After reading the printed German translation, I thought it advisable to attach a copy of the English typescript.- J.V.
Jiří Veltruský
BUHLER1S ORGANON MODEL AND THE SEMIOTICS OF ART
1. The Organon Model: An Interpretation ........ 1
2. Expressive and Conative Meanings ......... 11
3. Metalinguistic Function ............ 18
4. Phatic Function............... 22
5. Esthetic Function .............. 25
6. Art as Reference, Expression and Appeal     ....... 35
Footnotes.................. 47
Jiri Veltrusky
BÜHLER'S ORGANOID MODEL AND THE SEMIOTICS OF ART
The term "organon model" in the title of this article
Is not meant to indicate that the sign will be examined here
as a particular kind of tool; the question whether, or in what
respect,  the sign is a tool,  though very important, will be
left aside. The title alludes to Bühler's conception according
to which the sign is related not only to the denotatum but also
to the originator and the perceiver, and is therefore an
expression and an appeal at the same time as a reference.
Although Bühler developed and refined his organon model from
1920 on,  I shal*. refer only to the formulations he presented in
the 1935 Axiomatik and the 193^ Sprachtheorie;^ these formulations
turned out to be his last,  since the political ordeal of the
period drove him out of Vienna and made him engage in a quite
different activity. The purpose of my article is to study some
2
implications of Bühler s conception for the semiotics of art.
1.  THE ORGANON MODEL:  AN INTERPRETATION
Bühler defined the three relations not as mere aspects
of the situation in which the sign is used - therefore external
to the sign itself - but as semiotic facts.  In the organon
model expression and appeal are not psychological but semiotic
concepts, just as reference is;    each of the three functions
endows whatever fulfils it with the quality of sign (each, of
4
course,  in a different way);    the relations of the sign to the denotatum,  the originator arid the perceiver ore semiotic in nature and its referential,  expressive and appeal, or conative, functions are semiotic functions.
2
Three conclusions logically follow, in my opinion, from this conception.
The first is that in all three of its semiotic functio the sign does basically the same thing, namely bears a meaning, though a different one in each. It is no discovery that the referential meaning of the sign must not be confused with anything outside the sign and that it may as well be altogether fictitious as correspond to some reality.  But it is important that, mutatis mutandis,  the same applies to the sign in its qualities of expression and of appeal.
Whatever the sign indicates or hints at concerning its originator is a meaning - an expressive meaning - conveyed by its own properties or    features.  This expressive meaning may more or less correspond to something real in the originator but that need not be the case. At the same time, a sign through which the originator has really expressed himself need not necessarily signify that of which it is, psychologically, an expression;  it can even convey quite a different expressLvc meaning.
Through a minute analysis of Hölderlin's later poems, especially of "Die Aussicht" which may well have been his last, Jakobson and Lübbe-Grothues have identified extremely interesting correspondences between the language and the poetic techniques of these poems and the language peculiarities of schizophrenics. In addition to their importance for poetics and the theory of art  (they confirm that any fashion of using language can become an artistic procedure),  these findings also highlight the
semiotic nature of the expressive function.  There is naturally
no way of knowing whether or not Hölderlin was schizophrenic
or even mentally ill at all; psychiatrists'   speculations about
the mental state of somebody they have not examined directly
are no diagnosis.  But if they happened to be right,  and "Die
Aussicht" with its drawing on the language of schizophrenia
were therefore a genuine expression of Hölderlin's real mental
state, the most remarkable thing would be that the expressive
meanings of the poem do not point in that direction;  the reader
does not perceive "Die Aussicht" as a product of an insane
person,  even when he is familiar with Jakobson's and Lübbe-
Grothues*  findings.  Jaspers seems to have been aware of the
difference between the expressive meaning produced by a sign
and the possible status of the same sign as a psychological
symptom.  When in his speculations about iStrindberg,  Van Gogh,
Swedenborg and Hölderlin as possible schizophrenics he tried
to treat certain features of Van Gogh's pictures as symptoms
of the artist's hypothetical illness, he took pains to emphasize
that he did not mean these works were "sick" /krank/ and
admitted that the beholder need not see them the way he inter-7
preted them.
^ven when the originator intends to make the sign expressive, he may not succeed;  the sign may convey quite another expressive meaning.  Moreover,  the intended expression need not be in the nature of a real self-expression,  for instance, the artist who seeks to endow his work with a definite expressive meaning may be exploring and testing the structural potentialitie
4
of his medium to convey such a meaning, without; being himself otherwise involved.
oimilarly, the appeal directed by the sign at the perceiver is a meaning produced by various properties and , features of the sign and is perceptible as such even when the sign does not actually affect the addressee's attitude,
behavior or state of mind.  "Noting an injunction and doing
q
accordingly need not be the same thing.      The failure to respond is not necessarily a failure bo understand."^ The persuasive features of a speech may be observed with amusement by the addressee as may the implication of his ignorance about a topic, a devotional object may produce a comic effect and the connoisseur may admire a work of Tantra art without going into meditation.
The conclusion that in its expressive and conative aspects, too,  the sign has (or produces or conveys) mean! important not only because semiotics must carefully distingu1^ between'the meaning wliicli is an integral part of the sign the reality outsine it to which it relates 5  it is also important because it allows the meaning od the sign to be grasped in its whole wealth and complexity, whicli is not possible when the concept of meaning is reduced to referential meaning alone.
An example of this "reductionist" conception can be found in So'rensen's definition and analysis of meaning: ^" "The meaning of the sign 8 is the conditions to be satisfied by something in order that it may correctly be said to be denoted by o.  Or equivalently:  it is the conditions something must
5
satisfy in order to be correctly nomable by 8, or in order that S may correctly be said to apply to it." In his view, the English words father, dad and male parent "have exactly the same meaning". He explains that dad is primarily used by children and differs from father and from male parent by the effects it has on hearers but states that "'effect' elements are not part of the meaning of a sign". The only non-referential meaning this theory recognizes is the expressive meaning of interjections - and it does so on the wrong grounds. S^rensen claims that the English interjection bahl has a meaning because it "denotes" the speaker's attitude of contempt towards whatever has just been said. 8ince interjections usually have no referential meaning, their expressive meaning is misinterpreted as referential.
The second conclusion that can be drawn from Biihler's organon model is that the integral meaning of the sign, the signatum,  is made up of its referential,  expressive and conative meanings. The manner in which the,y combine in specific cases is extremely variable, but all these variations are within the framework of certain general relations deriving from the essence of the three types of meaning.
Referential,  expressive and conative meanings are related to each other by oppositions:
1. There is an opposition between the referential meaning
on the one hand and the expressive and conative meanings on the other because the first pertains to the object    and the second and third to subjects for whom it is an object. In extreme cases
6
the predominance of the referential meaning can altogether suppress  the expressive and conative meanings and, vice versa, the predominance of either one of these can suppress the referential. Two plus two makes four has neither expressive nor conative meaning. A pure interjection, whether expressive like bah!  or conative like hey 1, has no referential meaning. This is true of other categories of signs as well. When Robert Delaunay set out to give the dominant position among the meanings of the picture to the conative meaning which consists in giving the beholder the impression that the colors and shapes are moving (vibrating, rotating, moving in depth, etc.), he soon arrived at suppressing the referential meaning as conveyed by depicted objects.
But the oppositions that relate different types of meanings to eacn other are not simple but dialectical opposition One type of meaning may not only suppress but also enhance the opposite type. The referential meaning may bring out the
or bofcfe
expressive or the conative meaning1and either one of them may
A
bring out the referential. More about this later.
The opposition of the object-related referential meaning and the subject-related expressive and conative meanings must not be confused with the opposition between objective and subjective in Panofsky's scoliotic analysis of the artistic representation of human x)roportions. Panofsky is concerned-with quite a different problem when he identifies the three factoids  that make the proportions ' of the visual (pictorial or sculptural)  image of the human body differ from its objective
7
proportions and calls them "three factors of subjectivity".
They are the organic movement, which introduces into the
artistic composition "the subjective will and the subjective
emotions" of the represented beings; perspective foreshortening,
which reflects the subjective visual experience of the artist;
and the regard for the visual impression of the beholder, which
leads to altering "thai- which is right in favor of that which 15
seems right".  ' The three factors actually correspond to the referential, the expressive and the conative function; their description as subjective within the framework of what Panofsky analyzes in no way contradicts  the opposition between the function pertaining to the object and those pertaining to the subjects.
2. There is also an opposition between the expressive and
the conative meaning. Here, too, the predominance of one can in extreme cases suppress the other. In so far as bahl expresses the utterer's contempt towards what has just been said, it may but need not have a conative meaning;  it can relate, for instance, to something the utterer has just asserted himself. In so far as hey!  calls the addressee's attention to some danger, it may express the utterer's emotion but may also be uttered with total indifference,  for instance by somebody paid for giving warning when appropriate (such as a lifeguard on the beach). At the same time, the expressive meaning may bring out the conative and the conative bring out the expressive.
