1.1.

It has frequently been noted that music provides something of a problem for the student of semiotics, in that while it is open to analysis on what is often, if somewhat rashly, called the ‘syntactic’ level,Footnote ¹ it seemingly eludes all objective treatment on the semantic one. (Linguistic metaphors such as ‘syntax’ should, however, be used with a certain caution when talking about music: linguistic syntax, while it may be analysed independently of the semantic level, is nevertheless determined by it; whereas the formal structure of music, while potentially capable of communicating something beyond itself, is basically an end in itself. Furthermore, if this metaphor is maintained, the phonetic level of linguistic analysis becomes a syntactic system when viewed from the musical point of view: an unnecessary confusion.)

Naturally, for those who adopt a formalist approach to musical communication the problem does not exist: for them, music conveys no more than the relationships between its formal elements, so that an analysis of the way in which each element defines itself in relation to the other components of a set of morphological probabilities (in other words, a musical style) will be sufficient demonstration of the communicative process in question. Such an approach would seemingly carry music outside the realm of semiotics, strictu senso, unless one contends, along with [Umberto] Eco,Footnote ² that a sign denotes its position within a conventionally established system of cultural units, its meaning being supplied by a complex connotative process, and that therefore a musical element may be considered a sign, because it also denotes its position within a pre-established system.Footnote ³

Unfortunately this idea is open to the same objection as was that of musical syntax: a code is constituted by the assembling of a series of differentiated sign-vehicles that correspond to a pre-existing series of differentiated cultural units. So that if the terms ‘sign’ and ‘code’ are to be regarded as being equally applicable to communicative systems involving one (formally determined) level of articulation (music, abstract visual arts), and those (semantically determined) involving two (language, iconic signs etc.), they must be constantly qualified in order to indicate which of the two types of system is meant. It would seem less terminologically wasteful to limit ‘code’ and ‘sign’ to biplanar systems, and to evolve other terms to indicate the purely formal processes at present under consideration.

Furthermore, the communicative processes employed by much contemporary music would seem to undermine the whole basis of the above approach. The only set of aurally appreciable probabilities in operation within such processes is a negative one, in that elements that may give rise to false expectations (such as tonal progressions and regular rhythms) will tend to be excluded. In such circumstances, an account of formal communication might be given by applying the basic principles of Gestalt psychology to each parameter of sound-time (an approach already applied to the visual arts by R[udolf] Arnheim in his Art and Visual Perception).

1.2.

However, as soon as it is acknowledged that music may also potentially communicate emotive elements above and beyond those aroused by the appreciation of formal inter-relationships, one reaches a methodological impasse, in as much as it would seem impossible to establish by metalinguistic commentary what is communicated. The statistical methods of enquiry developed by R[obert] Francès that are advocated by J[ean]-J[acques] NattiezFootnote ⁴ as a solution to this problem might well serve to establish the basic affective codes applied to music within an explicitly defined sector of society, but since they must inevitably be based upon verbal communication, they are likely to be confounded, upon a more detailed level, by the metalinguistic impasse that they set out to solve.

There are however two aspects of musical communication, if subsidiary ones, that bypass this difficulty, and consequently lend themselves to a comparatively straightforward semiotic analysis: these being the use of iconic elements and the exploitation for communicative purposes of such social and historical connotations as will naturally be attached to any given musical style, or indeed individual composition.

2.1.

Iconic signs – that is to say configurations that reproduce such pertinent features of an original perceptual model as will serve for its recognition according to conventionally established codesFootnote ⁵ – can be effected in sound-time as they can in space. But owing to the differing characteristics of visual and aural perception, iconic communication in sound-time has had a very modest development in comparison with that enjoyed in almost every culture by the spatial icon. Aural experience is only concerned with a certain range of movements within the perceptor's environment – those that create vibrations lying within the auditory range – and although it may be selectively concentrated upon a particular phenomenon, its capacity for blocking the perception of other environmental sounds is limited. Visual experience, on the other hand, is capable of great selectivity when singling out the componential elements of the environment, whether they be at rest or in motion, and consequently establishes itself as the decisive factor in determining a conception of ‘reality’.

