Repository | Book | Chapter

File clean-up November 21, 2021, 4:53 pm sdvig press ( )

1Originally published in Semiotica 12/2 (1974): 97-128.

2Every language has various hierarchically ordered units such as distinc|tive features, phonemes, morphemes, words, sentences, and texts. These units are always discrete in nature, but not, however, disparate or ordered with no relation to each other. In the act of language they occur in manifold relations forming increasingly complex units at higher and higher levels. Nor do the relations occur unconnected and independently of each other. They originate partly out of the same motivations and obey the same laws; they partly imply each other; they are partly revealed through transformations. Thus the predicative relation in the statement

(1) My car is fast.

3can be expressed attributively by my fast car. Basically it can be said that we understand a language only when we have mastered two things: relations and transformations.

4Our main concern in this essay is not with the question of trans|formations, which at present dominates linguistic research, but rather with the determination of relations, which constitute language before all trans|formations. The theory of the two axes of language will serve as the frame of reference. N. Kruszewski (1886, 171–74) first introduced this theory; F. de Saussure (1931, 170–75) took it up with one restriction, which will be discussed later; and finally R. Jakobson developed it into “a harmonious, unified and exceedingly fruitful theory of language”.

5What we have said here of Jakobson are the words he himself used in praise of Kruszewski.1 Actually the most comprehensive description and typology of the two axes and the corresponding operations stem from Jakobson. Moreover, it was he who demonstrated the usefulness of this theory by his excellent application of it to two areas of linguistics, aphasia and poetry, which had previously been in a state of chaos from a linguistic point of view and which then suddenly proved to possess a simply structured skeleton (1971b, 239ff.; 289ff.; 1960, 358ff.). Finally, subse|quent to Jakobson’s contributions, Lacan (1966, 493ff.; cf. Rifflet-Lemaire 1970, 307ff.) applied the same theory to the mechanics of the unconscious.

6The theory essentially states that every linguistic unit is extended along two axes. It appears in combination with other linguistic units that together form its context, just as a unit itself serves as the context for the simpler units of which it consists. On the other hand every unit and every group of units in a message represents a selection from a storehouse of units, the code, which can be substituted for the unit without making the message meaningless. On the first axis, designated as syntagmatic by Saussure, we have for example the sentence

(1) My car is fast.

7On the second axis, for which Hjelmslev’s designation “paradigmatic” has asserted itself, we choose the noun car from a series of nouns, which could all take its place without preventing the syntagmatic combination from forming a meaningful sentence; vehicle, automobile, etc.

8The two activities, selection and combination, correspond to the two most treated forms of association in traditional psychology and to the two best known figures of speech. The units along the axis of selection are related to each other through similarity, and thus share an aspect with the figure of speech, the metaphor. The components of combination are related to each other through contiguity and therefore have something in common with metonymy. This results in two columns of polar determinants:

PARADIGMATIC AXIS	SYNTAGMATIC AXIS
selection	combination
substitution	contexture
similarity	contiguity
metaphor	metonymy

9We shall now examine this theory of the two axes of language from a phenomenological point of view, that is, through a descriptive analysis of the relevant aspects of language. Our starting point will be the use of this theory in psychoanalysis. A theory is most likely to betray its weaknesses, but also its potential through secondary applications. The revised theory will then be substantiated and become more powerful, if it is able to integrate the analyses of various linguistic phenomena that have been carried out, among others, by Jakobson himself, and if it can systematically fit these analyses, which have hitherto floated in “splendid isolation”, into a unified theory of language.

1 | Critique of the psychoanalytic adaptation of the theory of the two axes

10Lacan and his circle took over the determinants of the two axes to characterize the two mechanisms of the unconscious, condensation and displacement, as worked out by Freud.

11The first difficulty with this adaptation lies in the exclusive pairing of similarity with condensation and contiguity with displacement. Conden|sation occurs not only when a gentleman offers to begleitdigen a lady (Freud 1941, 77) (condensation of the two intentions begleiten “to accompany” and beleidigen “to insult”, the lady, based on the phonetic similarity of the two verbs in German), but also occurs in the case of so-called Mischpersonen, “composite figures”, as often appear in dreams (Freud 1942, 330f.). A is wearing B’s clothes and doing C’s job. We have here a condensation of three persons on the basis of complex relations: a relation between content and container (person-clothes), and a functional relation (gardener-hoe), both of which, from a purely external point of view, represent a spatial contact.

12In like manner, it is not only a matter of displacement when a conti|guous relation exists between the manifest content and the latent thought of a dream, as in the dream analyzed by Freud (1942, 154; 181) of a supper party, in which smoked salmon actually signifies the lady whose favorite dish it is. A displacement can also occur along a chain of similarity when for example Freud, after dreaming of turning over a color plate in a monograph on botany, is reminded of a picture book that he tore up as a child (1942, 178), or when a patient dreams of a broken bone when in reality his marriage is breaking up (1942, 413).

13If the exclusive pairing of condensation with the association of similarity and of displacement with the association of contiguity poses problems, then the pairing of condensation with metaphor and displacement with metonymy is consequently not valid either. The last example above demonstrates a metaphorical relation (often the subject of jokes) between a broken bone and a broken marriage, which supports an unconscious displacement.

14Nor can the two concepts, substitution-combination, be split up dichotomously between “the two governing factors to whose activity we may in essence ascribe the form assumed by dreams” (Freud 1942, 313), condensation and displacement. Freud writes in his book on dreams (1942, 344): “The displacements we have hitherto considered turned out to consist in the replacement (Ersetzung) of some one particular idea by another in some way closely associated with it.” After distinguishing two kinds of displacement, a first in which an important element is represented by and contiguously related to an indifferent one (e.g., the relation between the smoked salmon and the rival whose favorite dish it is), and a second in which a concrete, visual dream-content replaces an abstract, colorless dream-thought (e.g., when a piece of coal in the dream about a Wagner performance replaces the “secret love”), Freud goes on to say (1942, 344f.; emphasis mine): “... the outcome of the displacement may in one case be that one element is substituted by another, while the outcome in the other case may be that a single element has its verbal form replaced by another.” Furthermore, in his treatment of condensation Freud explicitly speaks of “fusion and combination” (1944, 176). How else could the example of condensation quoted above, namely the offer to begleitdigen “accompany” — “insult” a lady, be described?

15We come to the conclusion that the two columns set up by Lacan’s psychoanalytic circle, condensation — substitution/selection — similarity — metaphor and displacement — combination — contiguity — metonymy, lack validity. What is the explanation for this blatant mis|interpretation, which not only contradicts the facts, but also the text in Freud’s own analysis? We suggest that in pairing off the various concepts with the two axes of language not enough attention was paid to their explication and differentiation, which will be our concern in the following.

