An Anthropology of Stereotypes in TV Music?

by Philip Tagg

First published in Svensk tidskrift för musikforskning, 1989, pp.19-42.
[Swedish Musicological Journal for 1989, Göteborg, 1990]

View correct PDF file
Why did I land here?

Home page at this website (Philip Tagg, Liverpool) Home
Access to texts by P Tagg and other authors Online Texts

 


This text is especially for Bob Clarida of New York and Anahid Kassabian of Redwood City for having had the curiosity, friendliness and intelligence to help push this work in good directions. Thank you both.

Introduction

Background

In earlier publications I have argued that it is necessary to include musical semantics among the various approaches used to study popular music,1 partly with a view to offsetting the sociological bias of that field of research, partly to offer musicology the opportunity of developing methods of analysis that transcend the limits of theme spotting and harmonic digests à la Schenker. I have also argued that musicology of semantic persuasion has, with a few remarkable exceptions,2 fallen short when it comes to establishing viable relationships between musical structures and paramusical symbols, falling back on notated music and notationally storable parameters of expression to construct quite general theories and metatheories of musical meaning.3

For the last four years I have been working on a research project under the title `The Semiotics and Aesthetics of TV Music',4 in which the empirical base of perception and a discussion of musical sign typology has been added to semantic and hermeneutic methods developed at an earlier stage of research.5 During the course of this more recent work, it has been necessary to confront some anthropological problems, a few of which will be discussed here. The prospect of being anthropological about television music seemed quite attractive, bearing in mind that ethnomusicologists also have to analyse the structures of non-notated music and to relate these to the culture in which they are produced and used. Fuelled with such aspirations, I joined the International Council for Traditional Music (ICTM) and the European Seminar in Ethnomusicology (ESEM).

Ethnomusicology and TV `field work'

The ESEM and the ICTM both sent me a membership questionnaire containing a box in which I was asked to fill in details about my `field work'. Never having taken to the fields, let alone to the savanna or polar wastes, to study the music I was interested in, I was totally nonplussed. According to the associations' directories, most of my fellow members studied people living on coral reefs, in the tundra, the high Andes or the outback, only a few getting as close to home as immigrants in North American urban ghettos. So I was obliged to fill in the `field work' window of my questionnaire by stating that my academic safaris are conducted from the living room couch, whence I peer into the concrete and neon jungle of mass media entertainment on the tube: I don't even have to leave the apartment. This `field work' looks as incongruous as it is exceptional amongst the more legitimate jungles that other members seem to visit, for, as Nettl (1964: 1-2) points out, ethnomusicologists have traditionally been

`students of the music outside Western civilization and, to a smaller extent, of European folk music'. [They] `have worked, on the one hand, as the special kind of musicologist who investigates exotic music and, on the other hand, as the special kind of anthropologist who investigates music rather than other aspects of human culture, again outside Western civilization.'

Nettl is in fact quite clear about the issue, for, in discussing `the scope of Ethnomusicology' (op. cit: 5-7), he suggests that the subdiscipline deals mainly with music from three types of culture: (a) `preliterate' society, (b) Asian and North African `high cultures' and (c) oral (`folk') traditions within `high cultures'. Whereas Nettl does not so much as refer, even obliquely, to mass-media music, Jaap Kunst (1959:1) made no bones about the matter, stating that ethnomusicology `investigates all tribal and folk music and every kind of musical acculturation.' He continues: `Western art- and popular (entertainment-) music do not belong to its field'.

Such implicitly or explicitly restrictive views of which musics constitute legitimate jungles for ethnomusicological safaris may have been challenged by such scholars as Hood (1957:2), Chase (1958:7) and Merriam (1964:5-7)6 and the study of `popular music'7 may, as Skog (1977:617) points out, be considered comme il faut in certain circles,8 but this latter, culturally anthropological, view9 still does not seem to have made much impression on ethnomusicology as a whole, that discipline still remaining largely conformist in its choice of subject material (it is now 1990), at least judging from the articles contained in Yearbook for Traditional Music or Ethnomusicology and from the `field work' entries in the annual membership directories of the ESEM and ICTM.

If anthropology is `the study of man, his origins, institutions, religious beliefs, social relationships, etc.', if `ethno' is a morpheme `indicating race, people or culture' and if musicology means `the scholarly study of music', then the anthropology of music should be the study of music in relation to man, his institutions, beliefs, social relationships, etc., while ethnomusicology ought logically to mean the scholarly study of the music of one or more peoples or cultures. However, although there are no caveats or exclusion clauses in any of these definitions, ethnomusicology maintains largely ethnocentric, declaring by default any culture `ethnic' as long as it is not that of our own ethnos and deeming any anthropos worthy of serious study, as long as he/she is not one of us. Of course, such anachronism between rational definition of a discipline and its actual practice applies inversely to musicology which, by omitting any qualifier (`ethno', `anthropo' or whatever), defines itself in practice as the scholarly study of a certain type of music. Thus, with supreme `first world logic' we can summarise the following:

  1. ethnomusicology means studying the music of any culture but in practice it largely excludes our own;
  2. anthropology means studying humans and their culture but in practice it often excludes our own culture and the people in it;
  3. musicology means studying any music in any culture but in practice it excludes everything except a small part of our own.

Since this article becomes quite `ethno' (i.e. anthropological) about music used by the vast majority of our own ethnos, we shall henceforth be using dictionary rather than common practice definitions of the terms discussed above. In fact, to return to the matter that caused this polemic against theoretical and practical definitions of subdisciplines in music studies, we now hope to suggest how a combination of empirical, semiotic and anthropological approaches to the study of television music may be able to help us answer some of the $64,000 questions about the nature of music in our mass-media society. We shall, however, need to start at a much less grandiose level. We kick off therefore with a short survey of the empirical part of our current research.

Present research

Reception tests

Ten title themes for television or film, each lasting between thirty and sixty seconds, were selected as test pieces10 and played to persons attending or commandeered to a lesson or lecture that I was supposed to hold.11 88% of these respondents were Swedish, 2% Norwegian and 10% Latin American. About 30% were pupils or students with a special interest in music -- the `unmusical' populace constituting the remaining 70% -- and most respondents were below the age of thirty. On some occasions only three pieces were played, more frequently the first four and in a few instances all ten. This means that over six hundred people responded to the first three tunes, over four hundred to the first four and only one hundred and five to the last one. For this reason, all statistics quoted below are scaled down to comparable percentages.

The test pieces were played at the start of a course, lecture or lesson. Respondents were told `you will hear some12 pieces of music that have been used in film or on television. Please write down what you think is or ought probably to be happening on the screen along with each tune. There won't be to time to think, nor to write very much, so don't try to write sentences -- just jot down the words, pictures or impressions that come into your head.'

The results were collected in and reduced to single concepts (including important conjunctions and prepositions). Each concept from each respondent for each tune was interpreted as a `visual-verbal association' (abbreviated `VVA'). Thus, for the person who saw the first tune in connection with a girl in a white dress running slow motion through a meadow in nice summer weather as part of a shampoo advert, the separate VVAs were `girl', `white', `dress', `run', `slow motion', `through', `meadow', `summer', `nice weather', `shampoo' and `advert'.