This difference is reflected by the strong iconic tradition of the visual arts, where the reductive and schematic representation of a perceptual state reflects the departicularizing process essential to the perception and conceptual organization of visual reality; whereas iconic representation in music tends to be of peripheral importance, everyday aural experience not apparently being of such cognitive significance as to demand extended iconic commentary. Consequently, as might be expected, aural iconic codes have had a very limited development, and it is only in rare instances that they have moved towards a manifestly stylized relationship between sign-vehicle and meaning.

2.2.

A further difference between the visual icon and the aural one is that while the former tends to concern itself almost exclusively with the representation of other visual elements (the synaesthetic iconic vocabulary of the comic strip presenting a notable exception), the latter is as much concerned with synaesthetic elements as it is with aural ones. In fact, the instance[s] in which conventionalized codifications of aural phenomena have established themselves are few. As examples one might take the trilling flute figures that so frequently represent bird-song (Wagner's Siegfried, Beethoven's Sixth Symphony); the repeated arpeggio figures (developed from the purely formal device of the Alberti bass) that represent flowing water – but note that they are usually appropriately undulating, thus synaesthetically reproducing the visual form of the wave – (Schubert's Die schöne Müllerin cycle, Beethoven's Sixth Symphony); and the outbursts of kettle-drumming in representations of thunderstorms – though again, much depends upon the synaesthetic suggestion of a downpour (since it is not the sound of the rain that is reproduced, but its direction), and of flashes of lightning – (Rossini's II barbiere, Britten's Peter Grimes, and, once again, Beethoven's Sixth Symphony). Such cases apart, it is only when the perceptual model of a simple sound-time structure is reproduced with great fidelity (the cuckoo from Delius's On Hearing the First Cuckoo in Spring, or the sheep from Strauss's Don Quixote) that it proves unnecessary to resort to the metalinguistic intervention of words, whether sung or printed in the programme notes, or, in the case of opera and cinematic background music, action in order to clarify what is being represented. Consequently the tendency in programme music is not so much to represent the aural landscape as to suggest certain emotional atmospheres or moods, whose precise dramatic significance is revealed by some metalinguistic commentary. (As to whether and how music does successfully convey mood or emotion, see section 4.)

The synaesthetic elements that play an at any rate equally important part in iconic musical communication are almost all based upon a code that has been firmly established in European musical culture for at least four hundred years: that by which ‘high’ and ‘low’ pitches correspond to the vertical spatial dimension (note that the correspondence is preserved verbally in English – but not, for instance, in Italian, where a high note is described not as alta, but as acuta). Not that the resilience of this codification is particularly surprising in view of the fact that the inverse relationship has provided the graphic basis for Western pitch notation from Gregorian chant onwards.

Thus a visual, or in certain cases tactile, perception of movement will be represented by an iconic rendition of its rhythm, plus a depiction of its extension and form within the vertical dimension in accordance with the above-mentioned code. (See, for example, the depiction of flowing water discussed above.) Again, pitch structures are unlikely to be perceived as having a synaesthetic iconic function unless some metalinguistic commentary suggests as much. Consequently this form of representation tends to occur, as might be expected, in association with a sung text, where ‘high’, ‘low’, and all verbs indicating some type of vertical movement provide a seemingly irresistible invitation to word-painting. The examples are innumerable, but the ‘Et resurrexit’ section from almost any post-1500 setting of the Mass will provide an immediate illustration.

Another far weaker (and less important) set of synaesthetic codifications that may be noted in passing are those connected with luminosity. Lower pitches are associated with darkness, higher ones with light (as, for instance, in the melodic line of ‘The people that walked in darkness’ from Handel's Messiah). Similarly lower dynamics are associated with darkness, and higher ones with light; more interestingly, instrumental timbres that lack the higher harmonics are associated with darkness, and those that have them with light (this latter code being so well established that one quite habitually talks of darker and brighter tone-colours). All these, however, are basically examples of negative/positive associations normally employed to suggest emotional states, and only incidentally fulfilling an iconic role.

2.3.