2 | Differentiation of the system of the two axes of language

Metaphor and metonymy

16Jakobson ascribes a metaphorical character to the paradigmatic axis of language and a metonymical character to the syntagmatic axis. To avoid misunderstandings it is important to note that the concepts of metaphor and metonymy are here used in a metaphorical, i.e., limited, sense, namely in terms of their respective relations to the rules of association of similarity and of contiguity, but not in terms of the substitutive character which they both share as figures of speech. Instead of the sentence

(1) My car is fast.

17I can also say:

(2) My carriage is fast.
(3) My rattletrap is fast.

18Carriage and rattletrap are metaphorical expressions for car. In the case of rattletrap the comparison rests on the sound, while in the case of carriage form and function are involved in the comparison. However I can further say:

(4) My engine is fast.
(5) My Deux chevaux is fast.

19thus replacing car with metonymical expressions. In the case of engine, a complete part, a piece, represents the whole; in the case of Deux chevaux, which is an elliptical expression for a car with a two horsepower engine (thus the singular), an incomplete part, a feature, represents the whole. Although Jakobson uses metonymy to characterize the syntagmatic axis, we must not forget that it also operates along the paradigmatic axis when it functions as a figure of speech.

The combinatorial aspect of substitution

20The condensations described by Freud turned out to be combinations. However, a combination may not always be so overt as in the example of begleitdigen. The fusion can be such that one must grasp more than just the phonetic utterance in order to recognize the unconscious motivation in a dream or in parapraxis (Fehlleistung), for example a slip of the tongue. We can understand the sentence from an inaugural address, cited by Freud (1944, 27)

(6) I am not inclined [geneigt] to appreciate the accomplishments of my highly esteemed predecessor.

21only if we link the participle geneigt “inclined” with the similar sounding geeignet “qualified”, both of which often occur together. More or less overt combinations of this kind can also be encountered among the figures of speech in ordinary language. In order to grasp that reference is being made to a car in sentence (2) we must be able to conjure up the carriage-like features of a car, as cartoonists sometimes do, i.e., to visualize a cross between an automobile and a carriage. The same is true of metonymical expressions. We can grasp the full significance of the sentence

(7) My Ford is fast.

22only if we link the car with its manufacturer.

23What is linked together here need by no means be verbalized. It often happens that we have no direct way of saying what we have expressed metaphorically, and that we can understand what has been said long before we are capable of paraphrasing it, e.g., when we speak of light and dark sounds in phonetics.

24Within the code from which we choose successive entities for combi|nation in a meaningful message, more or less latent or — to use a positive and an active expression — virtual strings of combinations already exist, which are delivered along with the message. So far the relations of simi|larity and contiguity have acted as binding agents for these paradigmatic combinations.

25A comparison of the sentences

(1) My car is fast.
(8) My car is red.
(9) My car is outdated.
(10) My car is slow.

26soon reveals that the first predicate is related to the last in quite a different way than to the other two. Although we also grasp the characteristic of speed only insofar as we can distinguish it from those of, for example, color and age, speed does not imply these characteristics in the same definite way as it does slowness, without which it cannot even be thought — whether or not is verbalized in this or in another form (e.g., not fast). Opposition represents another form of paradigmatic combination. As a figure of speech, it is known as antonymy (palace-cottage).

27Opposition plays a dominant role at the lowest level of language in determining the distinctive qualities of phonemes. Phonemes can most efficiently be distinguished through bipolar opposition. Thus in the papa and tata of child language, the phonemes /p/ and /t/ are distinguished by the opposition of their light and dark qualities. The perception of darkness in the /p/ of papa implies the co-presence of a light phoneme. The co-present lightness is not manifest to the senses as the darkness is. Nor is it, at least not necessarily, imagined along with darkness. Lightness is present only in a virtual or intentional manner, just as the front view of a house implies the other sides which are not visible to the senses. Unlike Saussure and the French phenomenologists who have been influenced by Hegel, Husserl preferred the terms presence-appresence to presence-absence in order to describe the relationships of the implication of sensory perception. These mutually implied members are present in our mind in different modes: the one is sensory and direct; the other intentional, i.e., in a reference, and indirect. This reference is however not accessory to, but rather constitutive to the sensorily perceived member, without which it could not occur in the way it actually does occur, in our case as a dark phoneme. This reference is only virtually conscious, but it can at any moment be converted into an intuitive aspect or can be sensed in a kind of reflective activity, just as the distinction as such of a figure from its background, and not simply the colors of figure and background, can be actualized.

28The phenomenon of opposition can but serve to clarify the difference between a purely logical and a phenomenological analysis. Logical analysis restricts itself to what is given objectively and so defines opposition as exclusion. The same phoneme cannot simultaneously be light and dark. “In speaking of phonemes opposition implies incompatibility at a certain point” (Martinet 1960, 4.8). But a phenomenological analysis examines facts not only as they really are, but also as they appear in our mind. In such an analysis, which includes the subjective pole, opposition appears both as exclusion and as inclusion or implication. In the mind, too, dark and light exclude each other, but the excluded element, that is the process of exclusion itself, namely to set off dark from light, is itself included in the mind. “Opposition always combines two distinct things, which are however connected in such a way that the mind cannot conceive one without conceiving the other” (Pos 1938, 246) “The one implies the other” (Pos 1939, 76).

The associative aspect of the paradigmatic axis

29All three main proponents of the two-fold system, Kruszewski, Saussure, and Jakobson, have connected the paradigmatic axis with the association based on similarity. Kruszewski and Jakobson have done so in explicit contrast to the syntagmatic axis, to which they pair the association based on contiguity. With the figure of speech, metonymy, we have however granted this second form of association a place on the paradigmatic axis and have moreover introduced the third of the three classical principles of association, the principle of contrast.

30The association of similarity has an indisputable, but not exclusive right to a place on the paradigmatic axis. The typological pairing of paradigmatic relation and the association of similarity can thus be maintained aside from this one restriction. Similarity does in fact play a specific role in selection, which clearly distinguishes it from the other two forms of association.

31In our key sentence (1) we can replace the noun car with any number of other nouns. The construction remains grammatically meaningful as long as the unit chosen belongs to the same category and occurs in the same syntactic form:

(11) My room is fast.

32as opposed to:

(12) *My at is fast.
(13) *My rooms is fast.

33The construction remains semantically meaningful as long as the unit chosen belongs to the same subcategory, in our case to the subcategory of objects with the semantic features “concrete” and “mobile”:

(14) My breath is fast.

34Finally the entire construction is identical in meaning when the unit substituted is a synonym:

(15) My automobile is fast.

35Basically, the similarity of expressions, which is based on belonging to the same syntactic-semantic category and subcategory, is prerequisite to the operation of selection.

36But as mentioned, we occasionally speak of our engine instead of our car and of our palace instead of our cottage. Here we have, in addition to a fundamental selection within a basic category (noun) and a syntactic form (nominative, singular), a secondary selection which involves not cate|gorical but rather other formal aspects that are not of a logical or linguistic nature. Relations of association and the laws of Gestalt (whole-part, content-container) are essential to this secondary selection. Although they are not exclusively linguistic (syntactic and semantic) in nature, they are nevertheless of the greatest but also most confusing linguistic relevance. They do not follow the specifically logical and linguistic laws of compa|tibility, nor do they fit into the current systems of syntactic-semantic categorization and subcategorization. According to these laws only a noun with the semantic feature “liquid” could follow the verb to drink. A noun with the feature “hard” is incompatible with this verb.

