Musicology and the
semiotics of popular music

by Philip Tagg

This text was originally published as an article in Semiotica 66-1/3 (1987), pp. 279-298. Copyright P Tagg (author, Liverpool) and Mouton de Gruyter (publishers, Amsterdam). This text is basically the same as that sent unedited to Mouton de Gruyter in 1986, except that bibliographical references have been corrected. Footnote numbers are not the same as in the Semiotica version. Starts of pages in the Semiotica version are shown in small braces.

View correct PDF file
Why did I land here?


{279} This article aims to discuss some theories and to propose some methods which may contribute towards the establishment of a semiotics of popular music. Matters treated here will not be discussed within the general framework of (apparently rivalling) semiotic theories, since generalisations would be premature at this stage of research when empirical work on the reception of popular music, and analytical work on the accurate description of its structures is still far from complete. Nor is there room to present examples of popular music analysis, since meaningful treatment of even the shortest piece would require more pages than are at my disposal. However, the most important methodological aspects of semiotically biased popular music analysis are presented below in the form of a theoretical account and readers are referred to other publications to obtain a fuller account of method application (Alencar Pinto 1986; Tagg 1979, 1981; Tagg & Clarida 2000).

The problem

A semiotic musicology of the mass media: why?

Vast amounts of time and money are spent on and with music in advanced industrialised society. $70 per year per capita and three-and-a-half hours a day are average figures, not peaks (Chapple and Garofalo 1977, Fonogrammen i kulturpolitiken 1979, Blaukopf 1982, Fabbri et al. 1986). Almost all this music is technically mediated via loudspeakers and headphones connected to radios, TV sets, cassette recorders, car stereos, record players, videos, etc. Music has become {280} more or less ubiquitous. It can be heard at home, at work, in restaurants, cars, buses, lifts, aeroplanes, airports, stations, shops, even in public toilets. Explaining the nature, qualities and uses of this omnipresent music is a interdisciplinary task, involving everything from business studies to theology, from electronics and acoustics to semiotics and linguistics, not to mention sociology, anthropology, psychology and musicology.1

It is with musicology that our problems start. The vast majority of music in our society falls under neither of the headings `art' or `folk' -- the traditionally legitimate areas of serious music studies --, the only current terms available for denoting the music most used by most people being mesomusica or popular music.2 There is no room here to explain why, at least until quite recently, musicology has managed to ignore most of the music produced and used in the post-Edison era (see Tagg 1987), but it does seem that this discipline has had considerable difficulty in expanding its range of methodological tools (chiefly developed as a conceptual system for demonstrating the aesthetic superiority and mythologically supra-social, `eternal' or `absolute' quality of Central European art music styles) to deal with other music.3

This sad state of affairs does not only hamper the development of musicology; it is also an obstacle to interdisciplinary mass media studies. Several scholars approaching the area of popular music from other angles, often from the discipline of sociology, have explicitly asked musicologists for analytical models explaining relations between musical structures and the larger set of sociocultural reality in which these sounds occur (e.g. Laing 1969 and Frith 1982). They have received no satisfactory answers, because traditional musicology has tended to steer clear of viewing music as a symbolic system whose structures are considered as either references to or as interpretations, reflections, reconstructions or re-creations of experiences which are not necessarily intrinsically musical. This is, in short, a problem of musical semiotics.

What is the role of musicology in this context? If we want to know `why and how who is communicating what to whom and with what effect', the social sciences should be able to answer the questions `who', `to whom' and `with what effect', and even parts of `why' and `how'. However, when it comes to {281} the rest of `how' and, more specifically, to `what' music is communicating, we will be left in the lurch unless that discipline devoted to the description, explanation and systematisation of musical structures -- musicology -- is prepared to tackle the issue (Wedin 1972). Unfortunately, there is another problem here, for not only has musicology until recently ignored popular music (leaving its study mainly to sociologists and/or journalists): it has also been exceedingly slow to develop viable theories of music in general as a symbolic system. These two problems are, as we shall see, interconnected.

Musical formalism, notation and popular music

Allowing for certain notable exceptions (mentioned later), traditional music analysis (Werkanalyse) seems to show either scanty interest and/or little skill in relating musical structures consistently to the remainder of human existence and activity, the discussion of any possible levels of `meaning' in the music either taking the form of sporadically intuitive conjecture or being passed by altogether.4 Other symptoms of this congeneric formalism are: (1) the apotheosis of `autonomous' instrumental music; (2) a widespread guild mentality among musicians, exhibiting the inability or lack of will to associate music explicitly with any other form of experience; and (3) a time-honoured adherence to notation as the most acceptable form of musical storage.

The latter has sometimes resulted in an almost ethnocentric fixation on `notatable' parameters of musical expression. Standard European notation was developed for music in which (1) pitches are fixable as one of twelve points inside any mean or equally tempered octave and (2) rhythm is based on downbeat accents in monorhythmic symmetric metre. This means that analysis of music using our notation system as its starting point will focus on characteristics containable within that system and on their extension into longer, graphically presentable processes (e.g. harmonic progression, sonata form).5 Such parameters may be particularly important in the understanding of the European art music tradition, whose notation was presumably developed to record its own idiosyncrasies rather than those of, say, Maori or African-American musics; but such traditionally notatable parameters are not necessarily of prime interest in the analysis of many types of popular music whose `immediate' or `present-time' aspects of musical expression (timbre, electromusical treatment, ornamentation, etc.) make for intensional rather than extensional complexity (Chester 1970a, b).

Although formalist analysis, often in Schenkerian guise, still dominates many university departments of musicology (e.g. Salzer 1962), {282} it should be seen as a historical parenthesis in the area of verbal discourse on music, this being bordered on one side by the Baroque Theory of Affects and on the other by the hermeneutics of music (Zoltai 1970:137-215). The doctrinal straitjacket of Affektenlehre, a mixture of feudal absolutism and rationalist curiosity, and its apparent tendency to regard itself as universally applicable (Lang 1942: 438; Zoltai 1970: 177) render it unsuitable for use in popular music analysis which must deal with a multitude of codes, ranging from that of Hollywood film music in the late romantic symphonic style to punk from the late seventies and from middle-of-the-road pop to the Webernesque sonorities of Bernard Herrmann.