It should be noted that this test procedure is one of free induction, not of multiple choice, i.e. that respondents had to actively construct a picture in their head which they then had to put into words. This means that there is a far greater spread of responses for each piece than would have been the case if a multiple choice method had been used. It also means that each answer has greater cultural and symbolic significance since each response was actively created with the music as sole stimulus and not with the aid of ready made alternatives. Despite the use of free induction and the consequent spread of responses for each test piece, Poisson analysis of the results showed extensive clustering of responses, this suggesting that there was a highly significant consistency of response and a very low degree of random spread over the various tunes. To this extent the results can be considered as reliable indicators of differences in mood and scenario between the ten pieces, as constructed by the respondents in the form of VVAs (verbal-visual associations) to solely musical stimuli in the test situation.13 Since the results and statistics of these reception tests are so voluminous, I shall concentrate here on a few general aspects of the material and on certain musical and visual-verbal stereotypes of particular anthropological interest.

Classification of visual-verbal associations (VVAs)

As already mentioned, free induction produces a wide spread of individual responses. This impedes the use of such statistical techniques as varimax rotation and, if the results are to be presented in a meaningful way, requires that responses be grouped together in larger semantic fields than those represented by each VVA. Thus, even some of the most common VVAs, such as `love' and `romantic' (only 4% each of the total sum of responses given to all ten tunes), are bracketed together into the same three-digit category ("Love, Kindness"). However, since `love' is a broader concept than `romantic love' and would therefore warrant a wider range of different musics than the latter, `love' and `romantic love' were put into two separate subcategories, both contained under the more general `love and kindness' section.14

There is no room here to present the extensive listing of VVA categories,15 let alone the taxonomic rationale behind its construction. However, three aspects of response classification criteria are relevant to the question of musical meaning and, by extension, to the anthropology of mass media music in Western society.

The first of these criteria consists of widely accepted systematisations of the functions of film music, more specifically those codified by Lissa (1965: 115-256).16 Particularly important symbolic categories influencing our17 VVA classification are, freely translated, what she calls `emphasising movement', `stylising real sounds', `representing location', `representing time' (of day, in history), `expressing psychological experiences' (of actors) and `providing empathy' (for the audience).18 The second major influence on our system of response categories comes from classifications of musical moods made by compilers of `Library Music' or of anthologies of music for the silent film.19 We are referring here in particular to such synoptic classifications as `animals', `bright', `bucolic', `children', `comedy', `danger', `disaster', `eerie', `exotic', `fashion', `foreign', `futuristic', `grandiose', `happy', `heavy industry', `humour', `impressive', `light action', `melancholy', `mysterious', `national', `nature', `open air', `panoramic', `pastoral', `period', `prestigious', `religious', `romance', `sad', `scenic', `sea', `serious', `solitude', `space' (cosmos), `sport', `suspense', `tenderness', `tension', `tragic', `travel', `water' and `western' rather than to episodic functions like `openings' and `closes' or to genre-related classifications like `classical', `rock' or `jazz'.

The third response classification procedure relies on musical common sense acquired as members of this culture and on some skill in hermeneutics. We had to ask ourselves whether a particular VVA would, according to our experience as musicians, musicologists, composers and listeners,20 warrant different music to that corresponding to any of the categories we had listed up to that point. For example, we had to split up cities and towns, putting `big city' and `small town' into separate categories because they are different in terms of musical symbolisation, although both verbally categorisable as `urban'. This corresponds with mood music catalogue classification of such pieces.21

Conversely, we had to put `suspicion' and `jealousy' into the same subcategory because we could sort out no clear musically distinguishing characteristics between the two, although their verbal meanings show important differences of nuance and content. This problem of musically determined VVA classification demands further discussion, since it has direct bearing on what film and TV music can and cannot communicate. That discussion might, hopefully, make a small methodological contribution to the cultural anthropology of our society.

Towns, suspicion and jealousy: verbal and musical concepts

Assuming that music is particularly suited to the direct symbolisation and communication of affective states and processes,22 it seems strange that such prosaic and concrete words as `town' or `urban' call for at least two sorts of musical accompaniment, whereas such clearly affective concepts as `suspicion' and `jealousy' remain musically indistinguishable. Let us see how such a contradiction between valid notions about music and valid musical practice can arise.

Suspicion and jealousy

Using verbal definitions rather than relating to musical delimitations for the time being, we shall take `suspicious' to mean `believing' -- usually that something is wrong -- `without having any proof' of it, while the most common meaning of `jealous' can be rendered, more specifically, as `suspicious or fearful of being displaced by a rival'.23 Taking these definitions a step further into the realm of the reasons for and consequences of these states of mind, it appears that fear may be a common denominator for both -- for suspicion as regards `something going wrong' and `no certain proof', for jealousy according to the dictionary definition. This means that the only affectively significant difference between the two words seems to be the aspect of rivalry in jealousy, for jealousy seems to entail suspicion -- though not necessarily if the jealousy is grounded in certain knowledge of rivalry --, whereas suspicion need not include any jealousy. If we opt for the clear verbally semantic distinction of `rivalry' between these two affective states that share a common denominator of fear and consider their possible portrayal in musical terms, it would at first seem proper to provide some thing slightly more active and aggressive e.g. quicker tempo, more sforzandi, irregular metre or periodicity, greater variation in dynamics, more discords for jealousy and less for mere suspicion.

On the other hand, suspicion can be more aggressive than jealousy without involving any jealousy at all according to dictionary definitions. Just recall any TV detective who suspects, but is obviously not jealous of, a Jack the Ripper whom he desperately wants to catch and towards whom he feels aggressive and in whose pursuit he expounds much energy. Such suspicion demands the same sort of music we just thought might distinguish jealousy from suspicion. This means that we are back to square one, suggesting once again that suspicion and jealousy both call for basically the same sort of music. This conclusion is also supported by affective problematisation of the word `jealousy', the rivalry stage of which might well have passed and turned into bitter and suspicious fear and resignation, such an affective state demanding music expressing a correspondingly resigned, controlled and static type of unrest. We would not be dealing here with the sudden action or overt aggression as that found in the more violent early stages of jealousy, but something, once again, more closely resembling the gnawing, inactive type of suspicion that we presumed to lie more in the dictionary definition of `suspicion' and less under `jealousy'. It should in other words be clear that the factor influencing musical interpretation of suspicion and jealousy is the degree and quality of potential or actual energy and aggressiveness pertinent to those states of mind, whatever their verbal definition and connotations may be. To summarise this matter, it is reasonable to argue that music can distinguish between at least two different affective states that might include feelings verbally denotable as `jealousy' and/or `suspicion', whereas words systematise these affective experiences in a totally different way (e.g. according to the `rivalry' element). It is for such reasons that we have put both jealousy and suspicion together with guilt, reasoning that these three warrant similar music, distinguishable from that suitable for other unpleasant but related states of mind, such as feelings of absurdity, futility and madness.