(16) I’ll drink wood.

37as opposed to

(17) I’ll drink wine.

38is not a well formed sentence. There are things however that are not only hard, but hard as glass; yet we can use the verb to drink with them to form a meaningful and perfectly understandable utterance:

(18) I’ll drink a glass.

39In order to understand sentence (18) we must grasp the associative relation of spatial contiguity, or — to express it more precisely and concretely according to the critique of the classical theory of association by Gestalt psychology — the relation of the laws of Gestalt between wine and glass, namely contents-container.2

40Not only metonymy but also antonymy and the metaphor disregard the logical compatibility or the compossibility of subcategorical features. In the sentence

(19) My automobile uses too much gas.

41the choice of the noun rests on the semantic feature “equipped with a combustion engine”, which makes the utterance meaningful. Not so the metaphorical choice of rattletrap in the same context.

42The laws of categorically ordered compossibility of linguistic units on the syntagmatic axis mingle with associative and Gestalt-theoretical rules of compossibility on the paradigmatic axis.3

43In contrast to the adherents of a purely rationalistic philosophy, for whom the relationships among the objects of experience and thought are ordered exclusively by logical categories (the categories that are constitutive of thought would then also and exclusively be constitutive of sensory experience and language), Husserl insisted on additional laws of compossibility, which are not logical in nature, but nevertheless just as formal and a priori as the laws of logical categories. He includes among these the laws of time and especially the classical laws of association. Husserl intended these non logical-categorical laws of compossibility primarily for prepredicative if not even for preobjective (vorgegenständlich) experience as well as for reproduction, which is indeed largely associatively motivated (Husserl 1950, §§ 37ff.; Holenstein 1972a 198, 222). The examples cited lead to the insight that they are also effective in the original production of language.

44The associative ordering of selection is also of the greatest psychological significance. Those forms of transformation which dominate presentday linguistic discussion presuppose a “rational” mastery of the logical and grammatical categories involved, e.g., in the transformation of the subject in the following sentences:

(21) His ruddy face betrays his farming background.

(22) The ruddiness of his face betrays his farming background.

45No such mastery is required for the associative transformations. They take place involuntarily and unconsciously. One cannot only set up artificial systems of signs whose transformational rules coincide with the classical principles of association (x -> y, if x is connected by a relation of similarity, contiguity, or contrast with y); there also seems to be a natural system of signs in the human unconscious which can be reduced, to a great extent if not entirely, to such transformational rules. Perhaps greater justice can be done to the unconscious if it is not defined as language, as Lacan does, but is rather approached as a sui generis system of signs that is separate from language.

46An analysis of the paradigmatic axis in terms of associative theory leads to a further conclusion, that contrast, the third principle of association, is not merely secondarily and regionally involved in the process of selection, namely with respect to antonymic expressions. There is no such thing as choice unless one thing can be distinguished from another, i.e., unless the two are in some way contrasted. Contrast is, in this respect, no longer the motive for (associative) selection, but rather a prerequisite, a condition that makes selection possible. A glance at the second basic operation of language reveals that combination, too, requires contrast.

The associative aspect of the syntagmatic axis

47Looking through the literature on this point one is initially struck by Saussure’s silence. In contrast to Kruszewski before him and Jakobson after him, Saussure relates syntagmatic combination neither to association in general nor to contiguity in particular. Following Kruszewski’s system of the association of language, the fact that Saussure does not associatively determine the syntagmatic axis gives the impression of a deliberate, critical omission.4 We suspect that Saussure’s omission rests on the consideration that contiguity alone, a mere string of linguistic units, is not enough to produce a syntagm or a meaningful combination.

(23) *to of fast he.

48A mere relation of contiguity does not suffice to invest a sequence of words with a syntagmatic character. To this end particular and restricted syntactic and semantic relations between the individual word units are indispensable.5

49For a clear understanding of an interpretation of syntagmatic combination via the theory of association, the possible variants of this theory must be kept distinct, the physiological, the phenomenological, and the formalistic or descriptive in the narrowest sense of the word.

50The physiological theory accounts for the connection between mental pictures and ideas through correlative processes in the central nervous system. Let us suppose that two simultaneous stimuli A and B elicit a coupled response in the brain. When stimulus A occurs again later in isolation, it elicits both response A' and B' in the brain, so that the corresponding presentations A" and B" also appear together on the scene of the mind.

51The phenomenological and descriptive theories do not involve considerations regarding the physiology of the brain. In fact Hume, the father of the modern theory of association, made a clear distinction between the phenomenon of association and the additional interpretation just described (1739, 60), which may possibly turn out to be a chimerical construction. Hume states that the phenomenon of association itself consists of a connection between two events, which “we feel in our mind” (1739, 75). According to current phenomenological analysis, one cannot only sense as a phenomenon of the mind two events as conjoined, i.e., as belonging together and forming a unit, but also the motivation for the forming of this unit. The motivation lies in the relations of similarity, contiguity, or contrast, which exist between the events as far as they are sensed. These relations need not exist objectively or “in reality”. Not everything that is objectively similar need necessarily be subjectively similar, in perception for example — and vice versa. Subjective appearance alone is definitive for association. Similarity, contiguity, and contrast, are experienced as the motive for moving from one event to another and for the unification of both.

52The purely formalistic or descriptive theory of association, as W. Dilthey probably had in mind, avoids all explanative possibilities, causal as well as motivational, autonomous as well as heteronomous,6 and disregards the character of process and the psychic manner of givenness (e.g., that which can be felt). It restricts itself to a formal analysis and classification of relations. It is part of a pure theory of relation. This restriction stems from the dubious conviction that, as Dilthey says (1924, 184), “we do not directly perceive the elementary processes as a process within us or as the fulfillment of a function within us, but rather we are conscious of the result only”.

53The pairing of syntagmatic combination and contiguity is tenable only in formal terms, whereas selection is to be paired with similarity from the viewpoint of a phenomenological theory of association as well, which beyond formal description looks for a motivational explication. Contiguity is here the result only of a syntactically-semantically motivated process; it is not itself the driving force behind this process of combination, as similarity (of category and syntactic form) is for selection. The relations of selection/association of similarity and of combination/association of contiguity are in motivational terms chiasmatic. Strictly speaking only the first part of Jakobson’s statement “Similarity relations underlie the selective operation, whereas combination is based on contiguity” (1971b, 296) is tenable. The correct formulation would be paradoxical: “In the twofold, bi-polar constitution of language selection is based on similarity, whereas contiguity is based on combination.”