Musical hermeneutics, an anti-formalist, subjective and interpretative approach (e.g. Kretzschmar 1911: 140), is often violently and sometimes justifiably criticised. Indeed, from time to time it degenerates into exegetic guesswork and to imaginative but spurious reading between the lines, also when applied to popular music (for example, Cohn 1970: 54-55; Melzer 1970: 104, 153; Mellers 1973: 117-118). Nevertheless, hermeneutics can, if used with discretion and together with other musicological approaches, make an important contribution to the analysis of popular music, not least because it treats music as a symbolic system and encourages synaesthetic thinking on the part of the analyst, a prerequisite for the foundation of verbalisable hypotheses and a necessary step in escaping from the prison of sterile formalism.

Semiotics of music = Semiotics of Western art music?

The transfer of structuralist and semiotic methods, derived from linguistics, to the realm of music seemed initially highly promising (e.g. Bernstein 1976). However, several musicologists of semiotic bent (for example Francès 1972, Imberty 1976b, Lerdahl and Jackendoff 1977, Keiler 1978 and Stoïanova 1978), have pointed out that models constructed to explain the denotative aspects of verbal language can by no means be transplanted wholesale into the field of music with its connotative, associative-affective character of discourse. Unfortunately, some linguistic formalism seems to have hampered the development of musical semiotics, the extrageneric question of relationships (a) between signifier and signified and (b) between the musical object under analysis on the one hand and the concurrent ideologies, attitudes, norms of behaviour on the other, seeming not infrequently to be regarded as suspect (Nattiez 1974: 67), or at least as totally subordinate to congeneric relationships within the musical object (Grenier 1983, critiquing Nattiez 1974: 72-73 and 1975: 414-416).

One reason for this congenericity seems to be that many musicologists rallying under the banner of `semiotics' or `semiology' have drawn almost {283} exclusively on art music for their supply of analysis objects, examples and test pieces. This means that a limited number of particular musical codes -- those used in European art music --, developed during a particular period of one continent's history by particular sections of the population in a limited number of communication situations, have been used, implicitly or explicitly, as musical evidence for the possible validity of highly generalised theories and metatheories of music as a symbolic system. In short, many musical semioticians have in this sense followed the formalist tradition of university musicology mentioned earlier, no other alternatives being readily available within the ivory tower where it sometimes seems that music can still be treated as though it were clinically free from sociocultural contamination by symbols. This sort of musical semiotics exhibits three characteristic symptoms: (1) concentrating on works whose composition techniques must be considered as highly marginal, i.e. as the exception to rather than as the rule of current musical practices, codes and uses (for example, Stoïanova 1978); (2) resorting to notation as the only form of storage on which to base analysis -- for Nattiez, the score seems to be music's equivalent of Peirce's `representamen', the `note' its phoneme (Nattiez 1976: 50, 198); (3) taking `European Standard Average Music' to mean, implicitly or explicitly, works by art music composers of the eighteenth, nineteenth and twentieth centuries, as though songs like My Bonnie, Lilli Marlene, Pod Moskovnye Vechera, Satisfaction, Volare, or Je ne regrette rien -- not to mention film or TV music by such composers as Jarre, Legrand, Barry or Morricone -- were all somehow less European, standard and average than the music of Schütz, Schumann or Schönberg. This neglect of popular music as a field of serious study by musicologists and semioticians is of course a matter of cultural politics which we shall not discuss here. It is also a matter of importance to the development of both disciplines.

Disadvantages of neglecting popular music analysis

Popular music cannot be analysed using only the traditional tools of musicology developed in relation to European art music of the classical period. This is because the former, unlike the latter, is: (1) conceived for mass distribution to large and often socioculturally heterogeneous groups of listeners; (2) stored and distributed in non-written form; (3) only possible in an industrial monetary economy, where it usually becomes a commodity; (4) subject, under capitalism, to the laws of `free' enterprise according to which it should ideally sell as much of as little as possible to as many for as much as possible. According to the last point (nº4), popular music should, if commodified as a recording, elicit `love at first listening' on the part of presumptive customers if the song is to stand a chance of making a sell before competing product reaches them.6 This means that music produced under such conditions will often require the use of {285} readily recognisable stereotypes of musical code as a basis for the production of (new or old) complexes of affective messages (Tagg 1979:56-66).

Avoiding this vast corpus of familiar and almost globally available music can be considered tantamount to missing the truly `general' in the music of our society, to turning down the chance to analyse what the music around us usually communicates as a rule. It would seem to make more sense to start by analysing mainstream musical codes (such as middle-of-the-road pop, music for film, TV and advertising), and to have some viable hypotheses about how these work before proceeding to the discussion of meanings and functions in subcultural, counter-cultural or `alternative' musical codes -- be it hip-hop, folk revival, string quartets or any other genre `contradicting' or `complementing' rather than `belonging to' the dominant mainstream of musical uses in any culture. Using exceptions as a basis for establishing rules may be fun but it is not the most reliable intellectual strategy.

Apart from this rather obvious point, the neglect of popular music as an area for musical-semiotic study can also cause basic methodological problems. We have already mentioned tendencies towards notational centricity in which the score is treated as reification of the `composition', `ideal performance' -- the niveau neutre? (Nattiez 1976: 54-55, 239-396) -- or as objectivisation of the `channel' between `emitter' and `receiver' (Eco 1976: 33), when in fact it is no more than a more or less incomplete shorthand of intended musical events communicated within the `emitting' stage of the communication process (i.e. that between composer and performer). Such confusion is less likely in the analysis of popular music where notation, if used at all, is never acknowledged as a satisfactory graphic representation of a musical event and where transcriptions can only be used as a visual aid to the analyst and reader.