Small towns, big towns

These intrinsic differences between what we may call `musicogenic' and `logogenic' categories are also found behind the mechanisms that enable film and TV music to distinguish between large and small towns. Whereas the verbal concept of `town' denotes `a densely populated urban area' (in opposition to `rural'), film and TV music tend to be `rural' about small towns and `industrial' about big cities.24 Small towns are not necessarily more sparsely populated than large ones and do not necessarily contain less industry per head of the population or per square metre than large ones: they are in these respects merely smaller. From a logocentric viewpoint this should mean that small town music should be `smaller' than big city music. However, this would be a dubious musicogenic categorisation, for big-city-by-night music can be conceived in a quiet, slow and intimate solo-saxophone-cum-jazz-trio vein -- lonely gin at the bar and very sparse scoring --, while market day in a small town can call for a full string orchestra zipping tutti away at cheery forte semiquavers -- a veritable hive of activity and extremely dense scoring. Nevertheless, it is obvious to anyone brought up in our culture that the saxophone is more `big city' and the string orchestra more `small town'. In fact, the most significant musical differences between the two are not necessarily those of tempo, texture or volume but of the stylised use of instrumentation, melody, harmony, phrasing, etc. -- more jazz, rock or electronic for big towns, more classical or folksy for small ones. This means that operative parameters of musicogenic distinction between large and small towns are those that `represent historical time'25 or that can connote locations where particular genres are imagined to occur according to stereotypic musematic convention (e.g. jazz and rock night clubs on the one hand, country dancing in the town square on the other) and thereby connote the sort of people -- including their imagined or real affective attributes -- that frequent such locations (e.g. criminals, private eyes, prostitutes, factory workers on the one hand, local gentry, artisans or farmers on the other). In this way it is clear that, say, the four categories `loneliness', `bustle', `threat' and `hero' will ideally require eight different musical renditions if they are to occur in both a big city and a small town context.

Denotation, connotation, musicogenic, logogenic

The fact that `big city' and `small town' are words connoting complete constellations of time, space, people, moods, actions and gestures makes them far more musicogenic than `suspicion' and `jealousy'. These two words on the other hand, denoting certain characteristics of feeling and mood -- that which music is often deemed best suited to express --, rule out, by the very acts of verbal denotation and logogenic connotation to such musicogenically disparate categories as `rivalry', `fear' and `aggression', anything but the most general of affective common denominators; what remains is a sort of insecure and unresolved malaise that might just as well be part of guilt, boredom, nausea, illness, defeat, uncertainty, unfriendliness or loneliness. In this way music, at least that of title themes and underscore for film or TV, cuts reality into segments of experience that rarely correspond to those provided by verbal categorisation of the same experience. This in turn suggests that the anthropology of our own culture might draw considerable benefit from a discussion of the musical as well as verbal categorisations of human experience.

Of course, this is easier said than done, so, in addition to the discussion provided so far, let us see the third (and final) example of how semiotic studies of the structures and reception of film and TV music can contribute to an understanding of the musical systematisation of our relationship to social and natural environments (an anthropological ambition, if ever there was).

Women in film and TV music

To illustrate how musical symbols can categorise our experience of reality, let us return to the collective phenomenology of the reception tests described above, to the question of gender and more particularly to the musical categorisation of women.

The ten tunes

During the process of category classification sketched above, it became clear that our respondents had `seen' more male and less female figures in connection with some tunes, more female and less male in connection with others. The ten tunes used and their most common responses used were:

1. The Dream of Olwen (C. Williams). F major, andante, 4/4 (q =92), legatissimo, cantabile piano concerto with regular periodicity, classical/romantic harmony, Chopin left hand arpeggio sextuplets, melodic sighing sixths, sinuous melodic contour, long melodic phrases, regular resolution of suspensions, classical/romantic symphony orchestra, legato tied-over string padding. Plenty of romantic m 7-5 chords. A `nice' cross between a homely Warsaw Concerto (Adinsell), a subdued Variations on a Theme of Paganini (Rachmaninov), Schubert's An die Musik and Elgar's Nimrod.

2. The Virginian (P. Faith). E minor first section, regular periods, hexatonic Dorian or Aeolian folksy tune on clean Fender guitar with acoustic guitar playing 6/8 simplified son-style strum (q.=76), producing lively horseback effect. Bm and D chords used as dominants, also onbeat quavers plus crochets (resembling the inverted dottings of duple metre) in the melody. Second section even horsier in G major with string chords taking over the guitar strum figure. A cross between Apache (Shadows) and Bonanza (Livingston).

3. Monty Python's Flying Circus (J.-P. Sousa). Actually the march Liberty Bell. Full brass band, lively 6/8 march (q.=120) in F major à la Franz von Suppé's Light Cavalry. Includes a novelty-style `ding' on an unaccompanied tubular bells C as upbeat to the main tune.

4. Romeo and Juliet(N. Rota). Cover of title music for film version (1968). Slowish (q=84-88), poco rubato, legato, cantabile A minor 4/4 string lead with added ninth quaver oscillations in the piano part. Semi-classical, semi-modal harmony, including bass descent and Am-G-F-Dm-Em progression. Sighing sixths and short legato phrases in melody. Mainly strings. Short statements by solo flute and mandolin. Short, slightly livelier middle section in C major.

5. Sportsnight(T Hatch). Uptempo 4/4 (q=126), allegro non legato piece for brass, percussion and keyboards in C. Slightly syncopated and busy clavinette riff throughout. Melody is an almost one-note newscast or teleprinter rhythm stereotype on xylophone and trumpet. Chords of the eleventh supplied by brass accompanying riff in offbeat crochets. A cross between the themes for Hawaii 5-0(M. Stevens) or Magnum(Post) and either Shaft (Hayes) or Superstition(Wonder).

6. Emmerdale Farm(Hatch). Poco andante 6/8 (q.=72) idyllic pastorale music in D minor with circle-of-fifths progressions to F major and back. Classical harmony, moonlight sonata piano accompaniment (straight legato arpeggiato quavers) with no long suspensions but a precadential dominant pedal, oboe/cor anglais lead in wide but non-ecstatic, legato, cantabile melody. A cross between Les parapluies de Cherbourg, The Windmills of Your Mind(both Legrand) and the Pastoral Symphony from Händel's Messiah .

7. Sayonara(Waxman). Melodic and harmonic pentatonics in slow 4/4 (q=84) -- most note values crotchet or longer -- for legato and piano muted strings in and around B major/G #minor. No forehearable harmonic and little melodic direction. One or two pleading ascending melodic portamenti in first violins. Impressionist western view of a soft and misty orient.

8. A Streetcar Named Desire(North). Minor key jazz chords, lento and mainly 4/4 (q=58) with added (major) sixth and ninth, plunging trombones, plodding tonic bass crotchets, wholesale key shifts, considerable use of melodic and harmonic chromaticism, bebop flat fifths in melodic line played by strings, brassy jazz trumpet, clarinet glissando. Fades out over a West Side Story`rumble' riff (L. Bernstein). Whole piece contains elements resembling St James Infirmary, Harlem Nocturne(`Mike Hammer Theme', E. Hagen), minor key creole jazz and the themes for The Man With The Golden Arm(E. Bernstein), Perry Mason(F. Steiner) and The Untouchables(N Riddle).

9. Owed to `g'(Deep Purple -- instrumental). Never used in film or TV, at least so we thought!25a Rock hemiola (6/4 = 3x2/4 or 2x3/4), E minor, full-blooded, medium pace (q=120), heavy mid-seventies sound including extensive use of major ninth and sixth (detectives?). Cross between the James Bond Theme(Norman/Barry) and Purple Haze(Hendrix), but in 6/4 metre.

10. Miami Vice(J. Hammer). Semiquaver ridden 4/4 (q=116), uptempo, harmonically undirected (in A but ending in C), sequenced and synthesised collection of disco and blues-rock riffs with repeated note drive, Linn drums and Latin percussion. Self-consciously up-to-date at the time of its conception.