54If combination were associatively and not syntactically-semantically motivated, then a child would learn externally similar forms of combination simultaneously and an aphasic would lose his knowledge of them simultaneously. It has been shown that this is not the case (Jakobson 1971b, 295). An English-speaking child learns the plural of the noun dream, then the genitive singular, and lastly the ending for the third person singular of the verb, in that order, although all three are formed by adding an /s/ to the same root. The aphasic loses his knowledge of these forms in reverse: first the third person singular of the verb, then the genitive singular, and in an advanced stage the use of the phonetically identical plural.

55It was found that two motivations can play a role in selection, the one primary and fundamental, the other secondary and founded, and that not only similarity but also the relations of contiguity and contrast may intervene, namely in the case of figurative speech. Secondary motivation can be encountered in combination as well. There are two cases of this.

56Once a combination is constituted with the result that linguistic units enter into a relation of contiguity, then this relation of contiguity here too behaves in the usual way: the appearance of a part of such an activated combination or of a combination that has become familiar through repetition also evokes the remaining parts. The missing part of a sentence can be reproduced by association without regard to sense and grammar. Not combination, but recombination is based on contiguity.

57The other two principles of association can participate secondarily in the activity of combination as well, however not in temporal but rather in hierarchical terms. Poetic texts differ from prose in that their linguistic units are not only correctly combined syntactically speaking, but are also composed according to the laws of similarity and contrast.

Tabák da bánja,
kabák da bába —
odná zabáva.

Tobacco and bath
bar and woman —
the only pleasure.

58In this Russian proverb, analyzed by Jakobson in one of his studies in poetry (1970), the words are chosen and combined so that three of the lines have an even rhythm. The four kinds of pleasure are so arranged that the two direct names (tobacco and woman) and the two local, metonymical names (bath and bar) are chiasmatically opposed. Similarity and contrast in poetry also determine combination. For the sake of information alone it would suffice to list at random the four pleasures, ignoring both semantic contrast and the phonetic and rhythmic equivalences of their names.

3 | Extension of the theory of the two axes of language to a theory of four operations

Differentiation

59A contrast is prerequisite to both selection and combination. These two basic operations of language are accompanied and preceded by an additional operation, the differentiation of signs.

60The process of differentiation can best be illustrated on the phono|logical level. The phonological system is built, according to Jakobson’s thesis, on the maximal contrasts of the distinctive qualities that constitute phonemes. One of the first linguistic sounds, i.e., sounds that bear meaning, that a child acquires is papa with the maximally contrastive phonemes /p/ and /a/. /a/ affords the maximum opening; /p/, a labial stop, completely obstructs the oral cavity. The first consonantal variant besides the oral-labial stop occurs as a nasal or dental stop (/m/ and /t/ in mama and tata}. /m/ and /p/ are nasal and non-nasal respectively; /t/ and /p/ are distinguished by the contrastive characteristics light and dark. Vowels follow the same pattern; the wide open /a/ is maximally contrasted with the narrowest possible sound such as /i/ (papa — pipi). Phonemes can thus be further and further differentiated, always in terms of maximal contrast.

61The perception of phonological units leads to a differentiation of the “figure-background” structure, which both Gestalt psychology and phenomenology have shown to be a condition for the possibility of perception. In order to be able to perceive something at all, we must be able to distinguish it from a contrasting background. That which is thus distinguished, the “figure”, proves to be more differentiated, stronger, more solid, more “thing-like” than the background. In comparison with the figure, the background appears pale, undifferentiated, blurred (Rubin 1921).

62The perception of phonemes leads to a distinction between para|digmatic alternatives and syntagmatic contexts, both of which in their own way function as background. Both however contain elements which can make just as differentiated an appearance as the “figure”. “Did you say ‘pig’ or ‘fig’?” the cat asks in Alice in Wonderland. A phoneme is useless in isolation. Its only value consists in its being different, in its being distinct from all other phonemes, especially from those which could replace it and consequently change the meaning of the sequence of phonemes in question. Two distinct sounds that do not effect a change of meaning in a particular language can be kept apart only with great difficulty by speakers of that language (e.g., [i] and [ü] for a Czech).

63Moreover in the perception of signs an object can be made to stand out still more when its frame takes the strongest and most compact form. The figure is primarily distinct from and in sharp contrast to this frame and only secondarily together with the frame distinct from the general, diffuse background. A picture with strong contours stands out more than one with no contours. We are most likely to perceive a traffic signal when it is set off by a differently colored frame. A phoneme /p/ to which an /a/ is added is easier to distinguish from other possible noises than a solitary /p/. All these frames themselves can, but need not necessarily, be endowed with positive meaning. They serve primarily to set off by means of contrast. Their function is to clarify. A language contains not only elements that mean something (words, etc.) or that differentiate meaning (phonemes), but also elements that clarify what other elements mean. A context can serve as mere contour.

64The appearance of oppositional qualities in the perception of phonemes can be understood functionally as indicating otherness. We have seen that, according to Pos and Jakobson, an advantage of the function of opposition is the fact that one member implies its opposite, that it refers to its counterpart. The relation of opposition is not alone, however, in consisting of a reference to otherness. Among the idiosyncratic qualities of relation in perception, which have been treated by phenomenologists like Stumpf and Husserl, we find the “increasing sequence” (Steigerungsreihe), a noteworthy phenomenon with two aspects. Firstly increasing sequences, e.g., a spectrum of gradually darkening colors or a series of bigger and bigger triangles, form uniform figures of remarkable unity in the field of perception. The increase apparently has a unifying effect. Secondly every member of such a sequence refers to its modifications in both directions, a middle-sized triangle to its larger and smaller realizations. This idio|syncracy, which is particularly evident in increasing sequences, proves upon further analysis to possess a validity that applies to every act of perception. Every act of perception appears to be virtual as a member of an increasing sequence. We can imagine every picture on television more or less focused. In general every object refers to a more and to a less optimal realization of its qualities. Gradual differentiation, too, is a primary phenomenon like opposition, even though its development may follow later, genetically speaking. The implication of something else is specific to both phenomena. The phonological preference for a binary rather than a trinary relation is not phenomenologically rooted through the structure of the mind, but only functionally, teleonomically motivated through the goal of an efficient coding and decoding.

Signification

65In order to produce our key sentence (1) about the speed of a car, we have to choose each word from among several alternatives (car — carriage — automobile — engine, etc.), which can be substituted for each other. But at the same time we substitute the different signs for that which is being designated.

Commutation and substitution

66Significative substitution precedes selective substitution. Significative substitution occurs when two objects enter into a relation with each other as signans and signatum or — seen from the other side of the coin — become separated from each other. Selective substitution consists of the mutual exchange of two or more signantia in terms of their function. It is somewhat disconcerting that this function consists precisely in significative substitution. In the first form of substitution one thing replaces another, which it represents. In the second form one thing replaces another in relation to a third, which is represented by both in the same or similar manner. The other as such is not intended, but rather the third, in which they both meet.

67In the one case we have a meaning-relation. The substitutor signifies the substituted. In the other case we have a relation of equivalence. The substitutor can replace the substituted without having the combination of which they are a part sacrifice its meaning or the form of meaning.