Moreover, it seems that the almost exclusive study of European art music alone can also lead to sweeping notions about the `polysemy', `lack of precision', `non-referentiality', `absoluteness' and congeneric self-sufficiency of musical discourse. Not only does this tendency seem to hamper the development of nuanced, multi-level theories of musical meaning by bringing about a fixation on extensional structures, it also causes implicit scepticism about the degree of symbolic precision inherent at intensional or levels of musical communication.7 It is for example unfortunate that Imberty (1976a), in a highly interesting study of verbal association responses to Debussy's Préludes, only permits his respondents to use adjectives, thereby precluding nouns which can be the most common part of speech used in free induction test situations (Tagg & Clarida 2000). {285} Also, instead of using minimal musical-structural units against which to compare units of verbal associational meaning, Imberty also resorts to abstract parameters of musical expression as determinants for drawing the axes along which verbal associations are systematised into semantic fields. This procedure gives the impression that musical signification is vague, imprecise and contradictory. Similarly, Nielsen, basing his thorough empirical-statistical study of tension in music on responses to extensional aspects of musical structure only and by taking no more than a cursory glance at interesting verbal responses provided (1983: 247-248), gives the reader the impression that tension in music is almost exclusively created by congeneric means with no contribution to tension made by more `lexically' specific items of musical expression.8

However, instead of opposing extrageneric (emic, referential and multidisciplinary) with congeneric (etic, non-referential, unidisciplinary) approaches, it seems wiser to treat them as complementary. Only thus will it be possible to establish relations (extragenerically) between items of musical code and their respective fields of paramusical association and (congenerically) between these various individual parts of the musical code as processual structures too. These processual structures can then be discussed from an extrageneric viewpoint as well. What is being advocated here is simply a holistic, interdisciplinary approach to the analysis of popular music, an approach based on a thorough study of dialectical relations between the musical structure, its conception, production, transmission, reception and its social meaning, uses and functions.9

Suggesting solutions

Six general tenets

  1. In what follows, music is regarded as that form of interhuman symbolic communication distinguishing it from others in that individually and collectively experienceable states and processes are conceived and transmitted as humanly organised nonverbal sound structures to those creating these sounds themselves and/or to others who have acquired the mainly unreflected cultural skill of `decoding the meaning' of these sounds in the form of adequate affective response. {286}
  2. Direct imitations of or reference to sound outside the framework of musical discourse (iconic, programmatic, onomatopoeic ingredients) are extremely rare. Most elements of musical code are considered here as `referring' directly to either: (a) nothing outside themselves or (b) to their occurrence in similar guise in other music or (c) to their own intracontextual position within the same performance situation.10 At the same time, it is clearly absurd to treat music as a self-contained system of sound combinations, because changes in musical style are historically found in conjunction with (accompanying, preceding, following) changes in the larger sociocultural field of which music is a part. This would indicate that music is related to events outside itself after all.
  3. The contradiction described in nº2, above, is not antagonistic. A recurrent symptom observed when studying how music and musics vary inside society and from society to society in time or place is the way in which new means of musical expression are incorporated into the main body of any given musical tradition from outside the framework of its own discourse, i.e. as `Intonation Crises' (Assafiev 1976: 100-101). Such changes in musical code can work in a number of ways. They can: (a) `refer' to other musical codes (for example, social connotations of what sort of people use those sounds, how, when and where);11 (b) `reflect' changes in sound technology, acoustic conditions or the soundscape and changes in collective self-perception accompanying these developments (e.g. from clavichord to grand piano, from bagpipe to accordion, from rural to urban blues, from rock and roll to new wave) (Schafer 1977, Tagg 1987); (c) `reflect' changes in class structure or other notable demographic modifications (e.g. reggae influences on British rock, the change in dominance in US popular music from Broadway musicals to rock music of the South and West);12 (d) `reflect' changes in self-perception, group identity and ideology in society (e.g. from plainsong and polyphony to chorales and homophony, from sonata form to tone poems, from acid rock to heavy metal); (e) act as a combination of several of the above. After incorporation into the pre-existent musical tradition, new means of musical expression tend to lose their original referentiality and may even become so stylised that they end up as indicators, not of the style they borrowed from but of the one that did the borrowing -- for instance the way in which the `Hawaiian' guitar ends up as the `All-American' steel guitar of Country and Western music (Malone 1974:168-171).
  4. Cross-cultural `universals' of musical code are rare and mainly bio-acoustic, i.e. relations between human and musical pulse, between degrees of human bodily rates of movement, excitement, etc. and musical dynamics, rhythmic intensity, etc. For example, no lullaby will work if yelled jerkily at a brisk rate and {287} no war march will have the desired effect if crooned mellifluously at a snail's pace. Most other aspects of musical structure (pitch, timbre, texture, concepts of consonance and dissonance, etc.) share far fewer common connections to extramusical phenomena from one culture to another. Musical skills, be they passive or active, are therefore chiefly acquired by learning and are culturally specific.
  5. Musical communication can take place (a) between an individual and him/herself, (b) between two individuals, (c) between two or more members of a group, (d) between an individual and a group or vice versa, (e) between one group and another. Specifically musical (and choreographic) states of communication are (c), (d) and (e), i.e. those involving simultaneity of action between individuals producing sounds (or movements) perceptible as a unit. Compared with speech and the visual arts, music seems inherently suited to expressing collective messages and the affective identities and relationships of individuals in groups to themselves, each other and their surroundings.
  6. Music may be `polysemic' from a logocentric (verbal/linguistic) viewpoint because it is alogogenic, just as speech should be regarded as `polysemic' from a musicocentric viewpoint because it is not musicogenic. (Try producing music by reading this sentence in a normal tone of voice!). Thus, precision of musical discourse does not equal precision of verbal discourse since the two symbolic modalities are not interchangeable and have no quid pro quo relationship with each other (Francès 1972, Imberty 1976:36).

Analysing popular music

It should be clear that music is regarded here as a particular form of activity and communication in the much larger set of sociocultural reality of which it is a part. Apart from the problem of the analyst's `mentality' influencing preselection of method and approach (Törnebohm 1979), there is the problem of choosing an analysis object (=AO). At this stage of research, this choice is largely determined by methodological considerations mentioned earlier. The analyst must also have access to the AO in stored form and the final decision before analysing the musical structures is to determine which stage(s) in the tripartite (`emitter-channel-receiver') communication process should be studied in most detail. Notation has already been rejected as a primary storage form, the sounding object being the only accessible reification of the `channel' in the form of a recording. However, music as perceived by listeners and as conceived by emitters (composer, performer, producer, arranger, sound engineer, etc.) are equally important sources of information, since their relations to each other, to the sounding object and to the general sociocultural field are all vital parts of the perspective into which conclusions from the analysis of others stages in the musical communication process must be placed (Tagg 1982a).