The two oppositions are not on the same plane,  so that they do not necessarily affect each other.  When damn it all1,
8
an exclamation which is not entirely devoid of referential meaning,  is substituted for bahl,  or when watch out! is substituted for hey I,  it still remains true that the expressive exclamation need not have any conative meaning and vice versa. 3. Panofsky's analysis mentioned above calls attention
to yet another opposition concerning the two types of meaning that relate the sign to the subjects. His idea of the subjectivity involved in the "organic movement" of the represented beings points to the fact that the referential meaning, too, has its subject-related - thab is, expressive and conative -components. Jakobson has sbown that in language the semiotic functions which relate to the subjects operate on two planes,
one involving the participants in the "speech event", the other
14
the participants in the "narrated event".      The same seems to apply, mutatis mutandis, to all sign systems (so that it might be advisable to modify Jakobson1s terminology and speak of the "signifying event" and the  "signified event").  Thus Gombrich deals with the arousal potential of what the images depict, which is of course different from the arousal potential of the colors, lines,  shapes or composition of the picture, and stresse. that a great variety of conventional visual signs can have the same potential:   "The symbols of religion such as the cross or-the lotus, the signs of good luck or danger such as the horseshoe or the skull and crossbones, the national flags or heraldic signs such as the stars and stripes and the eagle, the party badges such as the red flag or the swastika for arousing loyalty or hostility - all these and many more show that the conventiona
9
sign can absorb the arousal potential of the visual image."
Strictly speaking,  the expressive meaning of the
"signifying event"  is opposed to  the expressive meaning of the
"signified event":  one is an immediate meaning of the sign
whereas the other is mediated by the sign's referential meaning.
The some kind of opposition relates the conative meanings
produced on the two planes.  Within each of these oppositions,
the immediate and mediated meanings can both overlap, even
tend to merge, and contradict each other.  Their interplay is
16
particularly complicated and intense in the theater. 4-. The referential meaning, too, has its internal
opposition,  reality,  even the same reality, may be referred to as external to the subject (any subject) or else as a situation of which the subject is an integral part (whether as agent or as undergoer). The first of these referential meanings pertains to reality as something continuous and coherent,  sometimes even self-contained, and unified by its own intrinsic order. In the second,  it is conceived as separated into its single elements and its unity consists in these elements'  respective relations to the same subject-, that is, to the subject's (actual or potential) action on each or to the effect each exerts on the subject or to both.      This opposition does not cancel the object-related essence of the referential meaning because in both aspects the reality referred to in an object:  in the first, it is the object of contemplation and thought and in the second, the object of action (or inaction).'7
The third conclusion which logically follows from
10
Buhler's organon model is that the concept of sign as such
implies    no hierarchy of the three functions. Reference is the
dominant semiotic function of language in general, though not
IS
in every particular case of its use.      But the same is not true of other sign systems, such as gestures or traffic lights (a system particularly cherished by many semioticiaris;.
This absence of any a priori hierarchy is particularly important with respect to the conative function. Traditional linguistics usually recognized the difference between the referential and the expressive features of language but tended to limit the study of appeal to the grammatical description of the vocative and the imperative. As regards semiotics, Ogden and Richards decided, for reasons so insignificant that they ore beyond understanding,  to disregard in most cases the difference between the expressive and the conative function although they were perfectly aware of it.^ Esthetics has been haunted, ever since the period of romanticism, by the idea that art consists in expressivity, which not only obliterated art's semiotic nature and its specificity among the semiotic facts but also brushed aside the conative function;  the effect a poem, picture, statue or musical opus tends to exert upon the attitude of the perceiver was reduced to a mere echo - or resonance, as the language of that conception would call it - of the artist's expression "embodied" in the work. In other terms, the conative function appeared as a mere satellite of the expressive. And, despite Buhler and the subsequent studies of J'akobson, MukarovsK Gombrich and others,  this conception probably still prevails,
11
quantitatively,  in the literature about art:  when such terms as "expression",  "expressive" or "expressivity" are used, the reader must try to find out if what the esthetician has in mind really is the expressive meaning,  or if it is the conative, or if he is confusing the two kinds of meaning,  in art history alone, the term "expressive" is used in at least three entirely different senses.  In traditional usage,  it means the expression of the emotions of the figures represented in a picture or statue. A more recent usage applies it to the whole work, interpreted as an expression oi' the artist's inner state. At the same
time, most inappropriately,  it is used to designate the work's
21
ability to arouse emotions in the perceiver.^
2. EXPRESSIVE AND CONATIVE MEANINGS
In its quality of expression and of appeal the sign can convey a great variety of meanings ranging far beyond the feelings, moods,  states of mind, emotions, passions, desires, etc., which have been traditionally associated with the concept of expression.
Jakobson has adopted Anton Marty's idea to broaden
the area of expressive meanings by calling them emotive instead 22
of emotional. But this still limits the expressive function to meanings pertaining to psychological phenomena, for example, one cannot possibly call "emotive" the meanings which certain verbs denoting mental dispositions and operations convey when they are used in the first person. According to Benveniste's analysis, the first person itself signifies neither a concept nor an individual but the individual act of speech and designates
12
the speaker. The first person of such verbs as suppose, presume or conclude - 1 suppose (that...), I presume (that...), I conclu (that...) - does not describe the speaker performing the mental operation in question but implies that he adopts a certain
attitude towards what is asserted in the following subordinate
23
clause. * In fact, the sign as expression can convey any meaning
concerning its originator,  including the relation in which he
stands ot the topic, to the addressee, to the material situation
in which he produces the sign, to the sign itself and, through
the intermediary of any of these, his more general relations to
reality, society, people, etc. The sign can also express the
social category to which the originator belongs, his profession
or craft,  ethnic or regional origin and such like. Rabelais
uses this type of expressive meanings to create a peculiar
narrative device. Through a series of twenty-one oaths uttered
by a Parisian crowd he evokes, as Bakhtin pointed out, a certain
number of characters whom such a crowd typically comprised:
Jt'o cab de biousl points to the Gascon, Pote de Christol to th
Italian, Das dich Gots leyden Schendl to the German landsknecht,
far sainct Fiacre de Bryel  to the (French) fruit and vegetable
vendor, Je foys veu a sainct Thibaudl to the cobbler and Par
sainct Guodrin qui fut martyrise de pomes cuyttesl to the 24
drunkard.      Fione of the oaths reveals the niood or the state of mind of its utterer,  since the crowd swore, as Rabelais says, les ungs en cholere, les aultres par rys ("some in anger, sui..-in fun").
The conative meanings are just as varied.  To mention
13
just a few examples, the sign may aim to elicit information, empathy or motor reactions from the perceiver, to arouse him or to provoke affective responses of different sorts, to communicate something of which he is deemed  to be ignorant, to imply or insinuate something about the addressee, to persuade him, to induce him to act in a certain way,  to make him adopt a certain attitude towards the thing to which the sign refers, to make him actively contribute to the construction of the sign's referential meaning, to locate him inside or outside the represented reality, to assign him a specific place in the material situation in which he perceives the sign,  and so on.
At least four of these conotive meanings require further clarification, namely the elicitation of empathy, the appeal to make an active contribution to the construction of the referential meaning, the placing of the perceiver inside or outside the represented reality and the invitation to him to take a specific position in the material situation in which the sign appears.
With respect to empathy, the problem under consideratioi is conative meanings eliciting empathy, not empathy as a psychological process.  This needs to be emphasized for two reasons. On the one hand, the"psychological analysis of empathy has been clearly inadequate. On the other hand, many estheticians, art historians and theater critics have used this concept in an arbitrary and impressionistic way. The vogue of empathy as an explanatory principle could not last under these circumstances. For some time now, however, the problem has again been attracting
14
the attention of scholars,  so that serious reexamination and
26
elaboration ma,y be expected.      Anyway, as a conative meaning
produced by certain features of the sign, the appeal to empathy
is perceptible even when it fails to provoke an actual response.
The study of this conative meaning may lead to some
broadening of its scope, for instance, there may be elicitation
of empathy in such conative meanings as those carried by.what
Mathesius calls the adjunctive type of the inclusive first perso
plural.  In'this use, the first person plural substitutes either
for the first person singular or for the second person.  In a
scholarly text, for example, In the course of our reflexions we
have found that...    refers to the author's own activity and
Let us, however,  keep in mind that...  is an injunction to the
reader or audience.  In both cases, the first person plural
elicits the addressee's identification with the originator: in
the first by implying that the addressee participates in the
originator's activity;  in the second by implying that the origi-
27
nator shares the addressee's perception of the exposition. ' If this broader conception is acceptable,  the elicitation of empathy comes fairly close to, but does not merge with, another conative meaning listed above, which consists in implying something about the addressee.
When I say that the sign may aim to make the perceiver actively contribute  to the construction of its referential meaning, I have in mind such phenomena as the "beholder's share" in making sense of a picture,   that is, the appeal to his projective activities to compensate - from the stock of images stored
15
in his mind as a result of his perception of the world - for
28
what the picture itself does not show;      or the necessity for the beholder of a statue to imagine those parts that are hidden from him (for instance the saint's face in Bernini's  ecstasy of St feresa in the oanta Maria della Vittoria church in Rome) and, more generally,  to reconstruct the statue in his mind out of the partes pro toto he actually perceives as he moves towards, along or around it; or the listener's transposition of program music into extramusical meanings (the sophisticated listener may feel frustrated when too little effort is required to that effect, as in the final part of fchaikovski1s 1812 Overture); or the typical ellipticality of colloquial language  (the lack of it may give the interlocutor the impression of being lectured).