68It seems appropriate to keep both forms of substitution conceptually distinct as well. Indeed linguistics has long had a specific concept for the second form of substitution: commutation (Jakobson, Fant, & Halle 1952, 1). The commutation test serves to determine the equivalence of linguistic data. It reveals, for example, that in English /b/ and /p/ can distinguish meaning and are thus two distinct phonemes, whereas in Finnish they do not and therefore function only as two variants of the same phoneme.

69The prerequisite for significative selection is again difference, an inequality between signans and signatum. The operation of differentiation is the fundamental and universal condition underlying the three other operations constitutive of language: significative substitution, paradigmatic selection, and syntagmatic combination.

The associative aspect of signification

70The connection between signification and association is a familiar theme. Saussure (1967, 2119; 2024f.) speaks of a relation between signifier and signified as association interne or association primordiale, to which he opposes paradigmatic relationships as association externe. Husserl’s Logische Untersuchungen (1913, 29) include a “Digression on the Associative Origin of Indication”. Some preliminaries are necessary in order to understand this argumentation.

71In the widest sense of the word any mental connection is called association. A narrower sense of the word refers to the nature of conscious experience out of which a connection arises. The narrowest and most precise use of the word refers to the “objective” data of consciousness which appear to motivate the connection. Noetically, that is in terms of mental experiences, those connections are called associative that originate not in wilful, rational acts such as position and apperception but are rather passive and unconscious. In noematic terms, however, that is in terms of the objects of the mind, a connection is called associative when it is not motivated by the sense of the connected elements, but rather by the formal structures and relationships of these elements, such as similarity, contrast, and contiguity. Table and chairs in a living room form a unit in relation to the other furniture not only because they “touch” each other, but also and above all because, from a meaning point of view, they functionally refer to one another. The unity of table and chairs is both associatively and sensefully motivated. An originally senseful connection often becomes in time a merely associative one. The sense-connection between a country and its national colors can be forgotten. What remains is the contiguity of both, which in turn becomes the connecting factor. We shall here use the concept of association only in its narrowest, noematic or objective sense for connections that are based exclusively or primarily on the relations of similarity, contiguity, or contrast.

72The phenomenon of association can itself be divided into various phases or aspects and consequences: intention, combination, repressive and signi|ficative substitution. In the first phase something refers intentionally to something else that lies more or less in the background, on the consciousness-horizon of perception, memory, expectation, or imagination and that is “awakened”, actualized by this reference. A tendency arises to turn toward or become aware of the intended. The direct consequence of this is a sensed connection between the awakening and the awakened datum. What was previously simply adjacent is now experienced as belonging together. The one forms a phenomenal unity with the other. Association has been designated as such from this second phase. Following the forming of a unity, but also instead of an actual forming of a unity, it can happen that the associating datum is relieved by the associated one. The one represses the other. If this repressive substitution is a chain reaction, then it is called a flight of ideas. As a further possible consequence the associated data can occur as signs for each other. The one represents the other.

73What interests us most here is the difference between the first phase of association, the merely intentional reference and awakening, and the pos|sible consequence, mentioned last, of a sign-reference. The latter cannot simply be reduced to the former. Not every associative reference can function eo ipso as a sign. Each member of an association does not “automatically” become the sign of its partner.

74We spontaneously interpret a curved line on a road sign as an indication of a bend in the road. In the same way everyone considers smoke as a sign for fire. But no one assumes that a black color sample on a package is a sign that it contains white material. Similar things refer to each other and spontaneously act as signs for each other. Contiguous things refer to each other and spontaneously act as signs for each other. Contrasting events also refer to each other, but never spontaneously become signs for each other.

75However, there do seem to be exceptions. (1) Secret languages, not only artificial ones but also natural ones, such as the dream, often depend on oppositional relations. Yet the choice of the opposite is motivated not by indication but rather by concealment, by deviation from the message. (2) In figurative speech we use antonyms. However we can speak of our palace instead of our cottage only because the house-quality, which is common to both, permits a spontaneous comprehension of the use of the one expression for the other. If the common base assumed in every opposition is not realized by the recipient of the message, if he does not already know or suspect that our house is not our palace, then the antonymy will miss its mark. The expression will be misunderstood.

76One could be tempted to deny contrast its own associative force or to allow it only very little strength. In fact one occasionally encounters the tendency to reduce the association of contrast to that of similarity and to consider the association as carried by the common core of that which is contrasted (cf. Jakobson 1971b, 244). If similarity alone were the binding factor, then a greater associative effect would have to be produced by what is similar in every respect than by what is completely opposed in one respect. The association of “almost synonymous” expressions (big — large) would have to occur more frequently than that of antonyms (big — small). This is factually and understandably not the case. Statistical studies have shown that word associations are most frequently determined by explicit polar qualities (Clark 1970, 275ff). Contrast clearly does not have less associative strength than similarity and contiguity. The frequency of the association of contrast makes sense if we keep in mind what has already been said: that the essence of oppositional qualities lies in not being able to think one without the other.

77This remarkable but unexplored fact that precisely the strongest association does not serve eo ipso as basis for a sign-relation could itself prevent a rash equivalence of associative and significative reference. There is a descriptive difference between association and signification. An asso|ciated datum is only appresented or copresented; a designated event is represented. The aliquid stat cum aliquo of association becomes an aliquid stat pro aliquo in signification.

78The three principles of similarity, contiguity, and contrast are related to the basic operations of language in different ways depending upon their performance. Similarity and contiguity establish a positive relationship between two data, which serves as the basis for the signification of the one by the other. Contrast is primarily a principle of dissociation and only secondarily a principle of association as well, due to the essential fact that the very distinction between two events also serves to relate them to each other. The differentiation of signantia in language is based on contrast. It is thanks to the associative aspect that differentiated events are not randomly adjacent, but interrelated by being felt. As we shall show in the following, an intuitively grasped signification also makes indirect use of contrast on the basis of this characteristic of reference.

Contiguity as the motive for the choice of signs

79Peirce first introduced a systematic analysis of signs in terms of their origin in the relations of similarity and contiguity. In phenomenological analysis contiguity seems particularly problematic as a principle of origin. According to Gestalt psychology only Gestalt-like wholes can function connectively and can together become a new, definite Gestalt. A part can become a sign for the whole only when it is perceived as a “natural” part that is itself distinctly set off from the whole. A single note, which alone does not make itself heard in a melody, is unable to call this melody to mind; a chord, on the other hand, heard in the music as an independent passage, can do so. Phenomenological psychologists like Merleau-Ponty and Linschoten have gone even further by questioning not only the connective effect of the principles of association but also of the principles of Gestalt. According to them a part refers to another part or to the whole only insofar as a unity of sense or function is formed.