Describing the composer's, producer's or performer's concept of the music under analysis (the AO) is often fraught with problems in the field of popular music. Apart from practical access problems to primary informants -- for instance, Hollywood film composers (Tagg 1979: 10-11) or international pop stars such as Abba (Tagg 1981: 12-13) --, emitters of popular music, unlike many of their colleagues in the avant-garde art music business, are often unwilling to be verbally explicit about the musical messages they naturally consider to be self-evident (Tobler and Grundy 1981, Pollock 1975). However, it is possible to glean some verbal information about an AO before it reaches the stage of performance on stage, film, record, video, etcetera.

An interesting example of how considerable semiotic information can be collected is offered by Hennion (1981) in an `ethnology' of a recording session during which the soloist, musicians, sound engineer and A&R man talk about mixing a `dry' guitar `up front', making the Fender `heavier' and `more explosive', `thickening' other guitar sounds, etc. Sounds are also referred to as `American', `hard and heavy', `tense and speedy' (yes, a rock number), `sensitive', `cool', `ironic', `throbbing with pain', `violent but discrete', `crackly', `sea green', `beyond the tomb', `drooling organ', `majestic with a hangover', etc. (in a dramatic ballad at the words `quand le mec devient fou'). It was in such pre-performance situations of communication between musical `emitters' that the generally accepted register of Italian terms describing the spirit in which the musical notation should be put into sound was developed (allegro, vivace, lento, largo, poco lusingando, scherzando, sforzando, lamentoso, etc.). However, before this veritable gold-mine of modern inter-musician pre-performance terms of popular music production can be used methodically in detailed semiotic analysis, many more studies like Hennion's `ethnography' will need to be conducted and the use of terms systematised into some sort of semantic fields corresponding to sets of musical structures.13

One readily accessible set of verbal descriptions of musical affect used by the emitters of popular music is that systematised on the basis of the long-standing practice of selling prerecorded music designed to evoke specific moods. We are referring here to collections of `library' or `mood' music.14 The nomenclature of main categories, subheadings and titles contained in these collections is reasonably consistent from one catalogue producer to another. For example, a {289} comparative study of the Nature category as semantically determined by the titles, descriptions and crossover rates of six hundred pieces in five mood music collections showed `nature' as a musical concept frequently qualified as pastoral, ethnic, bucolic, scenic, calm, quiet, serene, peaceful, romantic, tender, gentle, soft, sad, meditative, nostalgic, religious, broad, spacious, fresh, flowering, and healthy. The library music category `nature' was diametrically opposed to such concepts as industry, technology, science, atomic, machines, research, automation, crime, violence, turmoil, tense, modern, traffic, parade, enterprising, threat, sinister, space, eerie, laboratory (Tagg 1982b).15

This view of nature as an affective-associative sphere corresponds well with the `joy-serenity-calm' sections of Imberty's `scale of musical expressivities' (1976: 24) and such results could contribute to a deeper conceptual analysis of general concepts like `nature', `science' or `technology', since they deal not so much with verbally semantic or visually/gesturally symbolic fields but with affective and associative levels of cognition. Systematic study of such emitter verbalisation, if related to the sounding object and to music as perceived and used, might make a valuable contribution to a semiotic theory of popular music.


Most empirical studies of musical perception (or reception) have, as inferred earlier, been conducted with art music as stimulus for the responses reported, for example by Francès (1958), Karbušicky (1973), Imberty (1976), Rösing (1983) and others. Systematic studies of popular music reception are still very rare and seem generally to be concerned with psychosocial aspects of levels, types, or attitudes of perception (that is, with that which most obviously distinguishes popular music communication situations from the idealised Adornistic listening type) rather than with semiotic issues of musical discourse (Rösing 1983). Obviously, there are innumerable problems surrounding the empirical study of popular music reception, not least that of finding test situations which do not automatically negate the communication and perception contexts peculiar to most uses of popular music. Just imagine the methodological difficulties involved trying to glean reliable verbal statements about music while the informants dance, drive, eat, do their housework or homework or slouch in front of the TV after a hard day's work. Although these problems may seem insurmountable, there are other ways of carrying out empirical studies of popular music reception.

It is, for example, quite possible to elicit verbal associations by free induction to short music examples if, as in the case of standard film, TV or advertising scores played without accompanying {290} visuals, respondents are asked to jot down as quickly as possible (they should be given no time to `think') what they envisage (`see', `hear' or `feel' in their mind's ear/eye) to be taking place on an imaginary screen along with the music they hear. Results collated so far show considerable statistic reliability and homogeneity of response.16 In the case of The Dream of Olwen, for example, there was a 2:3 probability of any respondent describing a scenario in nature and a 1:2 likelihood of associating actively and explicitly to `romance, love or sentimentality'. Moreover, there was a 1:5 probability of associating to the `dance, walk, flow, stroll, float, undulate, sway' qualities of movement and one in five had the sun shining. In other words, musical meaning, as perceived by 560 Swedes, 10 Norwegians and 40 Latin Americans, in response to a standard piece of film music in the vein of a major key, andante, mp - f, molto legato e melodioso e arpeggiato Rachmaninov-style piano concerto is extremely unambiguous.

The relationship between scenarios, patterns of movement, personality types and other associated categories can be partly explained by semantic common denominators based on bodily movements and functions connected (culturally or biologically) with particular emotional states which can, in turn, be connected with the `general emotional' categories of association. These various types of freely induced verbal response can, moreover, be related to the musical structures of the piece and a typology of title music signs could be presented -- possibly using the Peircian model put forward by Boilès (1982) -- and drawing on The Dream of Olwen or similar excerpts of title music as a basis for such a theory. Nevertheless, however promising the results of reception tests may seem at this stage, it is ultimately by studying the sounding object (`the channel') that musicologists can contribute most extensively to the semiotics of popular music.