The choice of iconic musical material, whether aural or synaesthetic in its subject matter, is limited by two factors:

1. (a) Except in the case of the rare exceptions at present under discussion, music is not concerned with the dialectic between form en soi and form as sign-vehicle that is fundamental to the literary and representative visual arts. Consequently even when iconic elements are introduced, they are even more subject to the formal considerations that are music's primary concern than would be the case, say, in a representative painting. Thus the choice of iconic material will tend to be limited to such elements as are well adapted to repetition and formal development: flowing water, galloping horses, etc. When instead the music seeks to illustrate some passing and unrepeatable event – such as is, perforce, the guillotine at the end of the ‘Marche au supplice’ from Berlioz's Symphonie fantastique – the result tends to be both formally and iconically incoherent unless one is well acquainted with the programme notes. The same effect becomes immediately apparent in any over-elaborate attempt at word-painting in vocal music: if the iconic elements are not strictly subordinated to the formal exigencies of the structure, then the effectiveness of both communicative levels is quickly lost.

2. (b) Since, as was noted above, the primary concern of programmatic music is not so much iconic representation as the creation of a certain emotional atmosphere, or mood, such iconic elements as are introduced will be selected according to their capacity for creating such an atmosphere: musical seascapes will tend to ignore the scrunch of pebbles underfoot, or the squawk of seagulls.

Granted the further limitations imposed by these two factors upon a process that is in any case of only minor musical significance, the instances in which a sound-time structure fulfils a function analogous to that of the iconic sign in visual art are comparatively rare. Indeed, most musical icons are based upon such unstable and limited codes that it is only when some metalinguistic commentary indicates as much that their presence is suspected – a circumstance which, significantly, only occurs in the representative visual arts when they border upon the abstract.

3.1.

There is, however, one area, if a somewhat specialized one, within which metalinguistic intervention becomes unnecessary, and the limitations that beset other forms of iconic sign-function in music are greatly reduced. This is, in fact, the other sector mentioned in 1.2. as lending itself to a comparatively straightforward semiotic investigation: that of the exploitation of the socio-historical connotations associated with a certain style or piece, and thus of quotation, parody, and pastiche. Naturally, any piece of music will carry connotations relating to the historical period in which it was created and the use to which it was, and is, put. But in respect to such connotations the music acts merely as an index, and it is only when a deliberate quote or imitation is made, recalling the connotations to mind in a context that would not otherwise do so, that it takes on the function of a sign.

The three processes in question are unusual in that the iconic signs by which they are effected are realized in the same medium as were the elements that they represent, but otherwise they follow the normal pattern of iconic communication, reproducing such pertinent elements as will serve to indicate the perceptual model in question. (Indeed, the possibility of purely stylistic reference, as in the case of pastiche, clearly illustrates that the iconic sign indicates not a referent, but a perceptual model,Footnote ⁶ since it is only as such that style can exist.) In the case of parody, the reproduction of these pertinent features is accomplished in an ironically hyperbolical fashion: either, in its cruder form, by isolating the most obvious principal features from the hierarchy of subordinate features that support and lend variety to them, and insisting upon them to the point of creating an absurdly redundant communicative situation; or, in a more subtle fashion, by conserving all the elements within the hierarchy of pertinent features, but calling upon the subordinate ones to play an improbably major role, in failing to sustain which they give rise to insidious doubts as to the validity of the hierarchy itself, while obscuring the fact that it has been more or less maliciously turned upside down.

Naturally, a careful study of parody and pastiche will throw interesting light upon the perceptual codes involved in the appreciation of stylistic differences. In this respect an analysis of middle-period Stravinsky, or Poulenc, or Walton in his Façade mood might prove revealing.

3.2.