Concerning the ability of the sign to signify to the perceiver that his place is either inside or outside the represented reality, the jamb figures of the central porch of Chartres cathedral can be usefully compared to Bernini's bust of Cardinal Borghese (1632, Galleria Borghese, Rome). The appearance of each of the jamb figures is entirely coordinated with the corresponding column, their heads are straight forward and they are not related
to each other or to the beholder.  "Each figure lives,  as it were,
29
in its own world."      Cardinal Borghese,  on the contrary, "is shown in the act of speaking and moving." i'he mouth is half-open, his head turns slightly to his right, his eyes somewhat more, apparently fixing an interlocutor; the body is turned slightly to his left, suggesting other possible interlocutors  on that side. The beholder's basic position is straight front, which tends to
16
include him among the Cardinal's implied interlocutors.^0
Shapiro observes that Cezanne's landscapes do not
invite the viewer to enter or  traverse the space (the occasional
roads are empty and most often the vistas or small segments of
nature have no paths at all) and sometimes send him signals to
the effect thai; he must limit himself to looking (the path to
a house is blocked by a barrier in the foreground,  elsewhere a
deep pit or quarry lies between him and the main motif, or steep
rocks obstruct the way inwards, etc.).  In the Self-Portrait with
Palette of 188^-1887 (in the collection of Cezanne's family, .
Paris),  "the suspended palette in his hand is a significant
31
barrier between the observer and the artist-subject";      it may be added that the artist is seen here intently looking at a c canvas.on an easel which the viewer sees only from the back.
It is the other way round in Raphael's Saint Paul Preaching at Athens (carton in the Victoria and Albert museum, London), of which Gombrich says:   "...the artist turns us into participants of the momentous scene when the apostle of Chri^„ addressed the elite of pagan philosophers. We must envisage ourselves sitting on the invisible steps outside the picture, but the image shows us notiiing that would not be visible from one point at a given distance, a distance which could be worked out mathematically but which we feel instinctively.  (...) Every object in view is seen as it would be seen from the same point." The conative meaning described relies not only on the consistenc of the pictorial perspective but also on at least two auxiliary features, namely the platform and steps on the one hand and the two figures in the foreground on the other;  both are cut off by
17
the lower part of the picture frame, so that their invisible parts are located at the beholder's side of the picture plane.^ I have the impression,  but it is no more  than an impression, that the inclusion of the beholder in the represented scene is signified even more distinctly by the tapestry woven after Raphael's carton, in which,  of course, the.picture is reversed.
Wittkower singles out the features of (Jaravaggio' s Supper at Emmaus  (c.1600, National Gallery, London) which aim to draw the beholder into the represented situation: Christ's extremely foreshortened arm and the outflung arm of the older disciple, which seem to reach oat into the space where the beholder stands (actually,  only the disciple's arm does, not Christ's which only points in that direction), and the precarious position of the fruitbasket which gives the impression that it
'AIL
may at any moment land at the beholder's feet.
The sign's ability to attribute to the addressee a specific place in the material situation in which he perceives it manifests itself, for example,  in requiring the viewer of a picture constructed on the basis of linear perspective to stand opposite the center of the pjicture, lest the whole referential structure be distorted; some pictures tend to induce him to shift his station backwards and forwards,  some others to move along the picture plane, and so on.      Certain buildings are shaped m such a way as to be approached from one direction only. The i'igura serpentinata, which was so popular in the sculpture of the Mannerist period, incites the beholder to walk around the statue, in the direction the spiral turns from the bottom to the
18
top. Many of Henry Moore's sculptures incite to moving around,
backwards and forwards and up and down in variously alternating
directions. Still other statues call for a frontal view only.
Relief sculpture sometimes requires, to be fully seen, the
beholder's movement along its face, often combined with small
movements of  the head in varying directions. Brancusi incised
on two adjacent sides of a limestone base supporting a plaster
cast of Mile Pogany (1919, Brancusi Studio, Centre Georges
Pompidou, Paris) 3 frieze of a planar version of his Kiss (repea
ted three times on each side). As Geist points out, it can be
understood at a glance but the nuances caused by the accident
•56 •
of execution invite scrutiny;y    m other words, the beholder is invited to get closer to the frieze and move round the corner
3. METALINGUISTIC FUNCTION Jakobson introduces into the organon model a new function, which he calls metalingual, to single out such speeche or such moments of a speech as relate not to something externa to language but to the language itself - that is, to la laiigu . in de Saussure's terminology or "code" in Jakobson's. Language is used in this way whenever it serves to provide an "elucidatin interpretation of words and sentences - whether intralingual (circumlocutions,  synonyms) or interlingual (translation)...", whenever the speaker or the addressee or both "need to check up whether they use  the same code",  in "any process of language learning, in particular child acquisition of mother tongue", and so on. '
The role of metalinguistic operations in language
19
learning,  critical perception of statements and other utterances is manifestly crucial. Their relation to the semiotic functions defined by Biihler requires careful examination.  In Jakobson's famous diagram of the functions of language,  this one lies opposite the referential, just as the expressive ("emotive") does to the conative function. That, however, does not seem to imply a real antinomy between them since Jakobson points out
that the metalinguistic function is in diametrical opposition
-58
to the poetic.      In any event, it is conceptually difficult to give the metalinguistic operations the same status as the reference, the expression and the appeal. As mentioned before, Biihler found that each of these three functions endows whatever fulfils it with the quality of sign;  I fail to see how the same could apply to metalinguistic operations.
The "exasperating dialogue" imagined by Jakobson for illurtration is perhaps revealing in this respect. It goes like this:   ""The sophomore was plucked.'   'But what is plucked?1 'Plucked means the same as flunked.1   'And flunked?'   'To be flunkec is to fail in an exam. '   'And. what is sophomore?' persists the interrogator innocent of school vocabulary.   'A sophomore is (or means) a second year student.'" It is true that:  "All these
equational sentences convey information merely about the lexical
59
code of English:  their function is strictly metalingual."^y But is also true that all the questions and equational replies serve here to elucidate the referential meaning of the opening statement. Elam gives a different example of the use- of metalanguage in the construction of dramatic dialogue.  In Edward Albee's
20
Who's Afraid of Virginia Woolf? there is the following exchange: "G-eorge:  Well, you'll get over that... small college and all. Musical  beds is the faculty sport around here. Nick: Sir? G-eorge:  I said, musical beds is the faculty... Never mind. I wish you wouldn't go  'Sir'  like that... not with the question mark at the end of it. You know? Sir? I know it's meant to be a sign of respect for your (Winces) elders...  but... uh... the
way you do it... Uh... Sir?     Madam? Nick: No disrespect inten-
4-0
ded."      In this example the bits of metalanguage serve to elucidate the expressive meaning of an utterance. Karcevskij has shown that every application of a linguistic sign in a speech act involves a more or less complicated process of choosing along the homonymic and synonimic axes, therefore a metalinguistic operation (a term he did not use).^
It would no doubt be rash to conclude that metali-tic operations are auxiliary to the referential, expressi conative function, ot to any combination of the three, but this aspect would probably repay the linguists' attention.
Another question arising in this connection is wheB:
metalanguage is a specifically linguistic or, on the contrary,
a general semiotic fact (like the three functions defined by
Bilhler).  It is probably relevant in this respect that language
is the most universal means of communication and cognition and
that no other type of signs constitutes so vast, coherent,
diversified,  integrated and constraining a system as what
4-2
de Saussure called la langue.      Thus Benveniste declares that "language is the interpretant of all semiotic systems" and adds:
21
"No other system has at its disposal a langue in which it could
categorize itself and interpret itself according to its semiotic
distinctions, whereas language can, in principle, categorize
43
and interpret everything, including itself."      At this stage,
I am not ready to go so far; in my view, the various sign
systems must be thoroughly examined in this respect. It is
perhaps indicative that in pictures, statues and works of music
some of the metasemiotic functions which in a speech could be
fulfilled by metalinguistic operations actually fall on language.
Gombrich has demonstrated that the caption can be crucial for
44
the elucidation of the meaning of a picture      and his findings could easily be extended to sculpture and music: thus, to the perceiver who does not know its name, Aristide Maillol's The River (c.1939-1943, New York, Museum of Modern Art) represents a woman; Cooke observes that if Debussy had not given the first of Three Nocturnes its title, Nuages, the listener would be
45
uncertain about the meaning of the shifting patterns of sound; in tests concerning the perception of meaning in music, the subjects' responses are checked against the name of the composi-tion.^ Jakobson's analysis of Henri Rousseau's picture The Dream and of the poem the painter wrote to be attached to it has revealed the complex and subtle ways in which the poem elucidates the meaning of the picture, of its different parts,
47
compositional devices, etc. '
On the whole, it seems to me that the question of the exact semiotic status of the metalinguistic operations remains open.
22
4. PHATIC FUNCTION
The felicitous term "phatic", by which Malinowsk -labeled "the case of language used in free, aimless, social intercourse", is adopted by Jakobson to designate yet another function he adds to the organon model, a function pertaining specifically to the contact between the originator and the addressee ("to establish, to prolong, or to discontinue communication, to check whether the channel works..., to attract the attention of the interlocutor or to confirm his continued attention...").^
A considerable amount of analytical work would be needed to explore the relations of this function with the expressive and the conative functions so as to delimit it from the one and the other not only within the logic of the comuiunica tion theory but also on empirical grounds. Another question to be carefully examined is whether the phatic function is part of the language system (la langue) or merely a matter of the situation in which the utterance takes place. Thirdly, as for the metalinguistic operations, it is necessary to find out whether t/ie phatic elements of a speech are not simply auxiliary to the referential, expressive or conative function - as the case may be in any single occurrence.