Hermeneutic-semiotic method

Choosing the sounding object as our reification of the `channel', the first methodological tool is that of interobjective comparison (=IOC). IOC helps solve the analyst's riddle of music's alogogenicity by using other music instead of words as metalanguage for music. If an analytical approach establishing consistency of response to the same AO played to a number of respondents is called intersubjective, then an interobjective approach is that which establishes similarities in musical structure between the AO and other music. The researcher can carry out this procedure alone, but the scope and reliability of the interobjective comparison material (=IOCM) can be considerably increased if others are asked to associate musically to the AO. In this way, musicians with varying specialities, be they avid listeners or fans, producers, DJs, conductors, performers, composers or arrangers are co-opted as expert informants, racking their vocal and tactile (not verbal) conscious for snippets of sound, turns of phrase, timbres, gestures, textures, cadences, rhythms, etc. which they feel sure they have heard, played, sung, danced to or conducted before. This process of musical association can establish banks of IOCM which for one single AO can include references to hundreds of different musical works.

The next step is to search the IOCM for structural elements similar to those found in the AO. These elements can range from the snippets of sound, rhythm, timbre, melody etc. just mentioned to general sonorities and overall expressional constants. Structural correspondences of this type must then be tested for paramusical associations. This type of symbolic relationship can be established if pieces of music in the IOCM share any common denominators of paramusical association in the form of visual, verbal, gestural or any other paramusical message. If a particular set of musical structural elements in different examples of IOCM are found to bear consistent relationship to the same sort of paramusical phenomena, the latter may then be considered as an paramusical field of association (=PMFA) and the corresponding musical structural elements as items of musical code (=IMC). If objective structural correspondences are established between the AO and the IOCM and if there is demonstrable correspondence between the IOCM and its PMFAs, then the structural elements of the AO corresponding to those of the IOCM may also be considered as IMCs and the relationship between these elements and the PMFAs of the IOCM may be regarded as similar to that between the IMCs and PMFAs of the IOCM. For example: if our AO is a TV title theme including a forte horn whoop spanning at least a fifth, played unison a quatro, rising from an unaccented to an unaccented tone where it lands in a consonant relationship to the underlying harmony, all in moderato to allegro tempo, we could start by asking horn players to tell us where they had played such things before. Trying different keys and slightly different intervals and tempi, they might come up with such references as the male theme from Strauss's Don Juan or the sword motif from Wagner's Ring. Such references would initiate a bank of IOCM. The IOCM could be enlarged by listening to other title music. If in the process we discover similar musical structures occurring in the hero's themes of How The West Was Won, Star Wars, and The Saint, we are well on our way to making a hypothesis about heroism as an PMFA to this horn whoop, which now becomes a presumed IMC in the AO as well.

Having extracted the various IMCs of the AO in such a way and interpreted their associative meaning by means of IOC, the validity of {292} these temporary conclusions may be tested. Since it is impossible to construct reception test models isolating the effects of one item of musical code in a particular listening situation, hypotheses about its symbolic value may be tested by a commutation technique well known from such practices as `majoring', `minoring', `jazzing up' and `rocking up', i.e. by altering one or more parameters of musical expression presumed to be carrier of a particular symbolic value and by keeping the rest in tact. This technique of commutation is called hypothetical substitution and, by defining the musical structure(s) operative in carrying particular symbolic values, also establishes these structures as musematic. Such musematic structures occur as stacks or strings (see below).17

Having established musical `meaning' at the micro level, one may then proceed structuralistically to the explanation of how musemes are combined, simultaneously and successively. Unlike verbal symbols -- where complexities of affective association can generally only be expressed through a combination of the string of semantic discourse and simultaneous inflection, facial expressions, gestures, etcetera -- music can express such complexities through multi-layer sets of simultaneously heard musemes. Several theoretically discretisable musemes are generally combined to form what the listener hears as a single sound entity. These museme stacks can be seen as a vertical cross-section through an imaginary score. They have little or no phenomenological duration, never exceeding the limits of `present time' experience in music (Wellek 1963: 109). In popular music, museme stacks seem to correspond to the concept of `sound', one of whose chief characteristics is a hierarchy of dualisms, one of the most important in Western music (art and popular) being that between melody and accompaniment and interpretable as the musical version of the relations figure to ground (visual arts), individual to environment (social sciences), main character to story (novels) (Maróthy 1974: 22; Mühe 1968: 67; Tagg 1979: 123-124).

The general meaning of the main blocks of `now sound' thus established, it is now possible to climb up the structural hierarchy through the syntactic analysis of melodic phrases -- a question too complex to discuss here (Tagg 1979: 186-202) -- to the point where larger patterns of musical process can be examined. Here we have reached the happy hunting ground of formalist musicology with its sophisticated conceptual apparatus for describing the organic mutation and development of themes. However, as Chester (1970a, b) has suggested, there are clear differences between the extensional type of musical discourse found in the heyday of sonata form and the intensional blocks on which much popular music (not least rock and film music) is constructed. This does not mean that patterns of musical process in popular music analysis {293} are a simple matter (Wicke 1978). Although block shifts (i.e. simultaneous changes in many parameters of musical expression within the same AO) are reasonably clear in joins between verse and chorus, A and B sections (or, in films, between scenes), the total meaning of ostensibly straightforward patterns of recurrence, either as reiteration, repetition or recapitulation, can often be far more problematic than their deceptive simplicity suggests (Tagg 1979: 217-219, Middleton 1983). The situation becomes even more complex when there is incongruence between musical and paramusical processes in the same AO. Only a depth analysis of simultaneity, staggering, congruence or incongruence of change and return in both musical and paramusical processes can actually lead to any deeper understanding of musical meaning in the AO.18

As a final analytical measure, it is also possible to apply a checklist of parameters of musical expression (Tagg 1979:69-70). This allows for hypothetical substitution (commutation) of all structures under analysis (both IMCs and patterns of musical process) which have been postulated as corresponding to PMFAs. The checklist includes the following aspects: absolute and relative duration, order of sections, pulse and tempo, metre, periodicity, rhythm, melody, pitch, range, tonal vocabulary, orchestration, timbre, phrasing, accentuation, ornamentation, tonality, texture, harmony, parts or voices (type, number and interrelations) relative and absolute dynamics, ambient acoustics, electromusical treatment, rates and patterns of change in parameters.