Statistics for male and female person VVAs for these tunes are shown in table 1.

Gender specific VVAs for human beings

Table not included: see PDF version

The first thing to notice about table 1 is that men were seen far more often than women connection in the reception test as a whole. Even though it might be objected that a mere ten pieces can in no way be regarded as representative of film and TV signatures as a whole, it should be pointed out that five of the pieces were connected to moving pictures starring women in one of the two main parts and that women occur visually in allthe other productions too, except for the Deep Purple number, which has never been used for a film or TV series and could, at least theoretically, have been connected visually to persons of either sex. From this it should follow that women account for half the gender specific human beings mentioned in connection with six out of ten tunes and, say, one quarter of the other four productions starring men. In figures, this means that 40% of the male-female person VVAs ought to have been female. Instead only 18.5% of gender specific human beings mentioned by our respondents are female.26 In any case, this decrease in female representation from 51% in reality to 40% in the films whose title themes were chosen for this study and, more specifically, the further decrease from 40% to 18.5% from presumed visual representation of females to those visualised in responses to the same music, raise important anthropological issues that could be of particular interest to feminist theories of mass media culture.27 The second point to observe in table two is that each tune has a gender profile. Four tunes (The Dream of Olwen, Romeo and Juliet, Emmerdale Farmand Sayonara) clocked up more female than male person VVAs, while all the others were predominantly male in this respect. In fact, male person VVAs for these other six tunes outnumbered their female counterparts by a factor of between 12 (Miami Vice) and 73 (The Virginian ), except in the case of Monty Python(a factor of 6) and A Streetcar Named Desire(only 4 times as many men). To simplify matters in this paper we shall restrict the remaining discussion to the four `female' tunes mentioned above and, as comparison, to the four tunes with an overwhelmingly male predominance (The Virginian, Sportsnight, Owed to `g'and Miami Vice), which, for the sake of brevity we shall call the `male' tunes.

`Male' and `female' music

There is no room here to describe any musical structures in detail, let alone to present complete transcripts of the pieces. However, let us in passing and in general review the musical traits that distinguish the four female from the four male tunes, referring to the short musical sketches provided above. Please observe the caveat that `female' and `male' refer in the following nine descriptive points to the four plus four test pieces just enumerated and not to any wider cultural definition of the words.

  1. That the average tempo marking for female tunes is slower than that for male tunes (83 against 109) gives no more than a vague impression of men being conceived as musically faster than women as far as the test material is concerned. This is because consideration of the number of tone beats performed in note values faster than that of the tempo markings is far greater for the male than for the female tunes, this suggesting that stereotypically male title tunes are probably perceived as even more uptempo (and stereotypically female signatures correspondingly slower) than their metronome markings seem to imply.
  2. Legato e cantabilemelodies and legatoaccompaniments, especially of arpeggio or tied-over string chord type are exclusive to female tunes, staccato phrasing and quick repeated notes to male ones.
  3. There are changes in overall volume, mainly in the form of crescendo and/or diminuendo in the female tunes. No such variation occurs in the male tunes.
  4. The bass line of male tunes is almost always more rhythmically and intervallically active than that of the female tunes.
  5. Male melodic lines contain more rhythmic irregularities (e.g. 6/8 quavers plus crotchets, syncopations, repeated notes) and shorter note values, but less by way of `normal' dottings and 6/8 crotchets plus quavers than female melodies.
  6. The first destination of melodic motifs in male tunes (first accentuated tone beat in a complete motif) tends also to be the highest in pitch, while the female tunes almost never do this. Instead, their melodic contours have either `up-and-back-down' or `down-and-back-up' or generally descending tendencies, sometimes including drops of a sixth that never occur in the initial motifs of male melodic phrases.
  7. The female melodic leads are played by piano (The Dream of Olwen, Emmerdale Farm), strings (Romeo and Juliet, Sayonara), flute and mandolin (Romeo and Juliet) and oboe or cor anglais (Emmerdale Farm). Male melodic leads are taken by electric guitar or guitar synth (The Virginian, Owed to `g', Miami Vice) and trumpet plus xylophone (Sportsnight). Only in the second half of The Virginiando strings carry a male melody and only the second 8 bar period of the electric guitar's initial 16 bars of the same tune is doubled by oboe. The Virginianis of course a Western and therefore has its historical location set much earlier than those of the other male tunes. All this seems to suggest that female figures (melodies and characters) are generally more `classical' and `serious' than the more `up-to-date' males (even a nineteenth century cowboy -- The Virginian--is partially electrified).
  8. Similar observations can be made about the orchestration of accompanying parts: the female melodies are all backed by strings, either on their own or boosted by piano or woodwind. All male figures (melodies) move in the accompanimental environment of strumming or riffing guitars, of punchy brass chords (Sportsnight) or of bubbling synths (Miami Vice). Once again, only the The Virginiancontains any string writing and all male tunes feature percussion28 (least in The Virginianand Sportsnight, most in Owed To `g'and Miami Vice). Feminine accompanimental environments are on the other hand totally percussionless.
  9. The tonal language of the female tunes is either classical-romantic (Olwen, Emmerdale Farm) or classically modal(Romeo and Juliet, Sayonara). All male tunes except The Virginian(folksy modal) are couched in the harmonic language of rock or diluted fusion (Sportsnight).

Table not included (see PDF version) hypothetical and stereotypical polarities of musical male and female

These musical descriptions should lead us to expect further connotations of male and female in the sort of directions shown in table. These hypotheses seem quite sexist, to say the least. It may therefore be wise to check such connotations by examining what other VVAs occur consistently or exclusively in connection with male and female tunes.

In a memo handed to me after working on gender aspects of this study, Anahid Kassabian reported:29

`I began... by looking through the statistics on the verbal-visual associations to find out which ideas are associated with male characters and which with female characters. Certain ideas, such as tranquillity, were predominantly associated with women, or strength with men. Others, such as quiescence, were exclusively associated with women (or weapons with men)'.

After making a raw list of such `male' and `female' associations, Kassabian soon realised that they did not all belong to the same cultural image.

`For example, tranquillity did not seem to belong to the same set of associations as danger. I organised the associations intuitively into four lists each for male and female, quickly realising that I was using character `type' and genre images.'

For female associations she called these lists: `Romantic Nature/Reflection', `Breakin' up is hard to do', `Marlene Dietrich or Lauren Bacall', and `Innocent Girlhood'. These associative categories aligned conveniently with mood music catalogue classifications and, she points out, `fit very neatly with various theories of gender difference and power relations', theories that were `developed to explain the same cultural stereotypes that our test pieces seem to communicate'. Kassabian continues:

`The three most obvious theoretical alignments were with anthropological theories of the nature/culture dichotomy, structuralist narratology, particularly Jurij Lothman's plot typology, and theories of the ideology of separate spheres. The nature/culture theory argues that women in the cultural imagination are connected with nature: reproduction, maternal instinct, `bodilyness', `earthiness', witches etc. Men are associated with culture: production, art, architecture, law, social order, technology, modern medicine, etc.'

Female/male tunes: nature/culture dichotomy. Table not included, see PDF version

Kassabian continues:

`The differences in the categories `rural' and `urban' are particularly striking; the study's respondents were absolutely sure, it would seem, that women belong in `green' settings while men are most at home in concrete. Quite clearly, the study's respondents associated women with nature and men with culture'.