Language is evidently the privileged field for this analytical work. But the same question must be asked with respec to the other types of signs that also bring the emitter into direct contact with the receiver, such as gestures, music in performance and the theater. Even such transparent conventional
23
devices as turning off the light and raising the curtain in the
theater are not negligible,  especially when their much more
complicated equivalents in other forms of theater are taken into 49
account.      But I have esx^ecially in mind the actor's problem to "keep in touch" with his fellow-actor or fellow-actors when they are on stage together and the extraordinary gift and skill that alone enable an actor to "do nothing" for a certain length of time - that is, neither speak noJ~ move nor change facial expressions - without losing the audience's attention;  in this, inexperienced or mediocre actors invariably fail. The investiga- . tion of the contact between the emitter and the receiver in these areas other than language would naturally require that the term "phatic" be replaced by a more general one.
Ultimately, the enquiry would necessarily extend to such signs as do not bring the emitter into direct contact with the receiver - pictures, statues, works of architecture, films, etc. - so as to find out whether the factor identified by Jakobson is a universal semiotic phenomenon. To this effect, considerably more than a substitution of a more general term for "phatic" would be needed:  a redefinition of the whole problem, Logically, in the case of these types of signs, the contact between the emitter and the receiver passes through the intermediary of the sign itself. Therefore, the question is whether the sign has any features that serve specifically to establish contact with the potential receiver of its meaning, to attract his attention or to secure his continued attention.  In other words, whether the sign has any features that do these things
24
otherwise than just carrying the meaning of the sign. A single
example may be enough to illustrate the problem at this stage.
With all the necessary reservations due to the novelty of the
whole concept and even more to the lack of clarification in
the area from which it has originated - language - I may suggest
such an example. Palladio, who like the other Renaissance
architects and theoreticians of architecture was concerned with
the ways of making the temple convey its conative and referentia
meanings to those who are inside, considered that a temple
should stand on a high base. And, unlike Alberti, he favored
this not only as a means to isolate the temple from the everyday
life that surrounds it but also because to reach the temple by
50
mounting steps inspires people with devotion and awe      (in other words, makes them receptive to the meanings to be conveyed to them by the interior of the temple). But, of course, the questic immediately arises whether this is not another aspect of the conative function. Nonetheless, the matter is certainly worth pursuing.
Another aspect of what Jakobson calls the phatic function must also be kept in mind. When Malinowski used the term "phatic", he did not have in mind anything conceptually comparable to reference, expression and appeal. He spoke of a particular kind of discourse. He did use the term "function" in this connection but in the sense of utilization or instrumentality ("the function of Speech in mere sociabilities") and he called "phatic communion" the type of linguistic use he was examining.^1 Quirk speaks of phatic use of language, rather
25
than function in Buhler's sense and illustrates it, among other
things, by a passage from a detective story concerning a
sentence at tho dinner table "that made sense but had no mean-52
ing".      Even one of Jakobson's examples is a whole phatic conversation borrowed from Dorothy Parker.  ' All this concerns, in Malinowski's words,  "one of the bedrock aspects of man in society", namely "the fundamental tendency which makes the mere presence of others a necessity for inan";y    it hardly concerns the organon model of the sign.
5. ESTHETIC FUNCTION In 193^, the same year as Sprachtheorie was published, Mukarovsky, who was familiar with Buhler's organon model, read at the International Congress of Philosophy a paper which extended the scope of semiotics to the whole area of art and
55
even, at least potentially, of esthetic phenomena in general. The most original feature of this contribution was in that it conceived the work of art not only as a set of signs - as some other estheticians had done before - but also,  and chiefly, as a single sign. By the same token, the question of the specificity of art in the whole realm of signs arose. Mukarovsky1s bold answer was that, unlike the communicative sign /Mitteilungszei-chen/, the work of art was both an autonomous sign relating to the total context of social phenomena, or collective views, rather than to a distinct delimited reality, and a communicative sign; thectwo aspects were inseparably united in a dialectical antinomy, although the intensity of communicative elements varied from art to art. The boldness of Mukarovsky1s thesis
26
is well illustrated by the fact that even a scholar so close to the Prague Linguistic Circle as H.J.Pos mistook the dialectical antinomy for a contradictio in adjecto and argued, in the discussion at the Congress, that the sign belonged to the practical sphere.^
In this first outline, Mukarovsky did not refer to the organon model but when he developed and revised his conception two years later, in a paper read at the International
57
Congress of Linguists,  '  he identified the esthetic function as the factor that makes the work of art an autonomous sign and placed Buhler's model right in the center of the argument. The estehtic function focuses attention on the sign itself and so opposes the referential, expressive and conative functions which are oriented towards entities exterior to the sign. These functions, described by iiukafovsky as practical, are not suppressed but merely subordinated to the esthetic function in the work of art. On the other hand, the esthetic function, which has a dominant position in art, is potentially present in all human activity. The esthetic function thus appeared "as a ubiquitous dialectical negation of the three fundamental functio; and, therefore, as a necessary complement of Buhler's schema"; in the same paper Mukafovsky also referred to the esthetic function as "the fourth function".
Although he strongly rejected the theory that expressivity is the essence of art and gave Buhler's model such a crucial place in his own theory, Mukarovsky did not elaborate on the relationships of the esthetic function with the expressive
27
and the conative. On this subject, he did not go beyond stating that sometimes the "practical" functions are prominent in a work of art, for instance the referential in the novel and the
expressive in poetry; but Biihler, who did not account for the
58
esthetic function, had pointed that out himself.      In other words, Mukafovsky concentrated, perhaps unwittingly, on the antinomy between the esthetic and the referential function; in this respect he now pushed his conception considerably further. Without giving up his previous view that in its quality of sign the work of art related to the entire context of social phenomena rather than to a particular reality, he now stated that the "weakening of the relation between the sign and the reality it points to" /realite vis6e/, which is due to the esthetic function,  "does not exclude, and even supports, the existence of the relation between the work and the universe" and added that art influences in this way the perceiver's whole conception of reality.
In his later writings, Mukafovsky developed his ideas concerning the relationship between the semiotic functions in art. He observed that the orientation towards the sign, as brought about by the esthetic function, means a concentration of attention upon the sign "in all its variety, especially its functional variety" and stated that in its esthetic orientation the language of literature oscillates freely between the referential, the expressive and the conative function and that it can at any time attach itself to and also detach itself from any one of them, and combine them in various ways;   that is the
28
epistemological consequence of its emancipation from a unilatera bond with any particular one.^jTlere Mukarovsky no longer concentrates on tue antinomy of the esthetic function with the referential; the antinomy of the esthetic and the communicative aspect of the sign bears upon the entire relationship between the esthetic function on the one hand and, on the other, the referential,   the expressive and the conotive functions as well as their various configurations. Yet, without giving up this much broader conception, Mukarovsky somewhat attenuated "it in another article, written some years later and again dealing specifically with literature,  in which he based his analysis on the assumption that the referential function of the sign was the most important of the three. He explained, indeed, that to fulfil its semiotic function the work of art, like any other sign, needs two subjects:  the emitter and the receiver. But in other types of si&ns the most important is the relation between the sign and the thing it stands for (referential relation), whereas in the case of the work of art this relation is weakened it is therefore the relation between the sign and the subject that comes to the fore.    /j-laybe this assumption concerning the primacy of the referential function was due to the fact that though he meant to formulate a general semiotic principle, Mukarovsky had in mind more particularly the verbal art. As already mentioned, reference is generally the dominant semiotic function of language but the same does not apply to all sign systems.  In acting, for instance, the expressive and the conatii functions are prominent. Elaborate procedures are needed wheneve
29
the expressivity of the stage figure and stage action are to be diminished and the conative effect may go so far as to induce the spectators to participate actively in the performance;^"'" at the same time, the real possibility to scale down the expressive function in spite of the a priori expressivity of the human face, voice, delivery and bodily behavior confirms Mukarovsky's thesis that the esthetic sign can at any time attach itself to or detach itself from any one of the three semiotic functions.
In the same article, Mukafovsky defined the relationshi between the conative and the expressive function in the work of art in a way diametrically opposed to the theory of expressivity and the concomitant theory of resonance (according to which the state of mind the work of art provokes in the perceiver is a resonance of the artistrs state of mind as expressed by the work). Mukarovsky stated that the unity of the work of art, which is one of its most essential characteristics, is in fact the mental state which the work induces in the perceiver or, more exactly, the act by which he appropriates the work (an act that is not basically different from other kinds of apperception) Yet, since the work of art is external to him, the perceiver projects his own state of mind into the originator of the work. The image of the artist that arises out of this process need not correspond at all to the artist's real person.
As regards the entire relation of the esthetic sign to reality, according to Mukarovsky's ultimate 'formulations the esthetic function induces a unifying attitude to reality. Reality
30
is not the immediate hut the mediated object of the esthetic
function.  Its immediate object is the esthetic sign which
reflects the whole reality, organizing it into a unity in the
image of man's own unity;  and, without losing its autonomy, the
esthetic sign projects man's attitude realized in its internal
structure into all reality as its general law. It manifests
its autonomy precisely by always relating to reality as a
whole, not to one of its particular segments; in as much as a
work of art points to a particular, it demonstrates on it all
the other particulars as well as their totality,  that is, 63
reality.      fhe esthetic function constantly renews man's aware-
64
ness of how manifold and varied reality is.