This final stage of analysis can be further refined by using a checklist of paramusical parameters of expression (Tagg 1979: 68-69), allowing the analyst to see how conclusions arrived at by IOC correspond with the relation between AO/IMC and AO/PMFA. This check list includes non-musical sounds, spoken or written language, paralinguistic communication, graphics, visual symbols, movements, behavioural patterns, etc.


At this stage in the analytical proceedings it is possible to collate information from the emitting and receiving ends of the musical communication process (if available) with findings from the hermeneutic-semiotic view of the AO. This operation should provide enough information for a wider cultural or ideological discussion.

For example, detailed musematic analyses of the title theme for the TV series Kojak and of Abba's Fernando (Tagg 1979, 1991) showed such mainstream popular music to be carriers of messages which were able to relate types of personality, environment and events to emotional attitudes, implicit evaluations and patterns of affective response. In Kojak {294} the music was found to reinforce a monocentric view of the world and to affectively emphasise the fallacy that the negative experience of a hostile and criminal urban environment in North America can be overcome by adopting a Rambo-like individualist attitude of self-sufficiency and go-it-alone heroism. In Fernando, another sort of monocentricity prevailed in which the threat and worry epitomised by oppression, poverty and strife in the third world under imperialism were warded off by a tourist stance and by pleasantly nostalgic reminiscences heard against a familiar `home' backcloth of leisure.

In these cases, the historical position of the emitters in the socioeconomic context of their respective branches of capitalist cultural production allowed at that time for little or no real freedom to contradict existing dominant patterns of musematic or ideological stereotypes (Tagg 1979:81-84, 1991: 6-13, 105-144). There is no room to give details here, save to say that in the case of Fernando, knowledge of Abba's position and activities in the international music business and on the Swedish cultural scene at that time, together with an overview of relevant historical and cultural developments immediately preceding the release of the record in 1975 (fascist coup in Chile, victory in the Eurovision Song Contest) was essential in contributing to the interpretation of the song as picture-postcard exoticism rather than as anti-imperialist involvement (Tagg 1985: 260-261). It was in other words necessary to look to the general sociocultural field of study of which the artists, their music and its audience were a small part, in order to discover what particular phenomena within that larger set of reality were being referred to, dealt with, interpreted and thereafter affectively and ideologically recreated and communicated in the song.

In fact this is the heart of the matter, because the ultimate aim of the discussion presented here is to try and find ways of increasing our understanding about that vast quantity of music around us, about how it works and influences us. This means understanding music's references to and interpretations of experiences caused by phenomena in the larger sociocultural set of reality of which music is an integral part. Perhaps this is also what musicology and the semiotics of music are really all about.

Bibliography {296}

Alencar Pinto, Guilherme de (1986). Análisis: `Deus lhe pague' de Chico Buarque. La del taller 1985/4: 19-26. Montevideo: Talleres latinoamericanos de música popular.

Assafiev, Boris (1976). Die musikalische Form als Prozess. Berlin: Verlag neue Musik.

Bernstein, Leonard (1976). The Unanswered Question. Cambridge: Harvard University Press.

Björnberg, Alf (1984). `There's Something Going On' - om eolisk harmonik i nutida rockmusik. Tvärspel - 31 artiklar om musik -- festskrift till Jan Ling: 371-386. Göteborg: Skrifter fr. Musikvetenskapliga institutionen, 9.

Blaukopf , Kurt (ed.) (1982). The Phonogram in Cultural Communication. Wien: Springer Verlag.

Boilès, Charles L (1976). Signification in Cinema Music. Versus - Quaderni di studi semiotici, 13 (ed J-J Nattiez): 49-61. Milano: Bompiani.

-- (1982). Processes of Musical semiosis. Yearbook for Traditional Music, 14: 24-44

Brolinson, Per-Erik, and Larsen Holger (1981). ...And Roll. Aspekter på text och vokal gestaltning. Stockholm: Esselte Studium.

Chapple , Steve, and Garofalo Reebee (1977). Rock-'n'-roll Is Here To Pay. Chicago: Nelson Hall.

Chester , Andrew (1970). For A Rock Aesthetic. New Left Review, 59: 82-86.

-- (1970). Second Thoughts on a Rock Aesthetic: The Band. New Left Review, 62: 75-82.

Cohn , Nik (1970). Awopbopaloobopalopbamboom. St Albans: Paladin.

Desai , Meghnad (1974). Marxian Economic Theory. London: Gray Mills.

Druud-Nielsen , Per (1981). Studier i hverdagsmusikkens sociologi. Århus: PubliMus.

Eco , Umberto (1976). A Theory of Semiotics. Bloomington & London: Indiana University Press.

Fabbri , Franco, Fiori , Umberto, Ala Nemesio, and Ghezzi Emilio (1985). Patterns of Music Consumption in Milan and Reggio Emilia from April to May 1983. Popular Music Perspectives, 2 (ed. D Horn): 464-500. Göteborg, Exeter, Ottawa, Milan: IASPM.

Fonogrammen i kulturpolitiken (1979). Rapport från Kulturrådet, 1979:1. Stockholm: Liber.

Francès , Robert (1972). La perception de la musique. Paris: J Vrin (2ème édition)

Frith, Simon (1982). The Sociology of Rock: Notes from Britain. Popular Music Perspectives, 1 (D Horn & P Tagg):142-153. Göteborg and Exeter: IASPM.

-- (1983). Sound Effects. Youth, Leisure and The Politics of Rock-'n'-roll. London: Constable.

Grenier, Line (1983). S'inspirant des travaux de Nattiez, quelle définition peut-on donner de la musique comme système symbolique qui permette une approche sociologique? Unpublished paper. Université de Montréal, Département de sociologie.

Harker , Dave (1980). One for the Money. London: Hutchinson.

Hennion, Antoine (1981). Les professionnels du disque. Paris: Métaillié.

Horn D, and Middleton , R. (eds.) (1984) Popular Music, 4.

Imberty, Michel (1976a). Signification and Meaning in Music. Montréal: Monographie de sémiologie et d'analyses musicales.

-- (1976b). Le problème de la médiation sémantique en psychologie de la musique. Versus - Quaderni di studi semiotici, 13 (ed J-J Nattiez): 35-48. Milano: Bompiani.