After a short discussion of film theory and gender,30 Kassabian concludes that mobile characters (heroes) in traditional Western film narrative

[...] `must be men, since immobile characters, features of the topological space, are morphologically women. Consequently, men act, cross boundaries, etc. while women accept the hero's entry into and emergence from the space they represent.

This type of gender structure in film narrative is reflected in the VVAs provided by our respondents and set out in table X.

`In light of the responses in the categories reflection and dynamism alone, there can be no question that the study's respondents are sufficiently ideologically competent to understand that men move and act, while women sit by quietly and patiently'.

Narratological VVAs and female/male tunes. Table not included, see PDF version

The next classical gender distinction reflected by our responses is that between the private and public spheres. This distinction

[...] `is connected to gender ideology by the construction of a gendered division of labour and space. With the rise of the capitalist mode of production came the ideology of the separation of state and economy. Supporting this separation was the ideology of separate spheres: the public and the private. The public sphere includes free exchange of ideas (newspapers, town meetings, etc.) and social space (parks, town square, etc.) and is male. The private sphere, also known as the domestic sphere, is the home and all of its activities (food, clothing, child-rearing, affective values, etc.). It is female'.

Responses to our eight stereotypic film and TV signatures seem once again to agree with such a view of male and female characteristics. These ideological formations appear in table X.

Public/private sphere and female/male tunes. Table not included, see PDF version

Kassabian's comments to these responses are terse and to the point:

`The overwhelming association between women and love assures that affective values are still ideologically the domain of women; moreover, the strong connection between `maleness' of tunes and the public activities of the categories `festive' and `presentational' demonstrates that the public/private distinction is an operative one'.

Judging from this presentation, stereotypical film and TV music sorts our experience of gender reality into categories that are more traditional and conservative than those of visual narrative and far more so than any decent mainstream gender debate would permit. One obvious reason for this tendency is that we chose responses to the most obviously one-gender-orientated pieces of stereotypical film and TV music and ignored the male tunes in whose responses men did not so devastatingly outnumber women (Monty Python, Streetcar). In fact, A Streetcar Named Desireelicited more female VVAs (6.5%) than any other tune except Sayonara(7.5%) although its men outnumbered women by four to one (see table 2). Streetcar's women were, unlike the VVAs of the four female tunes, described either as women in black asking `got a light?' or as vamps, or, quite overtly, as prostitutes, all frequenting a club or bar, while Monty Python's females were drum majorettes, royalty and a cheerful soldier's girl. These female stereotypes nuance the picture to some extent, since their spaces and activities are different from those described earlier.

It might also be objected that all the test pieces are musically speaking extremely stereotypical and this will make their musical categorisation of social and natural realities correspondingly stereotypical and conservative. Indeed, one important criterion behind the selection of the ten pieces was that of semiotic and musicological methodology: just as there would be no point in trying to construct a semiotics of modern language on the basis of Shakespeare's sonnets, avant-garde poetry or slang, it would also be a dubious intellectual strategy to describe basic elements of musical signification in our culture on the basis of Beethoven quartets, avant-garde minimalism or hip-hop. For this very reason, then, the ten pieces were chosen partly becausethey were perceived before the test situation asstereotypical.

However, although there is much validity in this objection, it does not refute the hypothesis that musicogenic gender categories may well lie at a deeper level in the public conscious, symbolic levels that may possibly be more conservative -- maybe more archaic -- than visual or verbal representations of the same chunk of reality. This argument, already supported by the fact that responses to the music included less than half the number of women seen on the screen in connection with that music, can also be verified by analysis of films and TV series in which women are the main figures (`goodies' and/or `baddies') and in which they drive the action. For example, the Cagney andLacey theme (Conti 1981), though scored with considerable `big band = big city' drive, includes as its only conceivable female ingredient a phrase played by two saxophones cackling in parallel thirds.31 Neither Angie Dickinson as Police Woman,32 with her music resembling a rhythmically simplified minor-key version of the Starsky and Hutchtheme (Scott 1975), nor the leading Laura of Remington Steele(Mancini 1982), with her intimate sax and vibes variant of classical Hollywood minor-key jazz detectives like Mike Hammer(Hagen 1941), are specially convincing either: they both sort melodically under the `high-heeled city lady', `Marlene Dietrich/Lauren Bacall/Sadie Thompson' vamp category.33

It is even more surprising to discover similar trends when turning to music for heavier feature films sporting screenplays that seem to have taken deeper notice of the gender debate. For example, the leading roles of Black Widow are both female: the murderess is an attractive, clever, devious and worthy opponent to the courageous, humorous, equally attractive and clever (though in other ways) female detective-cum-journalist. Apart from the obvious visual differences, the only main genderdifferences of screenplay between this film and any decent male dominated multi-murder thriller are (a) that murder victims are poisoned rather than blown, hit, cut or mangled to pieces, (b) that, Humphrey Bogart notwithstanding, both murderess and heroine are portrayed as more credibly vulnerable human beings than most of their male counterparts. Nevertheless, both female characters show great ingenuity, determination and daring and the heroine acts in a truly heroic fashion. Of course, Michael Small's music reflects the different character of this film: there are no heroic horn calls à la Star Wars, no Superman-like trumpet blaring, because our heroine does not charge around on horseback, nor in a fast car, helicopter or space ship beating the living daylights out of slobs. On the other hand, she is an extremely strong character with definite goals, chasing her prey with both a sense of both justice and of divine retribution. A man with such non-gun-slinging characteristics (e.g. a young Maigret or philosophical Sam Spade) would almost certainly have warranted a thoughtful but melodically strong leitmotif, i.e. something along the lines of the theme for Hassel (Neglin 1989) or Philip Marlowe(Matlovsky 1989). However, Small provides the Black Widow's heroine with no such musical identity. For this reason she comes over as a much weaker person than she is in the verbal/visual narrative, almost as though it were affectively inconceivable for women to act with such determination and courage. The psychological thrust of the screenplay, showing that vulnerability does not imply weakness -- on the contrary --, is lost by omitting to give the heroine any clear musical (directly affective) identity. Although she `drives the action', she has no driving melodic theme.