The contention that the esthetic function is dominant in art cannot be defended on empirical grounds; it is contradicted by such important art forms as architecture, religious art, polemical poetry, didactice art, etc. Mukarovsky himself
was aware of it and tried out various ways of overcoming the
65
contradictions      His last formulation was that as an esthetic sign the work of art receives its "content" from its extra-esthetic functionj, which put it in direct contact with realities extrenal to it, while the predominance of the esthetic function merely prevents any of these extra-esthetic functions from completely prevailing and dap ting the organization of the work to a single extgp-nal end; the esthetic function, in other words, controls the others by organizing their mutual relations and the tensions among them so as to bring out the multiplicity of
66
the functions concentrated on a single thing, the work of art.
31
In this formulation, the dominant position of the esthetic function is reduced to little more than a way of speaking; as such, it can be useful as a shorthand formula.
Another question mark hangs over the comparability of the esthetic function with the three functions analyzed by Blihler. As already mentioned, Mukarovsky put all four on an equal footing and sometimes called the esthetic "the fourth function".^ Yet he apparently had some doubts about the adequacy of this conception since in the same paper he also distinguished reference, expression and appeal as "the three fundamental functions of language" /les trois fonctions fondamen-tales de la langue/ from the esthetic function as their "ubiquitous dialectical negation" /negation dialectique omni-presente/. ,|_J*1:10 term "function" seems to have been
applied indiscriminately to phenomena of two different orders. In the case of the referential, expressive and conative functions it designates three aptitudes of the sign to relate to realities external to it and three kinds of meanings the sign can have. In the case of the esthetic function, it designates the sign's relation to its end, that is, to the end for which it has been produced or to which it is used (whether by the emitter or by the receiver). To put it differently, with respect to reference, expression and appeal, the term comes close to the biological notion of function; it is nearly synonymous with such terms as "ability" or "performance". V/ith respect to the esthetic function   it is used in the specific sense it has acquired in linguistics and social science. The confusion arose out of the
32
fact that more often than not social science uses "function"
68
both ways;      still worse, as Mathiot and Garvin point out, linguistics and social scence seem to use the term indiscriminat
A
69
ly to designate the role, the purpose or domains of use.
In the case of the semiotic functions, an additional moment increases the confusion. Besides being the sign's basic, semiotic competences, reference, expression and appeal may also be, respectively, the main ends to which the sign is used: some signs are chiefly referential, others expressive and still other conative. The sign, then, assumes a function of reference, of expression or of appeal in the same sense as it can assume the esthetic function. Only in this connection would it be justified to characterize the reference,  the expression or the appeal as practical functions by opposition to the esthetic.
Still another complication which can easily create confusion must be taken into account. The sign's ability to relate respectively to the designatum, the originator and the perceiver is die to its distinct structural features (such as the three grammatical persons in language, to mention the most trivial example). But thejuse of the sign to an end which is chiefly referential or expressive or conative need not necessari rely on the corresponding structural features.
So for example the first-person pronoun is a typical vehicle of the expressive aptitude of language because, as Benveniste points out, it refers neither to a concept nor to an individual but to the individual act of speech and designates the speaker, yet Husserl uses the first person in describing
33
the relation of the subject to the object, although his writings are rigorously referential. Sometimes he even combines it with the use of^demonstrative pronoun which strengthens the link to the speech act by evoking the material situation in which it takes place. But under the impact of the marked referential end of the entire text, the expressive meaning of these features is so subdued that the first person can shift, without any disruption of the sense,  from the singular to the expressively much more neutral plural inside the same sentence:  Intention ist nicht Erwartung,  es ist ihr nicht wesentlich,  auf ein künftiges Eintreten gerichtet zu sein. Wenn ich ein unvollständiges Muster sehe, z. B. das dieses Teppichs, der durch Möbelstücke theilweise verdeckt ist, so ist gleichsam das gesehene Stück mit Intentionen behaftet, die auf die Ergänzung hinweisen (wir fühlen sozusagen, dass t£ie Linien und Farbengestalten im "Sinne" des Gesehenen fortgehen);  aber wir erwarten nichts (Logische Untersuchungen, II, 1901, p.512).
It is said that, speaking of Thomas ä Becket in front of his courtiers Henry II exclaimed in anger Who will free me from this turbulent priest?, whereupon four knights went away and murdered the archbishop; the king was shattered when he learned what they had done. This semiotic story shows how the conative structure of a sentence uttered to an expressive end can induce the addressee to mistake it for an appeal.
Bühler mentions that a cLiefly conative discourse can draw on the referential structures of language:' "Der Weg über die Darstellung kann einem Appel dienen und es gibt selbst
34
innerhalb der Darstellung Stellen, wo der Führer den Geführten
durch Kenntnis von Faktoren und Umständen hinter der Kulisse ode
ganz allgemain gesagt, eben durch eine überlegene Lenkungstechni
dahin bringen kann, wo er ihn haben will. Nietzsche hat das so
ausgedrückt: Man braucht die Lüge nich mehr, wenn man vob-der
Wahrheit genug hat und kann man mit ihr die Menschen führen und
70
verführen, wohin man will."      The same point was made by Ogden
and Richards:  "In symbolic /referential/ speech the essential
considerations are the correctness of the symbolization and the
truth of the references. In evocative speech the essential
consideration is the character of the attitude aroused. Symbolic
statements may indeed be used as a means of evoking attitudes,
but when this use is occurring it will be noticed that the truth
or falsity of the statements is of no consequence provided that
71
they are accepted by the hearer.1"
A critical examination of the different concepts lying behind the term "function" may also throw some more light on the problem of the phatic function. The difference between phatic elements inside a discourse which itself is not chiefly phatic and the phatic type of discourse - or "phatic communion" - has something in comuion with the difference between Bühler's and Mukafovsky's respective concepts of function. And the chief] phatic discourse has this in common with the chiefly referential expressive or conative use of the sign that it need not necessarily rely on specifically phatic elements of language (provide: i the phatic function is part of the language system, which of course remains to be demonstrated): the sentence "that made sent
35
but had no meaning" in the situation which Quirk quotes as an example of the phatic use,  "Have you had a busv day, dear?", could by its intrinsic structure be just as well conative (eliciting information) or even expressive (displaying concern).
6.  ART AS REFERENCE, EXPRESSION AND APPEAL
According to Mukafovsky the esthetic function transforn everything within its reach into a sign, and at the same time it is a dia^ctical negation of the outward-oriented relations of the sign to the denotatum, to the originator and to the perceiver in that it focuses attention on the sign itself. If this is true and if Buhler's thesis that reference, expression and appeal are the basic semiotic performances of the sign is also true, then three conclusions logically follow:
First, in transforming   something on which it rests into a sign, the esthetic function must by the same token transform it into reference, expression and appeal; to put it differently,  it must endow it with referential,  expressive and conative meanings.
Second,  the dialectical negation of the outward-oriented relations of the sign cannot consist merely in that the esthetic sign may but need not really relate to a denotatum, may but need not express something real pertaining to its originator and may but need not bring about the response corresponding to its intimations. As already pointed out, all this is true of any sign, not only of the esthetic; it is due to the fact that what the sign produces in its qualities of reference, expression and appeal are meanings - referential, expressive
36
and conative meanings. The dialectical negation must do somethin more than that. If it is truly dialectical, then it not only weakens or undermines the outward-oriented relations of the sign but also strengthens its aptitudes to enter into such relations;  in other words, it must somehow strengthen the referential, expressive and conative meanings of the sign.
Third, these two conclusions are not mutually exclusiv and are indeed complementary, for they follow from the same premisses.
in the present state of semiotics there is no question of providing proof that the esthetic function actually does, or does not, produce these effects. The two    suggested conclusions merely indicate directions for further exploration.
There is probably no better way of testing the first conclusion than to study the problem of referential meaning in music. This, however, is one of the most difficult among those which semiotics has to face. At the same time, musicology is en extremely specialized discipline; a considerable risk of misinterpretation, or at least of misunderstanding, is therefore* implied in any semiotician's attempt to interpret the musicologist's findings in a more general conceptual framework.
Music can have referential meaning. In spite of the tradition of formalism on the one hand and of the theory of expressivity on the other, which are very strong in their discipline (and would perhaps be overwhelming were they not mutually exclusive), musicologists keep coming up against this simple fact and looking for ways of analyzing it. In my view,
37
these quests could be more fruitful if what is probably the worst
methodological error in this field, namely attempts to translate
musical meaning into words, were altogether eliminated. Because
of the way it is produced and of its peculiar qualities, the
meaning of music cannot be translated into language without
gross distortions; language can serve only to analyze or, as
mentioned above, to provide a "metasemiotic" interpretation by
means of the name given to a musical work.
Referential meaning seems to arise in music out of
the context of the whole composition, that is,  the order in
which successive components follow on,  and lead up to, each
other. Strictly speaking, it is a succession of the elementary
units, the notes organized both in simultaneous combinations and
in sequence according to the same scale. The successive tones
tend to cluster into more complex units, but as the context
unfolds such configurations may be undone just as they have
arisen, and may also be reconstituted again, depending on what
comes next and what later; unlike language, the music system
comprises no principle of the obligatory combination of lower
units into higher ones (it has no morphemes, words, phrases,
sentences), and a sequence that is a unit in one respect need
not be one in other respects.  Following ilusserl's analysis of
72
the temporal object,      it could be said that the configurations that are undone by what follows them not only recede into the past to remain present merely in the perceiver's consciousness: they disintegrate as they elapse.  The perceiver's consciousness still retains them but in an ambivalent state as at once being
38
and not being. ^Similarity (with its complement, contrast) betwee
what precedes and what follows is an important though not the
only factor in the constitution and resurgence of higher units
in the flow of a musical composition. And it is also chiefly
though not exclusively by way of similarity under its two forms
of resemblance and isomorphism (diagrammatic similarity) that
73
these units are connected with their referential meanings.