Karbuicky , Vladimir (1973). Das `Verstehen der Musik' in der soziologischen-ästhetischen Empirie'. In Musik und Verstehen -- Aufsätze zur semiotischen Theorie, Ästhetik und Soziologie der musikalischen Rezeption, H P Reinicke and P Faltin (eds): 121-147. Köln: Volk/Gerig.

Kayser, Dietrich (1976). Schlager - Das Lied als Ware (zweite Auflage). Stuttgart: Metzler.

Keiler, Allan (1978). Bernstein's The Unanswered Question and The Problem of Musical Competence. Musical Quarterly: 198-222. {297}

Kretzschmar , H (1911). Gesammelte Aufsätze aus den Jahrbüchern der Musikbibliothek Peters. Leipzig: Peters.

Laing , Dave (1969). The Sound of Our Time. London & Sydney: Sheed & Ward.

Lang , Paul H (1942). Music in Western Civilization. London: Dent.

Lerdahl , Fred and Jackendoff , Ray (1977). Toward a Formal Theory of Tonal Music. Journal of Music Theory, 1977/1: 111-171.

Ling , Jan (1978a). Assafjevs intonationsteori: ett försök till analys. Stencil, Göteborg: Musikvetenskapliga institutionen, Göteborgs universitet.

-- (1978b). Hjort Anders' 12th Day March. Antropologiska Studier, 25/26: 122-139.

-- (1986). Hur klassisk musik blev klassisk. Unpublished paper. Göteborg.

Malone , Bill C (1974). Country Music USA. Austin & London: University of Texas Press.

Maróthy , János (1974). Music and The Bourgeois, Music and The Proletarian. Budapest: Akadémiai Kiadó.

Mellers, Wilfrid (1973). Twilight of the Gods. London: Faber.

Melzer , R (1970). The Aesthetics of Rock. New York: Something Else Press.

Middleton , Richard (1981a). Introduction. Popular Music, 1: 1-8.

-- (1981b). Reading popular music. Popular Culture: Form & Meaning 2. Milton Keynes: Open University Press.

-- (1983). Play it again, Sam -- the productivity of repetition. Popular Music, 3: 235-271.

Mühe, Hansjörg (1968). Zur Intonation des deutschen Schlagers: Untersuchungen an in der DDR gespielten Evergreens. (Unpublished doctoral dissertation). Leipzig: Karl-Marx-Universität.

Nattiez, Jean-Jacques (1974). Sur la relation entre sociologie et sémiologie musicales. International Review of the Aesthetics and Sociology of Music 5/1: 61-75.

-- (1976). Fondements d'une sémiologie de la musique. Paris: Union Générale d'Éditions.

Nielsen , Frede V (1983). Oplevelse af musikalsk spænding. København: Akademisk Forlag.

Pollock , Bruce (ed.) (1975). In Their Own Words -- Twenty Successful Song Writers Tell how they Write their Songs. New York and London: Collier/Macmillan.

Riethmüller , Albrecht (1976). Die Musik als Abbild der Realität: zur dialektischen Wiederspieglungstheorie in der Ästhetik. Wiesbaden: Beihefte zum Archiv für Musikwissenschaft, 15.

Rösing , Helmut (1977). Musikalische Stilisierung akustischer Vorbilder in der Tonmalerei. München & Salzburg: Katzbichler.

-- (1981). Die Bedeutung musikalischer Ausdrucksmodelle für das Musikverstndnis. Zeitschrift für Muskpädagogik, 16: 158-264.

-- (ed.) (1983). Rezeptionsforschung in der Musikwissenschaft. Darmstadt: Wissenschaftliches Buchgesellschaft.

Salzer , Felix (1962). Structural Hearing. New York: Dover.

Schafer , R Murray (1977). The Tuning of The World. Bancroft: Arcana.

Schmidt , H-C (ed. 1976). Musik in den Massenmedien Rundfunk und Fernsehen. Mainz: Schott.

-- (1982). Filmmusik: Analysen, Beispiele, Kommentare. Kassel: Bärenreiter.

Schuler , Manfred (1978). Rockmusik und Kunstmusik der Vergangenheit. Archiv für Musikwissenschaft, 35: 135-150.

Seeger , Charles (1960). On the moods of a musical logic. Journal of the American Musicological Society, 13: 224-261.

Stefani , Gino (1976). Introduzione alla semiotica della musica. Palermo: Sellerio.

-- (1982). La competenza musicale. Bologna: Cooperativa Libraria Universitaria Editrice.

Stoïanova , Ivanka (1978). Geste - texte - musique. Paris: Union Générale d'Éditions.

Tagg , Philip (1979). Kojak - 50 Seconds of Television Music. Göteborg: Skrifter från Musikvetenskapliga institutionen, 2. {298}

-- (ed. 1980). Film Music, Mood Music and Popular Music Research. Göteborg: Stencilled Papers from the Gothenburg University Musicology Department no. 8002.

-- (1981). `Fernando the Flute' -- Analysis of affect in an Abba number. Stencilled Papers from the Göteborg University Musicology Department, 8106 (radically revised and expanded as Tagg 1991).

-- (1982a). Analysing Popular Music. Popular Music, 2: 37-69.

-- (1982b). Nature as a Musical Mood Category. Göteborg: Stencilled Papers from the Gothenburg University Musicology Department, no. 8206.

-- (1985). Zur Analyse von populärer Musik. Beiträge zur Musikwissenschaft, Heft 3/4 1985: 241-264.

-- (1987). Reading Sounds. Göteborg: IASPM-Norden Working Papers. Revised and published in Popular Music Research (1990, K Roe and U Carlsson, eds.): 103-114. Göteborg: Nordicom .

-- (1991). Fernando the Flute. Liverpool: Institute of Popular Music, Liverpool University.

Tagg , Philip and Clarida , Robert (2002). Ten Little Title Tunes (forthcoming).

Tarasti , Eero (1978). Myth and Music. Den Haag: Mouton.

Thorsén , Stig-Magnus (1977). För musiken i tiden. Stockholm: Sveriges Radio.

Tobler , John and Grundy , Stuart (eds. 1982). The Record Producers. London: BBC.

Tokaji , András (1982). Das Massenlied in Ungarn 1945-1956. Budapest: Institut für Kultur.

