.e., two words that are slightly different, although they answer to the same Studies in the Psychology of Language 33 meaning; one form is in his 'active' vocabulary and the other, the new, in the 'passive.' Some additional cause, as that of fashion for instance, must then be active to induce him to use the new word. There may now be a period where the two forms are used inter- changeably, but if that additional cause is still activei, then the new form will soon gain the preponderance and replace the old one entirely. Imitation in such a case reduces itself to the substitution of a response of other individuals brought about by the fact that both responses, the one substituted and the old one, fit one single stimulus. Awareness of such a change is a common characteristic, since it does not take place within the consciousness of sameness for one sound. The second case of partial imitation is a necessary hypothetical process wherever a tendential change takes place as pointed out before. This kind of imitation must be supposed to work at any given time towards the leveling of the variations caused by the uneven strength with which a tendency probably works in dififerent members of the community. In addition it must be supposed to work also between variations that are due not to common tendencies, but to individual tendencies. Any one with a pronunciation differing from that of the others is a potential factor of imitative sound change for the rest of the com- munity. It is very unlikely, however, that he should really cause a common sound change, unless an additional cause is active. His own difference may also be so great that to a hearer's consciousness of sameness he pronounces a different sound and consequently different words. In such a case an imitative change may come about in the way already described, by imitation through substitution. The second case of partial imitation is, however, easily noticeable wherever one individual enters another community with a dialect slightly different from his. Without noticing it he undergoes a process of transformation until he is in all respects a regular member of that community. Since such a case of single absorption exemplifies in my opinion also the mechanism of leveling, which is an absorption of each by the rest, I shall try to analyze only such an adaptation of one individual. When he enters the new community the contents of some of his zones of variations are slightly different from those of the original 34 Studies in the Psychology of Language members. His differences will probably not be noticed as such by them, but they will enter their consciousness of difference in regard to his speech, just as he has a somewhat uniform consciousness of difference in regard to all the others. (Chinese faces are all alike to us as long as their racial difference is responded to by us as an indivi- dual difference.) Without doubt the type of imitation analyzed by Professor Mead will now play an important part in the thinking process of the strang- er. The roles in which he will respond to his own stimulations will all have an aspect of uniformity by reason of that uniform consciousness of difference. Such an experience cannot be without effect upon his own actions. But it cannot be all, there must still be another type of imitation when he acts in his own role, for it is this which is gradually transformed. This experience of taking the roles of others, important as it is, does not seem to me even necessary to bring about the transformation. What is necessary only is the hearing of the speech of the others. One of the functions of that complex of past sound experience which we may call the auditory speech image is its control of articula- tion. This does not mean that an ordinary image actually precedes an articulation like a picture set up as a model. On the contrary, the image is usually merged in the action and its incoming sensations. The auditory image is not a stable and perfectly rigid thing, it is a psychic process which is compounded of past experience, especially that of the most recent and the most vivid. If this experience varies, or more precisely, if it shows a preponderance of a certain variation in one direction, as in hearing the sound / often pronounced /'', it does not remain unaffected, it also changes in that direction. Whether the hearing of others' or of one's own pronunciation plays the dom- inant part in this experience is a moot question and difficult to determine. But at least we are certain that the whole of the ex- perience, including taking the roles of others, is a potent agent in this respect. The speech image thus affected will in turn influence the articula- tion. At first the variation in the direction of the change will become more frequent, and at last the center so to speak of the zone of varia- tions will have been moved in that direction. Such a change takes place unnoticed, unless it is experienced by a phonetician instead of an unsophisticated individual. It is a change Studies in the Psychology of Language 35 within the consciousness of sameness of a sound. The response in regard to that particular sound will have the same function all the time, and will consequently appear as the 'same' in consciousness, while its content has actually undergone a transformation. We have here then a type of imitation which consists in the change of a response, while its function remains identical, through the modification of one of