Ruwet gives an illuminating example, a tonal musical fragment
made up of two parts whfeich begin in the same way but the first
ends with a deceptive cadence while the second ends with a
perfect one:  this structure, namely a movement begun and suspend
ed,  then resumed and  brought to completion,  is homologous with
an indefinite number of structures existing in reality and
74
belonging to the listener's experience.      The way in which such meanings enter into  the much richer and more complex sense of the musical work as a whole is another affair, but in its referential aspect this integral sense of the work, too, is as a rule homologous with an indefinite number of realities and experiences.
Three qualities seem to characterize the referential meaning of music, i'lrst, its generality, which has just been mentioned:  it pertains to an indefinite number of empirical facts. To narrow it down to the particular, music ^resorts to special devices such as the title of the composition, sound painting,: the symbolism of numbers, conventional musical quotation, etc. Gecond, music refers primarily to processes, creates so to speak a Heraclitean image of reality; that which
39
moves or is in process is not referred to or only incidentally (often by special devices) or indirectly, through its -metonymic relation to the process itself. Third, the processes to which music refers are qualitatively indeterminate, reduced to their structure; the musical fragment described by Ruwet as "movement begun and suspended, then resumed and brought to completion" may pertain to the structure of a natural process, of animal movement, of human work, of intellectual effort, of an emotion, and so on.
The fact that music produces referential meaning is
very remarkable because of the unique nature of music among the
semiotic systems, which particularly strikes the linguists.
Whorf put it this way:  "Music is a quasilanguage based entirely
75
on patternment, without having developed lexation."'^ In Benve-niste's formulation, music has no units "directly comparable to the 'signs' of language".^^
It would be tempting to say that music owes its aptitude to produce referential meaning to the esthetic function, which of course attaches itself to the musical work, not to the system of music. But that would be premature until other possible explanations have been fully examined. Logically enough, the other possible explanations fall into the field of the second of the conclusions suggested at the beginning of this chapter, namely that the esthetic function also strengthens the sign's referential, expressive end conative meanings.
In any attempt to explain how referential meaning can arise in music, the relations between the referential and the
40
conafcive meaningsjseem to be especially important. Yet to my (admittedly very limited) knowledge they have not been adequate! studied so far, partly because musicologists frequently confuse the conative with the expressive meanings and sometimes the referential with the expressive.
Music has a strong conative potential ranging from such rather obvious things as elicitation of a great variety of motor reactions or provocation of arousal, tension and affective responses to much subtler meanings which are more difficult to describe. Paul Valery makes Socrates reflect on one of them in the dialogue Eupalinos ou 1'architecte. He declares that human beings undergo music as they undergo the space created by architecture; they are totally encompassed by it.  "We then are, we move, we live in the work of man!" /MNous sommes, nous nous mouvons, nous vivons alors dans l'oeuvre de 1'hommeiV And he touches upon various referential implications of this conative meaning of music. What appears to be the most important among them is described in the following question: "hid it not seem to you that an intelligible and changing space was substituted for the original space; or,rather, that time itself surrounded you from all parts?" /"Ne te semblait-il pas que l'espace primitif etait substitute par un espace intelligib] et changeant; ou plutot,  que le temps lui-meme t'entourait de toutes parts?V^£f'he referential reflex of this conative meanir can be due only to the esthetic function of music. A man in a busy street or in a printing shop also "is, moves, lives in the work of man" in the sense that he is totally encompassed by man-
41
made noise. And if the traffic in the street is heavy, or the printing shop big, the noise has a perceptible pattern.  Yet my paraphrase of Valery's beautiful formula is ridiculous, just because the esthetic function is not involved in this man's experience. For the same reason, he does not have the sensation that "an intelligible and changing space... etc." jXnother one of the subtle donative meanings of music consists in that the clustering of tones into more complex units, and their possible disintegration and resurgence, stimulates the listener's anticipation, at every point of the context, of what is forthcoming. This constant anticipation,  too, has referential implications in so far as referential meaning arises in music out of this process of constitution, disintegration and resurgence of higher units.
Gonative meaning can also reflect on the referential in architecture.  In Book I, ch.2, para.5,  Vitruvius declares that there is propriety "...in the case of hypaethral edifices, open to the sky, in honor of Jupiter Lightning, the Heaven, the Sun, or the Moon:  for these are gods whose semblances and manifestations we behold before our eyes in the sky when it is cloudless and bright. The temples of Minerva, Mars, and Hercules, will be Doric, since the virile strength of these gods makes daintiness entirely inappropriate to their houses. In temples to Venu3, Flora, Proserpine, Spring-Water, and the Nymphs, the Corinthian order will be found to have peculiar significance, because these are delicate divinities and so its rather slender outline, its flowers, leaves, and ornamental volutes will lend
42
propriety where it is due.  The construction of temples of the Ionic order to Juno, Diana, father Bacchus, and the other gods of that kind, will be in keeping with the middle position which they hold; for the building of such will be an appropriate combination of the severity of the Doric and the delicacy of the Corinthian."^[yitruvius himself uses these examples to illustrate his concept of propriety that arises "from prescription" as distinct from propriety arising "from usage" or "from nature" and Gombrich has made the crucial point that the meaning: he attributes respectively to the Doric,  the Corinthian and the Ionic style can arise onlv because the perceiver's response and his interpretation of what (or whom) the temple refers to are limited in number by the expectations of possibilities and probabilities based on a "matrix" or "scale that has intelligil
dimensions of 'more'  or 'less'" - the rigid orders of ancient
79
architecture.      fins is no doubt the sense of Vitruvius1 term "by prescription",  what matters here is that conative meanings are involved:  the absence of the roof hints to the beholder that his view of the sky is unobstructed, the relative lack of adornment in the Doric style and its relative sturdiness provoke the rejection of "daintiness",  etc. And these conative meanings have referential implications:  the virile strength of certain gods,  the delicacy of others, the middle position of the characteristics of still others;  the supremacy of Jupiter Lightning is referred to by means of continuity with the sky, the sun and the moon, oince in ancient architecture the esthetic function wa, linked to the norm of the orders, Gombrich1s interpretation show:
43
in fact that the conabive meanings owed their extraordinary referential implications to the esthetic function.
The case of architecture is, however, different from that of music in that,  as a rule, the building has a meaning irrespective of the esthetic function;  it signifies its practical functions.  Therefore,  the referential meanings deriving from the conative in Vitruvius1  examples do not go so far as indicating that the buildings concerned are temples;  they merely specify the characteristics of the gods to wAom the temples are respectively dedicated.
The conative meaning can also imply, or reflect.on, referential meanings in thematic arts which derive from referential types of signs,and the referential meanings can at the same time reflect on the conative. In Cezanne's Poplars (1880-1882, Louvre), for example, the relatively dark mass of short color strokes of more or less similar color1, shape, size and slant, which covers mu h    of the upper three quarters of the picture, refers to tree leaves.  The referntial meaning displays an opposition between the compact mass and its innumerable particles because the trees themselves,  thoup-;h obviously numerous, are not clearly distinguishable as single units and the whole mass is broken down not into separate leaves but into single brush strokes and fairly distinct but thematically unmotivated series of brush strokes. This referential meaning leads up to a conative one: the beholder's task to get a reasonably coherent sense out of the picture requires long scanning in the course of which various referential meanings arise as the eye wanders
44
through the mass of strokes in various directions, discovers certain elements of its articulation into single trees and at the same time certain elements of other articulator;/ principles, and so on. And this restless movement of the eye reflects back on the referential meaning. Though no movement is actually represented on the thematic plane, the opposition between the compact mass and its innumerable particles also appears as an opposition between stability and instability, of rest and movement, of statics and dynamics. This procedure is the more remarkable since it cannot be explained by the limitations of the referential potential of the sign systems involved. Pictures can produce such referential meanings without the detour through conative meaning.  The complicated construction of the picture's sense which Cezanne chose is linked to the esthetic function.
The ability of one type of meaning to reflect on another type of meaning of the same sign is by no means limited to the relationships between appeal and reference. Any one of the three types of meaning can reflect in this way on the other two and in. its turn reflect their implications. Rabelais' list of oaths, mentioned above,  shows how a series of expressive meanings makes up a referntial meaning and how this referential meaning brings out one of the many expressive meanings an oath can have. In Hölderlin's untitled fragment which begins with the; stanza Wenn aus der Ferne, da wir geschieden sind, / Ich dir not kennbar bin, die Vergangenheit / 0 du Teilhaber meiner Leiden! , Einiges Gute bezeichnen dir kann,  (Sämtliche Werke, ed. Franz Beissner, vol.II/1, pp.262-263) and in which the speaker is a
4-5
loving woman and the beloved man from which she is separated is the addressee, the expressive, conative and referential meanings are so intertwined and reflect so intensely upon each other that the reader cannot possibly distinguish between them;  I have tested the fact that this semantic structure does not depend on the reader's knowledge of the poet's biography.