80The question whether mere contiguity suffices for association and consequently underlies a sign-relation is not easy to decide, because there are hardly any pure relations of contiguity in concrete perception which are not also molded by relations of sense and Gestalt. Unlike the empty space of mathematics, not all moments in the space and time of perception are isotropic. There are content-determined “focal points” in concrete space and in experienced time, which exercise a centering and localizing effect. Among existing objects there are always those that overwhelm and dominate others due to their meaning or Gestalt (Holenstein 1972b). A village and a mountain do not lie next to each other as equals. They are not connected by mere contiguity, but by a centering and localizing relationship. A village lies “below”, at the foot of the mountain. There can be a sense-connection as well. Perhaps the village was erected near the mountain for the sake of economic exploitation. All these relations play a role when the mountain becomes the symbol for the village.

81To guard against an overemphasis of sense-connections, it is worth remembering that in the psychoanalytic sign-system of the unconscious, coincidental minor details that are indifferent for the whole of a situation can frequently function as a sign. Sense explanations in a process of secondary rationalization are often superimposed on common relations of association and of signs, which are actually or supposedly based on a relation of mere contiguity (cf. folk etymologies). The human mind apparently has great difficulty in accepting “sense-less” connections.

Contrast as the motive for the choice of signs

82It is a condition for a successful choice of sign that the sign chosen has not already been claimed by another object which also requires a name, that the sign contrasts with the signs already in use. b can become the sign for B if b differs from a analogously to the way that B differs from A. This condition alone can sometimes determine the choice of signs, b can become the sign for B BECAUSE b differs from a analogously to the way that B differs from A.

83The traditional explication of the relations between signantia and signata requires further elucidation. (1) The relations among signantia must be taken into account as well. (2) In addition to the principles of association of similarity and contiguity, the principle of contrast, too, must receive attention. In this way we can venture a step forward in resolving the appearance of chance and arbitrariness in the choice of signs.

84Saussure deals only with the negative connection between the contrastive relation of signantia in themselves and their relation to signata. According to him (1931, 163) arbitrariness and differential character imply each other. A modification inherent to the system of signantia can ignore the relationship with signata, since the latter are arbitrary anyway. For Saussure (1931, 180ff.), the arbitrary choice of signs is limited only by relations of similarity which exist among signantia beyond the relations of contrast. Thus he considers nine absolutely arbitrary, unmotivated; but nineteen only relatively so. Nineteen is a term that calls to mind the other terms of which it is composed and together with which it forms number sequences: nine, ten; seventeen, eighteen; twenty-nine, thirty-nine, etc.

85According to our thesis the relation between b and B only appears to be arbitrary, b and B are not directly connected in any way that could motivate a sign-relation. But they are indirectly connected with each other through their analogous relationship with a and A.

86To illustrate the connective quality not only of similarity relations but also of contrastive relations among signantia, we shall treat a special case where both come together, namely in a particular kind of diagram. In Peirce’s semiotics, “diagrams” designate a subclass within the class of “icons”, in which similarity refers not to the qualities of the referent as with the subclass of “imagines”, but to the relations among its parts. In a relation of difference, for example, it is naturally the difference that motivates the choice of the individual signs.

87If we have lost a pawn from a set of chessmen, then we will choose from among the knickknacks at our disposal the piece that most resembles the missing pawn in size and if possible in form and color as well. Should we later also lose a bishop, then the size of the second substitute in comparison to the size of the bishop is no longer the most important factor — perhaps the substitute that we selected for the pawn is about as big as the bishop for lack of a better substitute. The only thing that counts now is the difference in size from the substitute pawn. Just as the choice of the term nineteen is motivated by the prior choice of the terms for the ninth and tenth units of the decimal system, to which the number to be named is related, so too the choice of a representative for the bishop is motivated by the choice of a representative for the pawn, which is related to the bishop through difference. The difference in size between these two substitutes corresponds to the difference in value of the chess figures.

88As long as the differentiation of signantia and signata is of the same nature (e.g., gradual), we can reduce it to an intuitive relation of similarity (similarity of contrast). Not so when the two systems of signantia and signata are differentiated in their own way and suited to their own particular nature. The intuitivity of similarity is then lost and with it the possibility of association on the basis of similarity. Both systems then are still mathematically but no longer intuitively isomorphic (every instance of the one system corresponds to an instance of the other). Only the differentiation of each system within itself remains intuitively given.

89Historical number systems illustrate such diverging differentiations. A Cretan decimal system from 1200 B. C. uses simple vertical strokes for the units from one to ten, familiar to us from Roman numeration: |, ||, |||, ||||, etc. “Ten”, the second basic unit, which differs from the number “one” in two respects, shows a graphically optimal antithesis to the symbol for “one”; it is a simple horizontal stroke —. A circle represents the hundreds and a rhombus the thousands (Smith & LeVeque 1971, 756f.). The differentiation of the numerical units according to mathematical principles is expressed on the side of signantia by a differentiation according to configurative opposites.

90It seems appropriate to coin a special term for those signs which arise indirectly through a specific difference to another sign. These signs live as parasites on other signs, to which they are paratactically related: we can call them parasemes.

91According to Peirce (1932, 2.306) all signs have a multiple character. There is no absolutely pure index, nor do there exist signs that are completely and utterly without index-quality. The designations, icon, index, symbol, indicate not an exclusive but the dominant quality of a sign. This also applies to the newly proposed designation, paraseme. All signs possess a parasemic aspect as members of a system of signs. Difference is co-constitutive for each and every sign, be it primarily an icon, an index, or for the moment a seemingly arbitrary symbol.

92Since the signs ˅ (or ˄) and X, familiar to us from Roman numeration, designate a different basic unit in other number systems (˄ is the sign for “ten” in an Egyptian hieratic number system ca. 3400 B. C.; X indicates “four” among Chinese merchants), we may assume that these signs are primarily of parasemic rather than iconic origin (˅ [or ˄] as symbol for the hand, X as double ˅). Their existence is justified by their patent difference from the basic unit |, which is first multiplied three, four, or nine times until it is replaced by a sign with a contrasting form for the sake of brevity and to mark a further differentiating unit. The sign for fifty (L), which is usually described as purely arbitrary or at most as an arbitrary adaptation of the chi-sign (X, T, etc.), also possesses a dominant parasemic character. It fits into the Roman numeral system through its simple and striking contrastive relation to the older signs, |, ˅, and X. C (a hundred) and M (a thousand) are clearly exogenous signs borrowed from another system of signs, the alphabet, i.e. the lexicon (Centum and Mille). Borrowing, a common origin of signs, is encouraged by both similar and contrastive traits of the borrowed sign with the endogenous signs which are developed by means that conform to the system.

93In our Indo-Arabic number system a clear example demonstrates how a change in one sign entails a change in another in order to maintain a contrastive distinction. The system has two variants each for the numbers “one” (I, 1) and “seven” (7, 7). The first sign for “seven” is so close to the second sign for “one” that they can be confused. To make the distinction between the two unmistakable an additional stroke is added to the sign for “seven”. An analogous distinction has arisen in German for the word for “two”, zwei. To make it unmistakably distinct from the word for “three”, drei, the pronunciation has in many places been changed to zwo. Both cases illustrate not an absolutely arbitrary, but rather a contrast-oriented modification of the sign with means that conform to the respective system.