its controlling factors. DIAGRAM III Imitation (taking the role of another) gradual sudden (change of controlling factor of response) (substitution of response) The difference between this type and the one described above lies in the fact that here in regard to its function the response is the same, although slightly modified in contents, while there one response is substituted for another, both serving the same function. This difference accounts for the fact that the element of awareness is always absent in the one type and may be present in the other. Mathematically considered, both cases might be called substitutions, but processes of consciousness have to be considered from their own point of view, which is here that of conduct. Doing this we find a radical difference: Consciousness of sameness of response, and substitution of responses. The situation is quite different, if our hypothetical case occurs in a community whose speech contains sounds that are unfamiliar to the stranger. Unfamiliarity of a sound means that it is not con- tained in any of the zones of variations of the speaker's language. First there will take place the substitution of a familiar sound which comes nearest in acoustic effect to that of the unfamiliar sound in place of this one. Thus the word that may be pronounced dat by a Swede and zat by an Italian. Unless by some additional process the speaker learns the new sound, he will not be able to produce it. The matter seems paradoxical: A considerable variation which can with the proper guidance easily be detected by any one, is not 36 Studies in the Psychology of Language imitated, and a slight one which goes unnoticed is imitated. How can a modification be imitated that is not perceived as such? The answer lies in the mechanism of the control of habitual movements through the corresponding images. A zone of variations of a particular articulation contains all the possibilities of this set of movements. This zone is accompanied by a range of shades of the auditory image which is in a certain sense parallel to it. The proof for the existence of such shades, although we may not be aware of the difiFerences, is afforded by the possibility of intentionally producing a speech sound which fits only one situation, as in the case of acting or other kinds of taking the role of another. Now what is imitated unnoticed by an individual in a strange community must already be contained in his possibilities of articulation, i.e., in his zone of variations: One particular shade of the auditory image gains the preponderance over the others and along with it goes the cor- responding variation of articulation. In the case of an entirely diflFerent and unfamiliar sound there is nothing in the past experience of the individual that can be influenced. The imagery called out by the sensation, combining with it to form the perception, functions as in any ordinary auditory illusion. And there can never be an imitation until the different articulation has been produced by the speaker through a process of conscious syn- thesis, as in foreign language instruction based on phonetic methods. A foreign language course is full of examples of this kind; it proves this point almost every day: Familiar to many is the way in which such German sounds, as the so-called ich and ach sounds are acquired by American students. The only possible road is a conscious syn- thesis of the new articulation followed by a constant drill until it has become a habit, i.e., a new zone of variations. Summary 1. Imitation is not a specific process of response. While there may be outward resemblances, as the copying of a model, different types of processes, each with its own essential characteristics, are found in different sets of cases. 2. In addition to the fundamental type (I) of 'total' imitation analyzed by Mead as a process of taking the roles of another, we must on the basis of linguistic facts recognize two partly related types of 'partial' imitation; namely, (II) Imitation through (conscious) sub- Studies in the Psychology of Language 37 stitution of one response for another where both fit the same stimulus, and (III) imitation through the (usually conscious) modification of the controlling imagery of a response. 3. Imitation through a modification of the controlling speech image can lead only to an action that has previously been executed; a familiar sound can be imitated, but not an unfamiliar one. 4. Gradual phonetic change may be both tendential and imitative. To which of these a certain sound change belongs must be determined on the basis of the historical material. 5. Tendential change is possible in a majority of the group while the rest of the group may follow through imitation, just as the differ- ences between individuals will be leveled out by this process. Imita- tion in this cases will be of type III. 6. A purely imitative change is usually a sudden change. After it has taken place in the pronunciation of one or a few individuals it is accepted through imitation by the rest of the community, but not without some additional cause. Imitation in this case will be of type II. LIBRARY USE ^RY RETURN TO DESK FROM WHICH BORROWED LOAN DEPT. THIS BOOK IS DUE BEFORE CLOSING TIME ON LAST DATE STAMPED BELOW I elow. jr UBWVW USt i:eb-&«^ ML REC'DLD FFR LD 62A-30m-2,'69 (J6534sl0)9412A — A-32 General Library University of California Berkeley f 2»60 j^4tf^ 5 71-lPW^^ i^^ ,:63 -1 PW 5W?44 o7 -5 PM U.C.BERKELEY LIBRARIES b