There are, nonetheless, some reasons to pay attention in this respect to trie conative meanings in particular, for the conative performance of art seems to be connected with the esthetic function more intimately that  the other two are. According to Kris, the actual conative ei'fect of the work of art, in other words the response to its conative meanings, develops in successive stages depending on how long the work is viewed, and seems to be qualitatively different at evei'y stage — provided the length of the viewing is esthetically motivated.^ Combining a solid art historian's erudition with a long experience in active psychoanalysis, Kris was extremely highly qualified for this sort of observation. His findings should be checked by experimental methods;  and since he derived them mainly from painting and sculpture, they should also be examined with respect to other arts. None of this can of course rehabilitate the esthetic theories that try to explain the esthetic function in terms of psychological effect. The esthetic function stems from the perceiver and is directed at the sign; it is the other way round with the psychological effect.  Secondly,  the esthetic function is essentially a social fact merely subject to individual variations; the psychological effect is individual by its very
46
nature and has only the validity of a symptom.
The ability of each of the three types of meaning to imply, and reflect on, the other two is not limited to esthetic signs. How far its remarkable intensity and scope in art is due to the esthetic function is a question which must remain open. It will have to be subjected to analytical and comparative studies of the different arts, and of phenomena outside the real; of art as well. The study of the organon model and its implications is as important today as it was half a century ago. That only confirms what many scholars felt at the time - that Buhler', three-function model is one of the milestones in the history of modern semiotics.
4 7
FOOTNOTES
1. Karl BUhler, Die Axiomatik der Sprachwissenschaften. Frankfurt an Main; Vittorio Klostermann,  1969^, and Sprachtheriep Jena: Gustav Fischer, 1934.
2. In an earlier article I said that BUhler "derived his conception from Plato's Kratylos. where speech is defined as an instrument by which somebody tells somebody else something about things" (cf. my "Contribution to the Semiotics of Acting" in: Ladislav Matejka, ed., Sound. Sign and Meaning. Ann Arbor: Michigan Slavic Contributions, 1976, p„ 576). This was due to an oversight. BUhler claimed the organon model had been so clearly conceived by Plato that a logician could have deduced it from his common sense proposition. But he also pointed out that both him and Alan Gardiner had discovered the model, independently from one another, in the course of their respective explorations of language, not through an exegesis of Plato's dialogue (cf. Die Axiomatik der Sprachwissenschaftent pp. 94-95, and Sprachtheorief pp. 1-2 and 25-26).
3. Cf. Sprachtheorier p. 29.
4. Cf. ibid.. p. 28.
5. Cf. Die Axiomatik der Sprachwissenschaftenf p. 102, and Sprachtheorie, p. 32.
6. Cf. Roman Jakobson and Grete LUbbe-Grothues, "Ein Blick auf Die Aussicht von Hölderlin", in: Roman Jakobson, Hölderlin - Klee - Brecht. Zur Wortkust dreier Gedichte. Frankfurt am Main: Suhrkampf, 1976.
7. Cf. Karl Jaspers, Strindberq und van Gogh, Bern: Ernst Bircher, 1922, p. 124.
8. Cf. E. H. Gombrich, "Expression and Communication", Meditations on a Hobby Horse. London: Phaidon, 19712 (paperback), p„ 68.
9. I. A. Richards, "Factors and Functions in Linguistics", Poetries:
❖ 8
Their Media and Ends. The Hague: Mouton, 1974, p. 10.
10. Cf. E. H. Gombrich, op. cit.f p. 62.
11. Cf. Holger Steen S^rensen, "Meaning and Reference", in: A. J. Greimas, R. Jakobson, M. R. Mayenowa et   al.r eds., Sign - Language - Culture. The Hague: Mouton, 1970.
12. Cf. Hans Joachim Albrecht, Farbe als Sprache, K'dln: DuMont, 1974,
pp. 24-25.
-a
13. Cf. Erwin Panofsky, "The History of the Theory of Human Proportions as a Reflection of the History of Styles", Meaning in the Visual Arts. Harmondsworth: Penguin (paperback), n. d., pp. 128-129.
14. Cf. Roman Jakobson, "Shifters, Verbal Categories, And the Russian Verb", Selected Writings, II, The Hague: Mouton, 1971, pp. 133-136.
15. E. H. Gombrich, "The Visual Image", Scientific American, September 1972, p. 93.
16. Cf. my ^Contribution to the Semiotics of Acting", op. cit.f pp.
577-578.
17. Cf. my Drama as Literature, Lisse: The Peter de Ridder Press, 197
pp. 87-89.
18. BUhler insisted that his three function model tends to 1i» not deny the predominance of the referential function in language (cf. Die Axiomatik der Sprachwissenschaften, pp.  101-102 and 112, and Sprachtheorief pp. 30-31).
19. BUhler himself examined gestures (in the broadest sense) as a predominantly expressive and conative type of signs (cf. his Ausdruckstheorie. Jena: Gustav Fischer,  1933). As regards the traffic signals, he stressed that in the performance of the specific conative function which they assume they are
superior to language because of their simplicity and exactness (cf. Die Axioma-tik der Sprachwissenschaften, p.  115), but he did not always resist the temptation to interpret the appeal as an indirect, or mediated, type of reference: "Das durch ein rotes Signallicht Vertretene ist letzten Endes irgendein unter anderen Wahrnehmungsverh'a'ltnissen vielleich, vielleicht auch nich sichtbares Durchfahrtshindernis, Uber dessen Vorliegen der Signalgeber entscheidet" (ibid., p.  116). In fact, the obstacle is the red light itself and the signal giver can be a soulless machine which turns on the red light in regular intervals irrespective of the presence or absence of any factual obstacle. The obstacle is "in the last analysis" just a meaning conveyed by a feature of the sign (its color), so that the reckless driver who decides not to respond need not necessarily run into any accident.
20. "Under the emotive function are included both the expression of emotions, attitudes, moods, intentions, etc., in the speaker, and their communication, i.e., their evocation in the listener. As there is no convenient verb to cover both expression and evocation, we shall in what follows often use the term 'evoke' to cover both sides of the emotive function, there being no risk of misunderstanding" (C. K. Ogden and I. A. Richards, The Meaning of Meaning, New York: Harcourt, Brace & Co.  (paperback), n. d., p. 149.
21. Cf. E. H. Gombrich, "Action and Expression in Western Art", in: Robert A. Hinde, ed., Non-Verbal Communication, London: Cambridge University Press, 1972, p. 393.
22. Cf. Roman Jakobson, "Linguistics and Poetics", in: Thomas A. Sebeok, Style in Language. New York-London: The Technology Press of MIT and John Wiley,  1960, pp. 354-355.
23. Cf. Emile Benveniste, "De la subjectivite dans le langage",
Problemes de linquistique qenerale, I, Paris: Galliraard, 1966, pp. 261 and 264.- In the development of Benveniste's thought this article (first published in 1958) as it were anticipated his theory of the twin oppositions sign/discourse and semiotic/semantic, which he was to formulate from 1962 on. However, the particular analysis referred to in the text is largely independent of that theory.
24. Cf. Mikhail Bakhtine, L'oeuvre de Francois Rabelais et la culture populaire au Moyen Age et sous la Renaissance, Paris: Gallimard, 1970, p. 193.
25. The variety of conative meanings, which in reality is much greater than this far from exhaustive list indicates, casts serious doubt in Ingar-den's attempt to substitute two separate functions, namely the communication /j<ommunikations- oder Mitteilungsfunktion/ and the influencing function /Beeinflussungsfunktion/ for the conative (Roman Ingarden, Das literarische Kunstwerkf Tubingen: Niemeyer, 19602, Annex: "Von den Funktionen der Sprache im Theaterschauspiel"). This could severely restrict the range of the meanings pertaining to the perceiver or else necessitate the more or less arbitrary subsumption under the one or the other function of many meanings which really fit neither. Furthermore, Ingarden gives no reason, whether logical or noetic, why the semiotic function pertaining to the perceiver should be split in two, especially when he maintains the respective unity of both the expressive and the referential function; Richards rightly points out that the referential function has two aspects, "covering both what-is-talked-of and what-is-being-said-about-it" ("Factors and Functions in Linguistics?, op. cit.f p. 5).
26. Cf. for example Ernst Kris, "Approaches to Art", Psychoanalytic Explorations in Art. New York: Shocken Books, 19642 (paperback), pp. 54-56 (with bibliography in footnote 78); E. H. Gombrich, "Action and Expression in
57
Western Art", op. cit., pp. 388-390, and "The Mask and the Face: The Perception of Physiognomic Likeness in Life and Art", in: E. H. Gombrich, Julian Höchberg and Max Black, Art, Perception and Reality. Baltimore: Johns Hopkins Press, 1972, p. 35.
27. Cf. Vilém Mathesius, "Jazykozpytné poznámky k řečnické výstavba souvislého výkladu" /^Linguistic Notes about the Rhetorical Construction of a Continuous Discourse/, Slovo a slovesnost, IX (1943), pp. 118-119.
28. Cf. E. H. Gombrich, Art and Illusion, Londoni Phaidon, 19724, pp. 154-244; "Action and Expression in Western Art", op. cit., p. 390; "The Visual Image", op. cit., p. 89; "The Mask and the Face", op. cit.T p. 17; Sense of Order. London: Phaidon, 1979, pp. 107 and 143.
29. Rudolf Wittkower, Sculpture. London: Allen Lane, 1977, p. 51.
30. Cf. Howard Hibbard, Bernini, Harmondsworth: Penguin, 1965, p.
90.
31. Cf. Meyer Schapiro, Paul Cezanne. New York: Abrams, n. d., pp.
14-15.