Quotation, however, raises a more complex problem. When it is exact in every detail (which is to say, in this case, when it conforms to every detail of the notation of the original – since the other, non-notated elements are free variants in any case), then we are evidently confronted not by an iconic sign, but by an intrinsically coded act,Footnote ⁷ in as much as one part of a musical work is being used to signify the cultural unit formed by the whole of that work and its attendant socio-historical connotations. But to what extent is an intrinsically coded act involved in an inexact quotation? At this point, Eco's example of the drawing of a red flag,Footnote ⁸ where the ‘redness’ of the figure is regarded as an intrinsically coded act, and its rectangular shape as iconic, suggests a tempting musical parallel – rendered all the more plausible by the fact that, until the beginning of this century, rhythm and pitch were undoubtedly regarded as the two basic structural parameters, while timbre was relegated to the level of tone-colour, as noted in 2.2. Thus an analysis of the trumpet-calls from the Agnus Dei of the Missa solemnis would suggest that while the melodic and rhythmic formation constitutes an iconic sign (referring to the cultural unit ‘trumpet-call’), the timbre of the trumpet is an intrinsically coded act. This is confirmed if I now play the same theme on an oboe: the iconic representation of a trumpet-call remains intact, although no longer supported by the intrinsically coded act of trumpet timbre. Similarly, I can play any melody that I like on a trumpet, but unless it has the appropriate melodic and rhythmic features, it will not be perceived as referring to the cultural unit ‘trumpet-call’. It would therefore seem that, in this case, an intrinsically coded act is brought about when an element is included which, while not being a pertinent feature for the purposes of iconic reproduction, is nevertheless closely associated with the represented phenomenon, and thus aids the de-coding of the iconic sign. But if I attempt to apply this model of the functioning of a musical intrinsically coded act to the case of an inexact quotation, its limitations immediately come to light: when making a quotation exact in every respect except that of timbre, the elements that remain common to both the original and its reproduction, and that I might thus be tempted, on the model of the exact quotation, to label as intrinsically coded acts, are precisely those that are iconically pertinent.

The solution to this apparently paradoxical situation is to be found in the differing nature of the cultural units that are evoked by the processes of quotation and iconic stylistic reference (i.e. pastiche). In the case of quotation, the cultural unit evoked is characterized by all the pertinent features of the unique original, so that the quotation can only function as an intrinsically coded act when it is exact – it otherwise immediately becomes an iconic sign. But in the case of pastiche (this term being understood to include the representation of musical ‘types’ determined by function – e.g. march, waltz, trumpet-call etc. – as well as historically determined styles), the cultural unit evoked is characterized by a restricted group of pertinent features common to a large number of different musical structures, but of themselves insufficient to indicate any one such structure specifically. One can thus seek to establish, by commutation, what features are pertinent (and thus necessary for the constitution of an iconic sign), and what features are not (and are thus potentially capable of acting as intrinsically coded acts when included); though inevitably certain borderline cases may prove difficult to classify. The basic pertinent elements of the waltz, for instance, are its rhythmic structure and the characteristic formal structure of the accompaniment. Consequently Schoenberg and Berg were able to write Walzer and Ländler, whose melodic and harmonic content would have caused utter consternation on any Viennese dance-floor, but which were nevertheless recognizable to the concertgoer as iconic references because they preserved these essential pertinent elements. Had the formal structure of the accompaniment been realized in terms of the restricted circuit of common chords, based on the tonic, dominant, and subdominant, that was characteristic of the waltz (as was to be the case in various of the waltz references in Strauss's Rosenkavalier), then this particular feature could be justifiably classed as an intrinsically coded act whose presence aided the de-codification of the iconic sign in question, but was in no way essential to such a de-codification.

It is worth noting in passing that if a quotation is to function as such, the receiver must already be acquainted with the piece in question. If not, he will interpret it as a stylistic reference, i.e. pastiche. Consequently:

1. (a) if he hears a trumpet playing a fragment that he recognizes as coming from the trumpet-calls in the ‘dona nobis pacem’ section of the Missa solemnis, he will receive it as an intrinsically coded act;

2. (b) if he hears the same fragment played on an oboe, and recognizes it, he will receive it as an iconic sign;

3. (c) if he hears a trumpet playing it, but does not recognize it, he will receive the formal structure as an iconic sign for the cultural unit ‘trumpet-call’, and the trumpet timbre as an intrinsically coded act;

4. (d) if he hears an oboe playing it, without recognizing it, he will view it, simply, as an iconic sign for the cultural unit ‘trumpet-call’.

4.1.

As has already been noted, although the two aspects of musical communication dealt with in sections 2 and 3 (plus certain codified forms of emotive communication, with reference to which see 4.2.) are the only ones that lend themselves to an analysis along the lines already established by semiotics, they only occupy a very limited sector of the total field. However, iconic elements of a less easily demonstrable sort also play their part in the hypotheses advanced as to the nature of musical communication in general.