Conclusions

It should be clear from this short discussion of gender in title themes and underscore that such music can sort our experience of the visual and verbal into constellations of experience that hang together musicallyin a totally different fashion to that prescribed by our culture's visual or verbal symbols. For instance, instead of accounting for the `male' and `female' in such music, we could have discussed musical categorisation of other broadly connotative concepts, some of which -- like `Hero', `Nature', `Time' or `Death' (Tagg, 1979:123-136; 1982; 1984; 1990) -- might even be anthropologically basic enough to qualify as `archetypes' (Jung, 1964: 56,ff; Henderson, 1964:101-119). Crudely generalising the findings of such discussions of musically semantic fields, it seems that the musically defined heroes of our mass media culture seem often to commit hybris with impunity, that our musically defined concept of `nature' does not include man, that `time' is inextricably linked to the machines and metronomic social order in which we have enslaved ourselves and that `death' is only a matter of individual sorrow and bereavement, not a question of social or natural renewal and continuity. If this is so and if, as we have suggested, the portrayal of `Woman' in title music is as conservative as our respondents seem to believe musically (less so visually and even less so verbally), then it appears that music in our culture, its digital technology notwithstanding, can categorise shared subjective experience of and relation to our social and natural environment at deeper, possibly more `archaic', levels of consciousness than visual and, more notably, verbal symbols. This argumentation is supported by simple anthropological and bioacoustic observations of how nonverbal, yet meaningful, sound plays a central part in learning processes during the most formative stages of a human's development (from minus five months to two years -- the preverbal phase). The argument is further supported by observing how music functions in the construction of peer-group identity: neither classical devotees with their `absolute music', nor bebop freaks and their improvisatory nonverbality, nor `real' rebel rockers and their overt anti-intellectualism want words to invade the harbours of non- analytical well-being they have built as collective bastions against all the facts, figures, words and numbers thrust upon them by school, work, money, statistics, politics and all those other embodiments of social power. Such asylums of nonverbal symbolism may be psychosocial necessities in a culture whose ideology of knowledge so one-sidedly invests certain symbolic systems, notably numbers and words, with great power and status as legitimate carriers of knowledge while banishing others to the freaky realms of `Art' or `entertainment' -- the fact that this presentation about musicis mostly wordsillustrates that point quite clearly -- but they do not help solve the difficulties that made them necessary. This is where musicology, `ethno' or not, can be of some use.

Verbal discourse on music can either strengthen or weaken our culture's schizophrenic ideology of knowledge. One popular way of strengthening it is to canonise some musics and deprecate others (`classical' against `popular', `folk' against `art' and `commercial', jazz against rock, rock against disco, etc.). This canonisation strategy can involve mystification by exclusive concentration, for example on extensional complexity for `classical', on authenticity for `folk', on spontaneity for jazz, on street credibility and intensional complexity for rock. Another, complementary tactic is to study music made far away in time, space or social habitat, without relating that music to our own social and cultural realities. Without such relation, studies of the far away or long ago can all too easily become idealistic or escapist projections into imaginary paradises of either lost innocence -- `field work' on the `noble savages' of natural or man-made jungles -- or of artistic wisdom and skill -- Beethoven, Parker or Hendrix as gods. These strategies constitute the easy way out and the safest bet for a comfortable but boring career in musical academe. Unfortunately, however, the situation is more serious than these strategies seem to suggest.

Music is an almost omnipresent and ubiquitous phenomenon in this age of the rock video, the soap opera, video rental, CDs, Muzak and TV advertising. This means that meanings and ideologies are constructed and communicated in the audiovisual media in visual and musical rather than in chiefly verbal categories to a far greater extent than ever previously in the history of our culture. At the same time, our ideology of knowledge seems to tell us that all that music is `entertainment' (only occasionally `art') and cannot really communicate any `knowledge'. This dilemma and contradiction forces music scholars to choose sides.

Jung (1964:92), writing about dreams, puts the matter quite simply:

[...] we still complacently assume that consciousness is sense and the unconscious is nonsense.

That dictum might just as well apply to music and its pre- or subconscious levels of affective symbolisation as to dreams and the unconscious. If Jung's observation bears any weight, we are in a sorry state, for although it may seem easier (or more complacent) to abandon the task of trying to explain how music communicates ideas and ideologies in this culture, just because our tradition of knowledge presents that as the easiest way out, it is doubtful whether that is the wisest choice to make.

In view of this argumentation and with reference to the short summary of research problems presented, it no longer seems quite so facetious to suggest that ethnomusicology might trade a little `ethno' for a bit more `musico'. For what is needed is not so much an ethnology of other musics as a musicology of this ethnos . This means we need to know how music makes us think and feel about different sorts of people (e.g. cowboys, Indians, men, women, criminals, the Americans, Russians, communists, fascists, rockers, yuppies), places (e.g. space, the countryside, the sea, towns, cities, jungles, Antarctica, Spain, Japan, slums, suburbs), actions (e.g. war, peace, making love, driving, dancing), objects (e.g. atomic power stations, knives, cigarettes, shampoo), etc., etc. We also need to know what music can say about feelings like love, hate, suspicion, jealousy, power, weakness. How we shall ever be able to understand our own society and what it does to ourselves and to the rest of the world without a profound understanding of how its music communicates ideas and attitudes is beyond my ken. I just hope that musicology, ethno or not, will be able help in this Sisyphean task.

Verbal references

Alencar PintoGuilherme de (1985). Análisis: `Deus lhe pague'. La del taller, 2/1985:19-26.

AyesteránLauro & F. de M. R. de (1965). El minue montonero - danzas, canciones e instrumentos del pueblo del Uruguay. Montevideo.

BjörnbergAlf (1984). `There's Something Going On - om eolisk harmonik i nutida rockmusik'. Tvärspel: 371-386. Göteborg.

-- En liten sång som alla andra. Melodifestivalen 1959-1983 . Göteborg.

chaseGilbert (1958) A dialectical approach to music history. Ethnomusicology2:1-9

GorbmanClaudia (1987). Unheard Melodies. Bloomington.

HendersonJoseph L. (1964). Ancient Myths and Modern Man. Man and his Symbols: 95-156. New York.

HoodMantle (1957) Treaining and Research Methods in Ethnomusicology. Ethnomusicology Newsletter, 11:2-8

JulienJean-Rémy (1987). Défense et illustration des fonctions de la musique de film. Vibrations, 4:28-41.

JungCarl G. (1964). Approaching the Unconscious. Man and his Symbols: 1-94. New York [1968].

KarlinFred, WrightRayburn (1990). On the Track. A Guide to contemporary film scoring. New York.

KarbušickyVladimir (1986). Grundriss der musikalischen Semantik . Darmstadt.

-- (1987). The index sign in music. Semiotica, 66-1/3:23-36.

KunstJaap (1959). Ethnomusicology. Den Haag.

LingJan (1978). Hjort Anders' `12th Day March' - Stylistic and ideological change. Antropologiska Studier, 25/26: 122-139.

LingJan, RamstenMärta (1985). The Gärdeby Folk Melody: A Musical Migrant. Popular Music Perspectives, 2 (ed. Horn): 119-141. Göteborg, Exeter.

LissaZofia (1965). Ästhetik der Filmmusik. Berlin.

MerriamAlan P (1964). The Anthropology of Music. Northwestern University Press.

NettlBruno (1964). Theory and Method in Ethnomusicology. London.

RösingHelmut (1977). Musikalische Stlisierung akustischer Vorbilder in der Tonmalerei. München & Salzburg.

-- (1978). Fast wie ein Unwetter. Zur Rezeption von Pastoral- und Alpensinfonie-Gewitter. Festschrift W. Wiora: 65-89. Tutzing.

SkogInge (1977). Musikantropologi. Sohlman's Musiklexikon , 4: 617, ff.

TaggPhilip (1979). Kojak: 50 Seconds of Television Music . Göteborg.

-- (1981) On the Specificity of Musical Communication. Göteborg: Stencilled Papers from the Gothenburg University Musicology Department, n§ 8115.

-- (1982) Natur i massmediemusik. Naturen som symbol(ed. Allwood et al.): 161-199. Stockholm.

-- (1984). Understanding `Time Sense'. Tvärspel: 21-43. Göteborg.

-- (1987). Musicology and the Semiotics of Popular Music. Semiotica , 66/1-3: 279-298.

-- (1990). `Universal' Music and the Case of Death. La musica come linguaggio universale(ed. Pozzi): 227-266. Firenze.