32. E. H. Gombrich, "Standards of Truth: The Arrested Image and the Moving Eye", Critical Inquiry. VIl/2 (Winter 1980), p. 248.
33. On the two figures and their conative meaning cf. Heinrich Wb'lf-flin, "Das Problem der Umkehrung in Raffaels Teppichkartons", Gedanken zur Kunstgeschichter .Basel, 1941, p. 95.
34. Cf. Rudolf Wittkower, Art and Architecture in Italy 1600-1750,.
A. ~
Hrmondsworth: Penguin, 1973J (paperback), pp. 48-49.
35. Cf. my "Comparative Semiotics of Art", in: Wendy Steiner, ed., Image and Code. Ann Arbor: Michigan Studiejin Humanities, 1981, pp. 114-115.
36. Cf. Sidney Geist, Brancusi/The Kiss, New York, Harper and Row,'?T«
pp. 66.
37. Cf. Roman Jakobson, "Shifters, Verbal Categories, and the Russiar Verb", op. cit., p. 131} "Linguistics and Poetics", op. cit., p. 356; "Metalanguage as a Linguistic Problem", KUlönlenyomat a Nyelvtudományi Közlemenyak 76 (1976), pp. 349-352.
38. Cf. Roman Jakobson, "Linguistics and Poetics", op. cit., 357-358.
39. Ibid., p. 356.
40. Cf. Keir Elam, The Semiotics of Theater and Drama. London: Me-thuen, 1980, p. 154.
41. Cf. Serge Karcevskij, "Du dualisme asymétrique du signe linguis-tique", Travaux du Cercle linquistique de Prague. I, 1929.
42. Cf. my "Comparative Semiotics of Art", op. cit.f pp. 123-124.
43. Emile Benveniste, "Sémiologie de la langue", Problemes de linquistique generale, II, Paris: Gallimard, 1974, pp. 61-62.
44. Cf. E. H. Gombrich, "The Visual Image", op. cit.f p. 86.
45. Cf. Deryck Cooke, The Language of Music. London: Oxford University Press, 1959 (paperback), pp. 3-4.
46. Cf. Robert Frances, Michel Imberty and Arlette Zenatti, "Le do-maine musical", in: Robert Frances, ed., Psychologie de l'art et de 1'esthé-tic|ue_, Paris: Presses universitaires de France, 1979, p. 175.
47. Cf. Roman Jakobson, "On the Verbal Art of William Blake and Other Poet-Painters", Linguistic Enquiry,  1970/1.
48. Cf. Roman Jakobson, "Linguistic^ and Poetics", op. cit.P pp.
355-356.
49. Cf. Petr Bogatyrev, Lidové divadlo České a slovenské /Czech and Slovak Folk Theater/, Prague: Borový, 1940, pp. 95-96; Richard Southern, The Seven Ages of the Theater, London: Faber, 1968 (paperback), p. 92.
ST3
50. Cf. Rudolf Wittkower, Architectural Principles in the Age of Humanism. London: Academy Editions,  1973^,   (paperback), p. 22.
51. Cf. Bronislaw Malinowski, "The Problem of Meaning in Primitive Languages", Supplement I in: C. K. Ogden and I. A. Richards, op. cit.. pp. 313-316.
52. Cf. Randolph Quirk, The Use of English,, London: Longmans, 19682,
p. 63.
53. Cf. Roman Jakobson, "Linguistics and Poetics", op. cit. .pp. «S^y-.Sfi
54. Cf. Bronislaw Malinowski, op. cit.f p. 314.
55. Cf. Jan Mukarovsky, "L'art comme fait semiologique", Actes du Huitieme Congres international de Philosophie a Prague 2-7 septembre 1934, Prague: Comite d'organisation du Congres, 1936, pp. 1065-1072.
56. Cf. ibid.. p. 1085.
57. Cf. Jan Mukafovsky, "La denomination poetique et la fonction es-thetique de la langue", Actes du Quatrieme Congres international des linquistes tenu a Copenhague du 27 aout au Ier septembre 1936P Copenhagen: E. Munksgaard, 1936, pp. 98-104.
I
58. "Denn 'der sprachliche Ausdruck1 und 'der sprachlice Appell' sind Teilgegenst'ande der ganzen Sprachforschung, die nicht nur eigene BemUhungen und Methoden erfordern, sondern auch eigene Strukturen aufweisen. Die Lyrik kurz gesprochen und die Rhetorik haben jede etwas Eignes an sich, was sie unter sich und von der - sagen wir einmal Epik, um nicht aus dem Konzept zu fallen - unterscheidet und noch auffalender unterschieden sind ihre Strukturgesetze natürlich von dem Strukturgesetz der wissenschaftlichen Darstellung" (Die Axiomatik der Sprachwissenschaften, pp. 103-104). In the Sprachtheorie (pp. 32-33), he -somewhat precipitately - added drama to the narrative as an example of a lan-
guage structure proper to the referential function. In Jakobson's view, "The peculiarities of diverse poetic genres imply a differently ranked participation of the other verbal functions along with the dominant poetic function. Epic poetry, focused on the third person, strongly involves the referential function of language; the lyric, oriented toward the first person, is intimately linked with the emotive function; poetry of the second person is imbued with the conative function and is either supplicatory or exhortative, depending on whether the first person is subordinated to the second one or the second to the first" ("Linguistics and Poetics", op. cit.f p. 357).
59. Cf. Jan Mukarovský, "On Poetic Language", The Word and Verbal Art New Haven; Yale University Press,  1977, p.5.
60. Cf. Jan Mukarovský, "The Individual and Literary Development", ibid.r p. 163.
61. Cf. my "Contribution to the Semiotics of Acting", op. cit.r pp.
578-592.
62. Cf. Jan Mukarovský, "The Individual and Literary Development", op. cit., p. 162.
63. Cf. Jan Mukarovský, "The Place of the Esthetic Function Among the Other Functions", Structure, Sign, and Function. New Haven: Yale University Press, 1978, pp. 42-43.
64. Cf. Jan Mukarovský, "The Esthetics of Language", in: Paul L. Garvin, ed.f A Prague School Reader on Esthetics. Literary Structure, and Style. Washington, D. C.J Georgetown University Press, 1964^f p. 33.
65. Cf. my "Jan Mukarovsky's Structural Poetics and Esthetics", Poetics Today. Il/lb (1980/81), p. 141.
66. Cf. Jan Mukarovský, "Umění" /Art/, Studie z estetiky /Studies in Esthetics/, Prague: Odeon, 1966, pp. 127-128.
55
67. Jan Mukarovsky, "La denomination poetique et la fonction esthe-tique de la langue", op. cit. Similarly, Jakobson simply ranges the poetic function among the six semiotic functions of the linguistic sign which he examines, like the referential, the emotive, the conative, the metalingual and the phatic    (cf. Roman Jakobson, "Linguistics and Poetics", op. cit. )
68. Mukarovsky himself dealt with this ambiguity in a different context (cf. Jan Mukarovsky, "The Concept of the Whole in the Theory of Art", Structure. Sign, and Function, pp. 79-81).
69. Cf. Madeleine Mathiot and Paul L. Garvin, "The Functions of Language: A Sociocultural View", Anthropological Quarterly. 1975/3.
70. Karl BUhler, Die Axiomatik der Sprachwissenschaften, p. 112.
71. C. K. Ogden and I. A. Richards, op. cit.r p. 239.
72. Cf. Edmund Husserl, "Zur Phänomenologie des inneren Zeitbewusst-seins", ed. Martin Heidegger, Jahrbuch fUr Philosophie und phänomenologische Forschung. IX (1928), pp. 285-287.
73. I am inclined to think, though not yet ready to affirm, that in all semiotic systems the association of the signatum to the signans by similarity requires some support from the opposite principle of association by contiguity. In another article I have tried to show that in the mimetic picture similarity cannot work witout contiguity contributing as a selecting and classifying factor (see my "Some Aspects of the Pictorial Sigh", in: Ladislav Matejka and Irwin R. Titunic, Semiotics of Art, Cambridge, Mass.: The MIT Press, 1976, pp. 250-261). Gombrich, who has probably done more than any other scholar to expose the loose and naive way in which the concept of resemblance has been used in art theory and history, argues nonetheless against the extreme relativists and conventionalists among the semioticians that in the visual sphere the
56
association by similarity is deeply rooted in basic responses to the environment, not only in man but even in animals (cf. E. H. Gombrich,"Image and Code: Scope and Limits of Conventionalism in Pictorial Representation", in: Wendy Steiner, ed., Image and Code, pp. 20-22). That is most likely also true of the association by similary in auditory signs, including music. It does not tend to indicate that the sign can refer to something by similarity alone.
74. Cf. Nicolas Ruwet, "Avant-Propos", Langaqef Musigue, Poesie, Paris: Seuil, 1972, p. 14.
75. Benjamin Lee Whorf, "Language, Mind, and Reality", Language. Thought.and Reality, Cambridge, Mass.: The MIT Press, 1963 (paperback), p. £c '
76. Emile Benveniste, "Semiologie de la langue", op. cit., p. bJ.
77. Paul Valery, Eupalinos. L'ame de la danse. Dialogue de l'arbre. Paris: Gallimard, 1944, pp. 50-51.
78. Vitruvius, The Ten Books on Architecture, New York: Dover, 1960^
pp. 14-15.
79. Cf. E. H. Gombrich, Art and Illusion, pp. 316-317.
80. Cf. Ernst Kris, op. cit., pp. 54-63.