Törnebohm , Håkan (1979). Kunskapstraditioner och mentaliteter. (Nr 43, serie 2). Göteborg: Institutionen för Vetenskapsteori, Göteborgs universitet.

Vega , Carlos (1966). Mesomusic. An essay on the music of the masses. Ethnomusicology: 1-17.

Wedin , Lage (1972). Multidimensional Scaling of Emotional Expression in Music. Svensk Tidskrift för Musikforskning, 54: 115- 131.

Wellek , A (1963). Musikpsychologie und Musikästhetik. Frankfurt-am-Main.

Wicke , Peter (1978). Licht in das Dunkel - Popmusik in der Analyse. Beiträge zur Musikwissenschaft, 1978/1: 3-15. Berlin: Verlag neue Musik.

Zak , Vladimir (1979). O melodik'e massovoi pesni. Moskva: Vsesoyuznoye izdatel'stvo Sovietskii Kompozitor.

Zoltai , Dénes (1970). Ethos und Affekt, trans. by B Weingarten. Budapest: Akadémiai Kiadó.

1. Accounts range from notions of capitalist conspiracies and populist emancipation to descriptions of musical styles, genres, subcultures, functions, habits, behaviour, modes of production and reception, etc. There is a host of competent literature on these subjects (e.g. Frith 1978, Kayser 1976, Druud Nielsen 1981).

2. Both are unsatisfactory terms negatively denoting what is `left over' after `serious' art music and `authentic' folk music have been taken care of. Mesomusica is the concept used by Vega (1966), while `popular music', the term used here, has been delimited by Middleton (1981:1-7) and Tagg (1979:20-32).

3. `How Classical Became Classical' is the English translation of the title of an interesting paper by Ling (1986) dealing with the origins and development of notions about `classical' and `absolute' music in the late 18th and early 19th centuries.

4. For examples of such discourse, see Tagg 1979: 38,ff. The whole notion of `absolute music', even in connection with the classical tradition, becomes increasingly spurious after reading works by scholars like Tarasti (1978) and Rösing (1977).

5. Some empirical tests to prove this point: (a) try getting your average chorister to `swing' a birhythmic Byrd madrigal properly when he/she structures passing music time with the help of bar lines; (b) transcribe a Hendrix solo, an Aretha Franklin vocal line or a Keith Richards guitar riff; (c) transcribe a kwela, a gamelan piece in slendro or any raga performance; (d) sight-read some Pandered or transcribe the music to any murder scene on television. Good luck!

6. For a general description of the circulation of capital in the music industry, see Thorsén (1977). A more thorough description of these mechanisms in the economy at large is provided by Desai (1974).

7. Intensional aspects and symbolic precision in music (`musematic meaning') are discussed later on under `Hermeneutic-semiotic method'.

8. We are referring here to museme strings and stacks. For fuller explanation of the term `museme', see Tagg (1979:70-79, 102-154).

9. In this context, valuable impulses can be found in `Intonation Theory' (Assafiev 1976). Though the Assafievian term `intonation' is problematic because of the diversity of meanings ascribed to it apart from those it already possesses, intonation theory seems capable of putting musical analysis into historical, social, and psychological perspective, avoiding the pitfalls of sterile formalism and unbridled hermeneutics (e.g. Maróthy 1974, Zak 1979). Other interesting work combining semiotic with other musicological approaches has been provided by Francès (1972), Stefani (1976, 1982), Imberty (1976a), Rösing (1977) and Tarasti (1978). However, in none of these publications are the analytical models applied to popular music. Allowing for some exceptions -- e.g. Alencar Pinto (1986), Boilès (1976), Wicke (1978), Tokaji (1982), Ling (1978a,b), Middleton (1983), Björnberg (1984) and some articles in Schmidt (1976) -- there is a notable dearth of scientific approach to musical structures and their symbolic value in the modern media.

10. A `performance situation' might just be a `piece of music', but could also be a film, an evening's dancing, background radio while doing the washing up, the complete opera, the same aria within the opera, etc.

11. `Man darf nich Tango in der Kirche spielen' because the connotative symbols of tango (who does it where, why, how, when and with what effects) are in contradiction to the connotative symbols of church music (who does that, where, why, how, when and with what effects?).

12. Reflect' is used in the non-mechanist sense of wiederspiegeln, as argued by Riethmüller (1976).

13. Hennion (1981:107-134). An attempt to systematise descriptions of vocal timbre used by rock artists is offered by Brolinsson and Larsen (1981). A systematisation of semantic fields based on verbal-visual association responses to film music played without visual accompaniment will appear in Tagg & Clarida 2000.

14. Customers for such products are radio and TV stations requiring quick access to instant mood music (e.g. for radio plays or documentary underscores), producers of newsreels, 16 mm films, adverts, slide presentations, low-budget video promos, etc. For some interesting semiotic aspects of this sort of production, see Tagg 1980.

15. It is interesting that Nature, if defined as that part of the material world uninfluenced by Man, should be directly opposed to Space (cosmos). Even polar regions are dubious as Nature according to `musical' definitions.

16. The Dream of Olwen was composed by Charles Williams as title theme for the film While I Live (UK, 1948) and is recorded by Geoff Love and his Orchestra on Big Concerto Movie Themes (Music for Pleasure MFR 5261, 1972). The nature answers break down as nature/ countryside 1:7, meadow grass 1:8, woodland, farms, parks and gardens 1:8. Here, as in responses to other test pieces, there is a preponderance of `non-adjectival' associations. Provisional results show so far (a) an average of about 3 verbal associations per respondent and test piece, (b) that the 13 most frequently chosen categories account for 58% of all responses, the 17 least frequently chosen categories accounting for a mere 2%.

17. The term `museme' is coined by Seeger (1960:76) and explained as `a unit of three components - three tone beats -'... which can constitute... `two progressions and meet the requirements for a complete, independent unit of music-logical form or mood in both direction and extension'. There are some problems with this explanation regarding correspondences between phonemes/morphemes and the musical counterparts Seeger seems to provide them with (his `tone beats'/`musemes') (Tagg 1979:71-73).

18. The methodological problems here are considerable. For example, a melodic tritone motif in Abba's Fernando was found to be musically polysemic, its symbolic function of longing or resolution of longing depending solely on its syntactic position in the piece (Tagg, 1991:62-73).