TarastiEero (1978). Myth and Music - A Semiotic Approach to the Aesthetics of Music, especially that of Wagner, Sibelius and Stravinsky . Helsinki.

-- (1985). Music as sign and process. Analytica: 97-116. Stockholm.

VegaCarlos (1944). Panorama de la música popular argentina . Buenos Aires.

Musical references

AddinsellRichard (1942). Warsaw Concerto. London: Keith Prowse.

ArmstrongLouis & All Stars (1959). St James Infirmary. The Jazz Story(n.d., c. 1970)

BerliozHector. Symphonie fantastique. New York: W W Norton.

BernsteinElmer (1955). The Man With the Golden Arm. MCA Coral ORL 8280 (1959).

BernsteinLeonard (1957). West Side Story. New York: Chappell.

ContiBill (1981). Cagney and Lacey-- end title (TVM).

Deep Purple (1975). Owed To 'g' (Bolin). Come Taste The Band . Warner Brothers PR 2895.

ElgarEdward (1899). Nimrod. The Enigma Variations, Op.36. London: Novello.

FaithPercy (1962). The Virginian. Golden Hour of Favourite TV Themes. Golden Hour GH 845 (1976).

GershwinGeorge (1925). An American in Paris. George Gershwin - Rhapsody in Blue. DGG Privilege 427203-2 (1997).

GrainerRon (1958). Title theme for series Maigret. BBC TV.

HagenEarle (1941). Harlem Nocturne. New York: Shapiro. Also as recorded for the Mike Hammerseries (CBS Columbia TV, 1983).

HammerJan (1984). Miami Vice Theme. Miami Vice. MCA 252 493-1 (1985).

HändelGeorg Friedrich (1741). The Messiah. London: Novello (1902).

HatchTony (1974). (a) Emmerdale Farm; (b) Sportsnight. Hit the Road to Themeland. Pye NSPL 41029.

HayesIsaac (1971). Theme from Shaft. Stax S45.

HendrixJimi (1967). Purple Haze. Are You Experienced?Polydor RS 6261 (Benelux).

HermannBernard (1960). Psycho(colonna sonore originale). RCA Cinematre NL 33 224 (1975)

LegrandMichel (1966). Les Parapluies de Cherbourg. Philips 72233.

-- (1973). Windmills Of Your Mind. Movie Memories. Music For Pleasure MFP 50438 (n.d.).

LivingstonJ. (1959). Theme from «Bonanza». Golden Hour of Favourite TV Themes. Golden Hour GH 845 (1976).

ManciniHenry (1982). Remington Steele- main title (TVM/NBC/MTM).

MatlovskySamuel (1989). Philip Marlowe, Private Eye -- TV theme. Jon Slam Music.

NeglinAnders (1989). Roland Hassel- theme for Swedish TV detective series.

NormanMonty, BarryJohn (1962). The James Bond Theme. Big Screen Hits of John Barry. CBS S CBS 31682 (1980).

NorthAlex (1951). A Streetcar Named «Desire». Fifty Years of Film Music. Warner Brothers 3XX 2736 (1973).

PostMike (1982). Magnum P.I. Theme. Television Theme Songs . Elektra K52372.

RachmaninovSergei (1934). Rhapsody on a theme of Paganini. Big Concerto Movie Themes. Music For Pleasure. MFP 4261 (1972).

RapéeErnö (1924). Motion Picture Moods for Pianists and Organists. New York (reprint 1974).

RiddleNelson (1959). Main theme from The Untouchables. (ABC/Desilu TV).

RotaNino (1974). Romeo And Juliet. Hit the Road to Themeland . Pye NSPL 41029. {59-60}

SchubertFranz (1816). An die Musik. Ausgewählte Lieder . London, New York, n.d.

ScottT. (1975). Theme for «Starsky & Hutch» (TVM/ABC). Covered on Police Movies.Cinevox CIA 5027 (Italia, distr. Ricordi), 1981.

SedakaNeil (1962). Breaking Up Is Hard To Do. RCA Victor 47-8046 (USA), RCA 1298 (UK).

ShadowsThe (1960). Apache. Columbia DB 4484.

SmallMichael (1987). Music for The Black Widow (20th Century Fox). CBS Fox Video (1988).

SousaJean-Philippe (1972). «Liberty Bell» - Theme for «Monty Python's Flying Circus». The World of TV Themes. Decca SPA 217.

SteinerFred (1957). Theme from Perry Mason(CBS/Paisano TV, 1957). As re-recorded with original scoring for comeback series Perry Mason Returns(NBC/Viacom/Intermedia, 1985).

StevensMort (1963). Hawaii 5-0. Golden Hour of Favourite TV Themes. Golden Hour GH 845 (1976).

SuppéFranz von (n.d.). Light Cavalry. Used as «The Mounties Musical Ride» in The Benny Hill Show,Vol 2. Thames Video [VHS] (1978?).

WaxmanFranz (1957). Sayonara - main titles. Fifty Years of Film Music. Warner Brothers 3XX 2736 (1973).

WagnerRichard (1856). Die Walküre. Philips 6769071 (1981).

WilliamsCharles (1944). The Dream Of Olwen. Big Concerto Movie Themes. Music For Pleasure. MFP 4261 (1972).

WonderStevie (1972). Superstition. Talking Back. Tamla Motown 1C 062-93880.


1. See Tagg 1979, 1987. `Semantics: the study of relationships between signs and symbols and what they represent' (New Collins Concise English Dictionary, London 1982). This is very similar to the Saussurean definition of la sémiologie: `science qui étudie la vie des signes au sein de la vie sociale' or `science étudiant les systèmes de signes' (Le petit Robert, Paris 1970), the same source defining la sémiotiqueas `théorie générale des signes'. As Karbušicky (1986, 1987) suggests, semanticsis probably the best word to use when talking about music and we shall be using it according to the (Peircian) definition offered above.

2. e.g. Alencar Pinto 1985; Björnberg 1984, 1987; Ling 1978; Ling & Ramsten 1985; Rösing 1977, 1978; Tarasti 1978, 1985.

3. For more detailed critique, see Tagg 1987.

4. Project financed by the (Swedish State) Council for Research in the Humanities and Social Sciences.

5. e.g. interobjective comparison and hypothetical substitution (commutation), see Tagg 1979, 1987.

6. Kunst, Hood and Chase are all quoted in Merriam, loc. cit.

7. For definition, see Tagg 1979:20-28.

8. Skog distinguishes two main schools of thought on this matter: `enl den ena'... `studeras vissa typer av musik, fr a utomvästerländsk musik och folkmusik'... `Enl den andra huvudinriktningen utmärks musikantropologin av sitt speciella synsätt på musik; karakteristiskt för detta är att musiken ses i sitt socio-kulturella sammanhang, och därmed är även tex västerländsk konstmusik ett tänkbart studieobjekt'.... We might call representatives for these two schools the `traditionalists' or `conformists' and the (musically) `cultural anthropologists'.

9. Without such a view it would, curiously enough, be impossible to legitimise important work on mesomúsica, such as that carried out by Vega (1944) or Ayesterán (1965).