It was observed in section 1.1. that, for those adopting the formalist approach (most notably Hanslick, Stravinsky, De Schloezer, Langer, and to a certain extent MeyerFootnote ⁹), music is a purely formal mode of communication with no reference beyond itself, so that within it ‘the signified is immanent in the signifier, and the content in the form, to such an extent that, strictly speaking, music does not have a sense, but is a sense’.Footnote ¹⁰ Despite the appealing coherence of such an attitude, it tends to give deceptively little weight to the fact that a large number of composers have deliberately used music to express an emotive or dramatic situation (albeit with the well-nigh invariable accompaniment of some form of verbal, kinesic, or even proxemic commentary), and that their audiences would seem to have appreciated the result in precisely the manner intended. De Schloezer, for instance, remarks further on in the same study that ‘the psychological sense that we come to attribute to it [music] is only an addition due to a complex of conventions and associations which habit has fixed and rendered unconscious’.Footnote ¹¹ Whether or not this is the case, it does not alter the fact that this type of musical experience is of undoubted importance for the majority of listeners (and was indeed fundamental to the Romantic conception of music).

It is not by chance that T[heodor] W. Adorno, in the admittedly arbitrary, but nonetheless highly penetrating analysis of types of musical behaviour that introduces his Einleitung in die Musiksoziologie (1962), distinguishes six types of listener of which only two really concern themselves with the formal qualities of the music (this approach evidently being for him, as for De Schloezer, the only truly viable one); furthermore, both of these groups represent very small minorities, whereas the three types of listener who basically regard music as a type of emotive stimulus belong to notably larger ones. This may at any rate serve to indicate that any study of musical communication that seeks to preserve a sociological orientation, rather than merely attempting a definition of the most desirable mode of musical perception, must perforce try to give some account of the way in which music successfully creates an emotive atmosphere.

4.2.

Music may be experienced as communicating an affective state, but, as has been noted above, it proves impossible to give an exact verbal definition of the affective state in question except in certain specialized cases. This difficulty is made all too evident by Deryck Cooke's brave but doomed attempt to prove the contrary in The Language of Music (1959) where, within the framework of the Western tonal tradition, certain basic melodic formulae are shown to recur in association with certain emotive situations. Naturally, most of the examples are from vocal music, or programme music, where the words make clear what the emotive situation is; when purely instrumental pieces are cited, they are well basted with adjectives by the author in order to fit in with his original suppositions, the inference being that when such formulae recur without words, they will ‘mean’ the same thing. Unfortunately, apart from attempting to ensure a watertight argument by simply ignoring the examples that do not conform with his thesis, the author does not give adequate consideration to the respective levels at which music and words communicate. The words undoubtedly serve to channel and focus the emotive impression made by the music, but were not the affective import of the music alone so ambiguous (at least from the point of view of metalinguistic description) as to allow a precise account of what is communicated only when some unambiguous verbal formation is associated with it, the problem which Cooke sets out to resolve would not exist in the first place, unless (and the possibility is not entirely to be discounted) music represents a seemingly unique example of a conventionally established code whose meanings are beyond the metalinguistic grasp of that other conventionally established code, language.

On the other hand, it is possible to codify certain relationships between a musical formation and an affective state that have been conventionally established amongst a specific cultural group during a specific historical period. Bach's music, for instance, has attracted several notable studies along these lines.Footnote ¹² Furthermore, the statistical studies suggested by FrancèsFootnote ¹³ could do much to clarify the present situation in this respect. However, the question immediately arises as to whether, in such cases, the relation between musical formation and signified affect is arbitrary, as in the relation between linguistic sign-vehicle and meaning, or whether, as the considerations to be put forward in 4.3. would suggest, it involves an iconic element that has been reinforced by an arbitrary and conventionalized codification.

4.3.