94The contrastive extension of an iconically based sign is common in mathematical logic: ˄ (conjunction), v (alternative). A special case is illustrated by the seals for the universal operator (Alloperator) ∀ and the existential operator (Existentialoperator) ∃. The inversion of the first letter of these terms indicates their specific use.

95Language, too, is to a great extent a parasemic sign-system, a system that is held together by immanent contrasts and that is connected to the universe of the referent via iconical and indexical bridges, while the referent in turn is differentiated according to its own structural laws. We discussed the contrast-oriented structure of the phonological system in 3.1. The contrasting process begins with sounds whose sign-characteristics are built up on clear, iconical or indexical aspects (mama as onomatopoetic production of the infant’s sucking activity, papa as a nonaffective designative expression — cf. Jakobson 1971a, 524f.).

96The thesis of the contrastive origin of signs, especially of linguistic signs, will of course remain speculative, based on relatively few and seemingly arbitrarily collected instances, as long as the various possibilities of contrast have not been schematized. Differentiations can come about not only by splitting up characteristics analytically but also through synthetic combinations (lane-pane-plane). Extensive systems of signs reveal not only differentiations of individual elements according to a unified code, but also contextually determined division of the code itself into subcodes: characteristics redundant in one context may play a distinctive role in another. The originally contrastive origin of a sign can become concealed by semantic shifts through time. The indirect relation between signans and signatum, originally borne by a relation of contrast, is later carried by a direct relation of contiguity, inaugurated however by the indirect relation of contrast.

Summary

97An examination of the major axes of language within the framework of the theory of association, including the third principle of association, contrast, led to an extension of the theory of two axes to a theory of four basic operations: differentiation and signification as primary operations, and selection and combination as secondary operations founded on the two primary basic operations (Fig. 1). This examination also led to a differentiated pairing of the three classical principles of association to these four operations.

4 | Evaluation of the theory of four operations

98The extension of the theory of two axes to a theory of four operations has the advantage of embracing — in a coherent theory oriented on all three classical principles of association — all basic relations that have been dealt with in the field of structural linguistics since Saussure. By involving contrast as a principle of association, the relation of opposition, the most important discovery of this branch of linguistics, can be included in a classical theory of relation.

99The validity of the extension is evident in two fields in particular, aphasia and poetry, which Jakobson schematized and ordered with his pioneering theory of the two axes. The extended theory integrates structural phenomena, which Jakobson himself discovered or analyzed, but which do not fit into the theory of the two axes.

Aphasia

100In Jakobson’s studies on aphasia two stages may be distinguished. In the first stage (1971a, 367ff.) he demonstrates that aphasia develops in reverse order to the child’s individual acquisition, based on maximum contrast, of a system of sounds as well as to the development of sound patterns in the various languages. Aphasia first undermines the most recently acquired sound differentiations and progresses to the older and increasingly simpler ones. In the second stage (1971b, 244ff.) Jakobson isolates and orders disparate aphasic phenomena in view of the two operations of selection and combination. He distinguishes two kinds of aphasia, similarity disorder and contiguity disorder. The former is tangent to the paradigmatic axis of language, which is based primarily on a relation of similarity; the latter touches the syntagmatic axis, which implies a relation of contiguity.

101All three types of aphasia are embraced by the theory of the three principles of association and the four operations of language: contrast disorder or disintegration touches upon the operation of differentiation just as similarity and contiguity disorders interfere with the operations of selection and combination. Furthermore we have gained insight into why an aphasic suffering from the two latter disorders is still capable of performing the basic operation of signification. The significative relation is fundamentally based on the relations of both similarity and contiguity. If the one form of association is lost, the other can always take its place. One of Goldstein’s female patients, when asked to list a few names of animals, disposed them metonymically in the same order in which she had seen them at the zoo; also, despite instructions to classify certain objects according to color, size, and shape, she arranged them on the basis of relations of contiguity as home things, office materials, etc. (Jakobson 1971b, 249). One of Head’s patients failed to recall the name for black. He first used the circumlocution What you do for the dead and later simply shortened this to dead (1971b, 250).

Poetry

102According to Jakobson’s analyses, stretching over a period of fifty years, it is the “orientation toward expression” (1972, 30f.), “the set (Einstellung) toward the message as such" (1960, 356) that distinguishes poetic language from other modes of speech. The immanent structural laws of the medium take on a “life of their own” in poetry; they become auto|nomous and as such, manifest and perceptible.

103What are these immanent structures? Following the theory of the two axes, Jakobson assigns a prominent position to the two operations of selection and combination, paired with the associations of similarity and contiguity. However, these associative relations do not play the same role in poetry as they do in the other modes of language. Could this be a consequence of the set toward the immanent structural laws? “The poetic function projects the principle of equivalence from the axis of selection into the axis of combination” (1960, 358). A poetic sequence is characterized on all levels, the phonetic as well as the meaning level, by the repetition of the same and similar elements and of their contrastive variations (alliteration, rhyme, rhythm, homonymy, synonymy, antonymy, parallelism, etc.).

104Jakobson (1965, 10f.) assigns contrastive opposition only a secondary and subordinate role in poetics and does not consider it very illuminating, at least with reference to one of its possible effects, alienation. We shall try to show in the following how the principle of contrast makes itself felt in poetic language as a principle of form that is equally as important and powerful as similarity and contiguity are. In our argument we intentionally confine ourselves to Jakobson’s studies on poetry. Our thesis: In addition to the projection of the principle of equivalence from the axis of selection into the axis of combination, there is also in poetry a projection of the principle of contrast from its usual level of a latent prerequisite for the significative, selective, and combinatorial operations into a level of a patent, “palpable” and “perceptible” form.

105On the paradigmatic axis every choice of an element by the poet and correspondingly every apperception on the part of the listener or reader takes place against a three-fold background: “(1) the existing poetic tradition, (2) the practical language of the present, and (3) the poetic tendency underlying the expression in question” (Jakobson 1972, 18ff.). “The unknown can be recognized and effective only against the background of the known” (1972, 62f.). The Russian formalists designate this particular contrast on the paradigmatic axis either as “difference-quality”, Broder Christiansen’s descriptive concept suggestive of Gestalt theory (1909), or as “alienation”, a functional and now famous expression. Speaking of Pasternak’s iambic tetrameter Zemlja “Earth”, Jakobson writes:

Since the overwhelming majority of downbeats occur with word stresses, the listener or reader of Russian verses is prepared with a high degree of probability to meet a word stress in any even syllable of iambic lines, but at the very beginning of Pasternak’s quatrain the fourth and, one foot further, the sixth syllable, both in the first and in the following line, present him with a frustrated expectation. ... Quite naturally it was Edgar Allan Poe, the poet and theoretician of defeated anticipation, who metrically and psychologically appraised the human sense of gratification for the unexpected arising from expectedness, both of them unthinkable without the opposite, “as evil cannot exist without good” (Jakobson 1960, 362f.).

