Article MT159

Cantometrics:

Song and Social Culture

A Response

I found this essay by Fred McCormick only recently, though it apparently dates back to the '90s.  I was impressed by his knowledge of so many aspects of Cantometrics as well as his many insightful criticisms.  As I'm now undergoing my own personal reconsideration of this methodology, I decided it would be interesting to present a response, however belated.  My interleaved comments are in boldface.

Victor Grauer - 24.7.05


Musical Traditions No 12 contained a review of Alan Lomax's book, The Land Where the Blues Began, in which the author touches on the topic of cantometrics.  As your reviewer seemed somewhat uneasy with the concept, I thought he and other MT readers might appreciate a nonspecialist word of explanation.

Cantometrics arose out of 'The Cross Cultural Study of Expressive Style', a multi-disciplinary research programme, formed under the direction of Alan Lomax in 1961.1.  The study was lodged at Columbia University until 1982.  It was later transferred to Hunter College, New York.1 The term refers to a system for the measurement of singing style which had been devised jointly by Lomax and musicologist Victor Grauer, to test a number of hypotheses formed by Lomax during his collecting work.  These were; that the dominant values of societies fundamentally influence how their members sing; that this influence applies wherever folksong is found in its natural state; and that the world distribution of singing styles is patterned on the world distribution of human societies.

The survey method was as follows.  First, 37 style elements were identified, covering solo vocal technique plus related features such as group singing and accompaniment.  Next a sample of 2,527 recorded songs, from 233 cultures, was prepared.  Each recording in the sample was analysed aurally by a pair of research workers who recorded their impressions for each applicable element on a rating scale of 1 to 5.

Some of the scales were 1 to 5, others 1 to 3, still others were larger, as many as 1 to 13.

The results were then compared statistically with the cultural traits of the societies they represented, from which conclusions were drawn about the relationship of singing styles to social norms.  Further research has involved the sample's being widened to 4,000 songs and 400 cultures, extending and reinforcing the team's initial conclusions.  It also laid the basis for development of the data into a multimedia software survey of the world's musical and choreographic cultures, which Lomax calls The Global Jukebox.

However, given the wide parameters on which the original study was set, it is not surprising that the findings confirmed initial hypotheses, or that further conclusions were reached.  Firstly, style is significantly affected only by certain cultural elements: i.e., subsistence type, political structure, sexual conventions, modes of social order and complexity of class structure.  Secondly, these elements affect song cross-culturally.  That is, where otherwise dissimilar societies resemble each other in respect of these elements, they will also resemble each other in respect of singing style.  Thirdly, individual cultures share features with their neighbours, which merit their grouping into much larger regions: Eurasian, African, North American Indian, etc.  Finally, singing varies with social structure, solo singing and unified choirs being found in centralised, cohesive societies, while leaderless group singing and diffuse choruses identify egalitarian groups and individualised cultures.

Not quite.  Solo singing and diffuse or poorly coordinated group singing were allegedly characteristic of highly centralized societies.  Highly integrated group singing, featuring interlock or overlapping call and response, was allegedly found in egalitarian, highly cohesive societies.

The above only scratches the surface of the complete programme, which includes equivalent studies of dance (choreometrics) and folksong texts.  The full story can be checked out by reading the survey team's writings on the subject, but unwary browsers should be warned that they could be in for a stiff and none too well organised read.2.  Principal works dealing with world singing styles are: 
Alan Lomax (ed.), 'Folk Song Style and Culture' (Transaction Books, New Brunswick: 1968)
Alan Lomax, 'Cantometrics: An Approach to the Anthropology of Music' (University of California Extension Media, Berkeley: 1977)
Alan Lomax, 'Folk Song Style' American Anthropologist, LXI (1959).2  Nowhere could I find a comprehensive summary of the central argument, or what criteria were applied in selecting the survey data, or any attempt to say what is meant by such core terms as 'culture' or 'folksong'.  Clarification of this latter term is particularly important here as the programme was dependent on donations of sound recordings from archives and individuals all over the planet.  Since usage of the word varies at least as widely as the sample data sources, do we have any certainty that the researchers were comparing like with like?