Since, as has been seen, no general system of codification is available for emotive communication in music, the only possibility open to verbal commentary is that of advancing hypotheses as to the processes involved in such communication. Perhaps the most objective material available for consideration in this respect is the research that has been done upon brain processes in relation to musical experience, a summary of which may be found in T[erence] McLaughlin's Music and Communication (1970). While this is not the place for extended physiological descriptions, the author's basic argument may be summarized as follows: granted that the messages from all sense organs are converted into patterns of electrical impulses in the brain, where they are checked against other patterns held in the memory, and granted further that the patterns of impulses are only differentiated from one another as sight or sound etc. because they are received at a certain part of the cortex, and not at another (a clear illustration of this being the synaesthetic impression that results from the misdirection of one of these patterns: a loud bang being experienced visually as a flash of light, for instance), then it is perfectly possible that a musical pattern may, while being processed for recognition in relation to the innumerable other patterns stored in the brain, activate a number of these – even though of diverse origins – and synthesize them. Certainly, it would seem a reasonable proposition that an abstract formation which, while possessing certain perceptual pertinent features common to a wide variety of conscious and unconscious experience-patterns, does not possess the full hierarchy of such pertinent features that would cause it to be related to any one of them, should thus be able to function as a sort of synthetic ‘Ur-gestalt’. But until a great deal more work has been done upon the processing of sensory data in the brain, it will be impossible to attempt any objective verification of such a hypothesis. (It may be noted in passing that a related idea, formulated from a philosophical point of view, may be found in the writings of two such determined formalists as Hanslick and Langer, both of them acknowledging that music can reproduce the dynamic structure of emotional processes.)

Thus if an iconic sign is a structure in either space or sound that concentrates within itself the pertinent relational features of a perceptual model, itself derived from a reductive synthesis of various differing examples of the entity in question, then an abstract form standing in close relation to the basic patterns to which a wide variety of experiences have been reduced in the brain, and thus effecting a synthesis amongst them, may be regarded as an ‘unconscious icon’. Naturally, certain formations will lend themselves far more readily to such a process than will others, and furthermore certain of these will tend to synthesize a far more closely-knit series of associations than will others, thus communicating a more specific experiential complex. So that in the light of such a hypothesis one may establish the following triangular relationship between the functions that may be assumed by either visual or auditory structures:

Of these three elements, conscious iconicity is the only one that may be analysed along the lines established by semiotics, although certain aspects of abstract form may be analysed in terms of a similar methodology (cf. 1.1.), as long as they relate to a conventionally established series of probabilities such as the tonal system. Subconscious iconicity, by its very nature, only allows the formation of hypotheses as to the nature of the communicative process involved. In the case of conscious iconicity, all their elements may, and probably will be, co-present. In all other cases, the abstract form may or may not tend towards iconicity, conscious or unconscious, the type of communication involved being determined as much by the listener's approach as by the composer's. Thus the listener intent upon finding emotional stimulus (an attitude probably representative of the majority of cases) will tend to be unreceptive to musical structures created as purely formal interplay, and will respond only in the presence of what he regards as ‘expressive’ elements. At the other extreme, the arch-formalist may find himself merely repelled by a structure that sacrifices richness of formal interplay to emotional effect.

At this point, a clear theoretical distinction must be made between emotion caused by the appreciation of formal relationships, and such emotive complexes as may be communicated by these formal relationships in their function as unconscious icons (though whether the distinction can be maintained with such clarity on the level of ontological analysis is a matter for question). It would seem, ironically, that only in the former case – that of an immediate and spontaneous reaction to a formal element – can one accurately talk of ‘emotion’: as regards the latter case, L[eonard] B. Meyer quotes the psychologist H[arry] P[orter] Weld's observation that the affective responses that arise from listening to music are ‘to be characterized rather as moods than as emotions in the ordinary sense of the term […]. The emotion is temporary and evanescent; the mood is relatively permanent and stable’, adding that ‘most of the supposed studies of emotion in music are actually concerned with mood and association’.Footnote ¹⁴ Certainly the essential characteristic of this latter type of musical experience would seem to be that of prolonged concentration upon an affective complex that provides a synthesis and summation of past emotional (in Weld's sense) experience. But such excursions into the psychology of aesthetic experience, however suggestive, are not the concern of the semioticist …

4.4.

If therefore a hypothesis such as the one outlined above is accepted, all forms of musical communication in which the formal structure conveys to the listener something above and beyond itself would seem to be based upon iconic processes: explicitly in the case of the iconic sign, instrumentally when seeking to manipulate the social or historical connotations of a certain style or piece, and unconsciously in the case of emotive communication. These categories may, furthermore, be applied with equal validity to spatial structures (although the functional arts introduce further levels of analysis). So that while the formal level of musical communication may, in most cases, be analysed in terms analogous to those of linguistic syntax, the various semantic levels that may also be potentially brought into play would seem to be best approached in terms of the iconic process.