106Earlier (1935), also writing about Pasternak, Jakobson says that his metonymical style can really be fully appreciated only in confrontation with Majakovskij’s metaphorical style, just as the emancipation of a sign from its object, characteristic of their generation of writers, can be understood only as an intended antithesis to the foregoing nineteenth-century naturalism, which made a fetish of objects.

107Marked relations of contrast on the prosodic level join the constitutive relations of contrast of the distinctive phonological elements on the syntagmatic axis through the means of accentual (sequence of regularly stressed and unstressed syllables), “chronemic” (long and short syllables), and “tonemic” (modulated and unmodulated syllables) versification (Jakobson 1960, 360). The best-known forms of contrastive combination on the sense level are “negative” and “reverse” parallelism. In the former a metaphorical expression is negated in favor of a real one (in speaking of a head: not a keg of beer, but a head), in the latter a real expression is negated in favor of a metaphorical one (1972, 44f.).

108For the axis of signification Jakobson himself expressly states that poetry “is a province where the internal nexus between sound and meaning changes from latent into patent and manifests itself most palpably and intensely” (1972, 373). Signs cannot enter into a relation unless something is there that is different from the referent with which it can be exchanged (Saussure 1931, 159). As soon as something “dissimilar” is given, it can enter into a positive sign relation through a partial similarity or a real or imputed contiguity with the referent. Poetry thrives on this dialectical relationship between dissimilarity and similarity. “In poetry, any conspicuous similarity in sound is evaluated in respect to similarity and/or dissimilarity in meaning” (Jakobson 1960, 372). This holds for the gulf between phonetic and semantic form as well as between phonetic and syntactic form. The disconcerting distribution of grave and acute vowels in the French jour “day” and nuit “night” is evaded either through the choice of substitute expressions or through the choice of epithets which phonologically and semantically equilibrate the collision between sound and meaning. In the same article (1960, 367) Jakobson quotes two lines from “The Handsome Heart” by Hopkins to illustrate the conscious discrepancy between phonetic and syntactic form:

(25) “But tell me, child, your choice; what shall I buy
You ?” — “Father, what you buy me I like best.”

109The recitation of these pentameters may be strictly metrical with a manifest pause between “buy” and “you” and a suppressed pause after the pronoun. Or, on the contrary, there may be displayed a prose-oriented manner without any separation of the words “buy you” and with a marked pausal intonation at the end of the question. None of these ways of recitation may, however, hide the intentional discrepancy between the metrical and syntactic division.

110The hermeneutic confinement to sense, to the sens vécu, seems to neglect the fact that the form of poetic texts is also sensed and that the poeticalness of a text lies in the very interplay of joining and opposing, of sens vécu and forme vécue, or more precisely, the form of sensory elements and the form of sense elements. The units of sense that occur in language are just as far from being atomistic and amorphous data as sounds are. Both levels of language obey specific laws of form, laws of compatibility and incompatibility, which poetry makes conscious, in a unique way.

111The only form to which hermeneutics, following phenomenology and existential philosophy, has done justice is that of temporality and to a lesser extent that of spatiality. Hermeneutics is above all blind to those forms which are prominent in the theory of association. This blindness is even more striking, since for years a number of scholars of hermeneutics have been concerned with psychoanalysis, where these forms play an indisputable role (Holenstein 1972a, 327ff.).7 With the opinion that the laws of association are constructions that are intellectualistically superimposed upon phenomena, these theoreticians approach their greatest opponent, nominalistic conceptualism. According to the latter as well, such laws of form are not principles inherent to phenomena, but are rather constructs for their scientific classification. According to the analyses of Husserl’s phenomenology, however, the principles of association are transcendental-phenomenological principles, i.e. principles that are just as immanent to as they are constitutive for the objects of consciousness. As universal principles of consciousness they make themselves felt in the various outgrowths of consciousness, in perception, in thought, in language, etc., in accordance with the nature of their respective objects.

Notes

• 1 1971b, 435: “The theory of the two linguistic axes, inspired by the classification of associations by the English psychologists and their radical proponent, Troickij, was raised from a mechanistic to a phenomenological level in Kruszewski’s work and grew into a harmonious, unified and exceedingly fruitful theory of language.”
• 2 Sentence (18) is still more complex than shown above. It is equivocal in that it can indicate not only what but also how much I am drinking. Two kinds of nouns can follow the verb to drink in a logical language: either a noun with the features “concrete” and “liquid” or a noun that designates a unit of measurement, in other words with the feature “abstract”. This distinction does not affect our argument. The proponent of a strictly logical language would find a glass out of place in either case. In colloquial speech however both possibilities occur as metonymies. In the one case a glass as a container represents the contents (wine); in the other as a measuring instrument it represents the unit of measurement (two deciliters).
• 3 These rules are however subject to certain restrictions. A glass cannot replace wine in all sentences, as in for example (20) He changed the water into wine.
• 4 R. Engler, Bern, kindly drew my attention to two passages in the lecture notes of Saussure’s students, according to which Saussure does seem to ascribe an associative character to the syntagm. 1967: 1995 C: “Cette association [paradigmatique] est tout a fait différente de la premiere [association ? syntagmatique].” 2063: “[Au moment où] le syntagme se produit, le groupe d’association intervient, et ce n’est qu’à cause de lui que le syntagme peut se former.” Neither passage is explicit enough to allow an unequivocal interpretation. The second could be viewed in the sense of the Cours (1931, 179), i.e., that the syntagmatic string consists of associative groups that are situated along the paradigmatic axis and successively make an appearance: “Notre mémoire tient en réserve tous les types de syntagmes plus ou moins complexes, de quelque espèce ou étendue qu’ils puissent être, ct au moment de les employer nous faisons intervenir les groups associatifs pour fixer notre choix.” Even if such fragmentary passages were to be interpreted in terms of an associative determination of syntagmatic relation, an explanation is still required for the unusual fact that Saussure organizes the associative aspect of this axis differently than that of the paradigmatic axis. Why did he neglect to give it a definite place in his balanced system of polar counterparts?
• 5 Saussure (1967, 199 C): “C’est la combinaison de deux ou plus unites également présentes qui suivent les unes les autres. Si elles se suivaient sans offrir aucun rapport entre elles, nous ne les appellerions pas syntagmes, mais plusieurs unités consécutives ayant un rapport entre elles ou avec le tout forment un syntagme.”
• 6 An adherent of a heteronomous theory of association is, for example, M. Merleau-Ponty, according to whom relations of similarity and contiguity are based on a precedent sense-relation (Holenstein 1972, 307ff.).
• 7 It is clear that the third principle of association, which has been given prominence in this essay, also plays a decisive role in the work of the unconscious, as manifested in the dream and in faulty acts (cf. Freud 1941, 67; 1942, 316), in view of the unconscious "intention" to repress, to conceal through distortion. Furthermore, it is not without reason that followers of Lacan point out that repression originates with instauration of the relation signans-signatum, for which dissimilarity has proven to be a prerequisite.