10. Criteria for selection of test pieces were quite random. A few of them happened to be pieces on an anthology of TV tunes I had acquired in 1978 and had already used in analysis classes. Others were chosen for crosschecking purposes. For example: how well could people distinguish between a romantic and tragic love story, between homely and lonely pastorale idylls, between detective and bitter-sweet minor triads with added ninths? Methodological criteria are mentioned later on in this article.

11. At state schools, sixth form colleges, music schools, weekend courses and in adult education between autumn 1979 and spring 1984.

12. `Some' -- the number of pieces played depended on the time available for the test situation. See previous paragraph in main body of text.

13. Of course, the test situations were, as usual, unsatisfactory. Sitting in a classroom or lecture hall is not the same as lounging in front of the TV. However, since respondents had so little time to think or indulge in any time-consuming reflective activities -- the examples were short and came in rapid succession -- and since only 7% of answers to all tunes were blanks, it is probable that the answers chiefly constitute immediate respondent associations to the musical stimuli, associations learned unconsciously (yet cognitively) as a result of many hours audiovisual training in the presence of music and moving pictures.

14. The sort of low level classification distinguishing `love', `in love', `requited love' from `lots of love', `romance', `very romantic' (different music for those two areas of experience) gives rise to two separate four-digit categories (1111 and 1112) which come under the more general three-digit heading 111 ("love, kindness"), a less musicogenic category that also includes such subcategories as `sensual', `seductive' (1113) and `gentle', `tender' (1117), both musically distinguishable from the other. The three-digit category `love, kindness' sorts in its turn under the even less musical two-digit heading `culturally positive' (11) which also includes `tranquillity and security' (112), `joy and festivity' (113), `beauty and attraction' (114), etc., all of which contain their more musically determinable subcategories. The two-digit category `culturally positive' (11) is in its turn part of the highly unmusical main one-digit heading `general attributive affects' (1) which also includes the `culturally ambivalent' (10), `culturally negative' (12) and `culturally neutral' (14). The other one-digit headings are (0) `test statistics and time positions', (2) `beings -- human, mythical, animal --, props and gatherings', (3) `location, scene, setting', (5) `explicit space-time relations, movements, acts, interaction', (8) `media immanent' and (9) `evaluative'.

15. 798 four-digit categories, 212 three-digit ones, 46 two-digit ones and 7 one-digit categories were used to classify the 8552 responses collected (2472 unique). The hierarchical systematisation of one-, two-, three- and four-digit categories classifying the 2472 unique responses occupies 25 pages on its own and the actual response results and statistics account for another 150.

16. For discussion of other film music function classifications, see, for example, Julien (1987:28-41), Gorbman (1987), Karlin & Wright (1990).

17. The `we' of that `our' are Robert Clarida (New York), who, as visiting Fullbright scholar, worked with me on this project, and the author of this article.

18. Translated from `Unterstreichung von Bewegungen; Stlisierung realer Geräusche; Representation der dargestellten Raums; Representation der dargestellten Zeit; Ausducksmittel psychischer Erlebnisse; Grundlage der Entfühlung'. This, in turn, has been translated into German from the Polish original. However, judging from Lissa's commentary to these functions, the English captions seem reasonable, even though `empathy' is Einfühlung rather than Entfühlung, because the `unfolding' of a particular mood or feeling (Entfühlung) by music to the audience is no more than an inverse dialectic complement to the audience's process of empathisation (by music) with particular aspects of paramusical film narrative.

19. By `Library music' or `mood music' is meant those productions, consisting of several hundred phonograms and an extensive catalogue, in which recorded music is classified under headings for use in such audiovisual situations as advertising and low-budget film, TV or video. Mood classifications have been assorted from the following productions: Boosey and Hawkes Recorded Music for Film, Radio and Television (London), Bruton Music (London), CAM (Creazioni artistiche e musicale, Rome), The Conroy Recorded Music Library (London), The De Wolfe Recorded Music Catalogue (London), KPM Music Recorded Library (London), Éditions Montparnasse (Paris), Selected Sounds Recorded Music Library (Hamburg), The Southern Library of Recorded Music (New York) and The Major Mood Music Library (Thomas J Valentino, Inc., New York). The silent film mood classifications stem mainly from Rapée (1924).

20. `We' are Robert Clarida (New York), classically trained guitarist and composer with considerable practical experience from the worlds of jazz, rock and disco, and the author, classically trained organist, erstwhile jazz, blues and rock musician, with some experience as a composer of choral works, one musical, several rock numbers and a few TV signatures.

21. For example, Boosey and Hawkes put the small town pieces Elm City USA, Fillmore Junction Before 1885 and Market Day in Martinique into their `Pastoral/Scenic Grandeur' section, whereas Metropolis, Oxford Circus and Rio Rhythm are classified as `Industrial'.

22. Tagg 1981.

23. All verbal definitions according to The New Collins Concise English Dictionary, London 1982.

24. See footnote 22.

25. `Representation der dargestellten Zeit' (Lissa: 1965:115, ff.) - see footnote 16.

25a. Update, January 2008: David Dean (London) just informed us by email that the Deep Purple tune was in fact used as outro music to the BBC TV spy series Quiller (1975)!

26. The 40% is calculated as follows: if six films out of ten co-star a man and a woman in equally important roles, that means that (6x10)2 = 30% of gender specific person VVAs should be female. If women are visible for only 1/4 of the time in the other four productions, this means that another (4 x 10)4 = 10% of such VVAs should also be female: 30 + 10 = 40. The 18.5% is calculated as follows: the male percentages in table 2 are summed to 134.0, the female ones to 30.4, making 164.4 in all. 30.4 is 18.5% of 164.4.

27. This is even clearer from the scores for the following male and female person VVAs with their total% occurrences (in brackets) and their order of frequency as type of response [in brackets]: one male (9.96) [2], several people (9.05) [5], several males (4.29) [20], one person (3.93) [23], couple (3.14) [28], one female (2.23) [40], many people (1.34) [66], several females (0.77) [94], two persons (0.73) [98]; two males only clocked up 0.13% and there were no two females. According to Kassabian, such responses seem to follow the cultural patterns that feminist scholars like De Lauretis have described in her critique of Lothman's typology, i.e. that human beings are defined men, and everything else is `not even `woman', but not-man, an absolute abstraction'. Kassabian writes in a memo on this reception test material: `By a huge margin, the study's participants found it easier to imagine a single individual as specifically male, and in general found it difficult indeed to picture specifically female people in the scenes they imagined. This is, of course, related to the `men act and women sit' distinction; humans (men) are agents of narratives, individual/individuated subjects, and consequently easier to imagine'. (Text supplied on diskette to the author).

28. `Percussion' in the everyday, popular music sense of the word, i.e. excluding the piano, vibraphone, etc.

29. Anahid Kassabian, postgraduate student in cultural studies from Stanford University, assisted me with gender aspects of this research while on a scholarship to Göteborg in the autumn of 1988. All the indented text on the next few pages is by her, unless otherwise stated.

30. Kassabian quotes Juri Lothman and Teresa de Lauretis but gives no further details of reference.

31. Transcript in Karlin & Wright (1990:479).

32. This is one occasion on which I failed to note the composer's name. Police Woman (NBC/Columbia) ran between 1975 and 1979. The 1975 theme is probably by Mort Stevens (of Hawaii 5-0 fame) or Jerold Immel (of Macahan and Dallas fame).

33. This subgenre and its connotations are discussed under the Streetcar named Desire analysis in Tagg & Clarida (forthcoming).