Some of Lomax's presentations could certainly have been clearer and better organized.  As far as 'folksong' is concerned, however, I don't think any special importance should be given that term in this context.  You are correct, it is certainly problematic, but it would be a mistake to read too much into its usage in the Cantometrics literature.  We were interested, essentially, in what could be called 'traditional' music and above all traditional performances, whether from the 'folksong' or 'ethnic' or 'tribal' repertoires it didn't really matter.  And one of the things we were able to demonstrate, I think, is that there are no hard and fast boundaries among these three areas.  We were not, however, interested in 'folk music' as performed by non-native musicians, not necessarily because there is necessarily anything wrong with such a practice, but because Cantometrics is the study of (authentic) performance style, NOT 'songs' per se.

While I can claim some familiarity with arguments which relate to culture and singing style, appraisal of the statistical method on my part has to be limited and subjective.  However, although the number of cultures sampled seemed sufficiently large and diffuse to represent the entire planet, and the style features analysed all significant aspects of song performance, I felt that the sample of ten songs per culture was too small to be reliable.

Lomax himself responded to this and other objections in his essay 'Factors of Musical Style' [in Theory and Practice: Essays Presented to Gene Weltfish, ed. Stanley Diamond, The Hague: Mouton, 1980, pp. 32-33].  My own view is basically consistent with his, but I'd like to take this opportunity to go into a little more detail.  The sampling issue has been raised many times, but the sampling problem has rarely been understood.  Let's say you have 1,000 recordings from a particular group and you extract a random sample of 100 from that.  That might look good on paper, but if the original 1,000 recordings aren't representative then the random sample will inherit that deficiency and its value will be equally questionable.  So it isn't the size of the sample or its randomness that's important.  What's important is whether or not it's representative.  Wherever possible Lomax, therefore, requested that the collectors themselves choose representative selections for our sample on the basis of their own experience with the role of music in the culture.  Is this a perfect method?  No, because clearly some were better informed than others and even the best informed might have missed something.  Could the method be improved?  In principle yes, but in practice no, because there is simply no way to determine for sure whether any sample is perfectly representative.  Which is one reason why Cantometrics must be treated as a heuristic, rather than a purely objective, method.  (More on this below.)

Why only ten?  Because, as Lomax has written, 'we found that we were recording little new information when we coded more.' (ibid., p. 32)  Actually a sample could be even smaller  than ten and still be representative, if the style of the group in question were sufficiently homogeneous.  And in many cases the sample was considerably larger than ten.  What must be remembered is that Cantometrics is a measure of performance style, not the structure of individual songs.  The songs themselves can be extremely diverse in many ways, but performance style usually contains many redundancies.

A potential problem does arise if the sample isn't sufficiently homogeneous.  You might, for example, have a sample of songs from a particular area which range all over the place, solo, group, embellished, unembellished, polyphonic, unison, heterophonic, etc., depending on where they were recorded and by whom.  In such a case a sample of 100 or even 1,000 songs wouldn't improve matters, because the resulting 'modal profile' (a summary of all codings from a particular place on a single chart) wouldn't represent any of these styles adequately.  What we usually tried to do in such a case was not make the sample larger (no help there) but break it down into smaller subsamples of at least ten songs each, one from one village, say, and another from another village.

Which is not to say we had no sampling problems.  In some cases we didn't have access to the original collector and had to go only by what was available on a particular disk or tape.  In other cases, fortunately quite rare, we were forced to settle for only a few samples since no more were available.  And in any case there will always be a problem in determining accurately and meaningfully whether any sample, no matter how large or how carefully chosen is really, truly representative.  This is a problem anyone doing ethnomusicological or folkloric research is faced with.  Which is why all such research must be regarded as work in progress, open to revision whenever new information or new sources become available.

Also that analysis rested too heavily on subjective judgement and that interpretation reflected the researchers' culture, rather than that of the performers.

Cantometrics was designed to be subjective, partly because it was intended for use by non-specialists, partly because we had no need for the precisely defined observations required by specialists, partly because there are hidden pitfalls inherent in almost all attempts to be objective when evaluating style.  Let me give you some examples, based on specific objections raised in the past.  How can loudness be determined from a recording?  Well, how can it be determined from a live performance?  Isn't the objectively measurable loudness (from, say, a decibel meter) going to depend on how far you are from the performer?  The only way to control for that would be to make all the recordings yourself with the microphone always at the same distance.  But even there you'll have problems due to the different acoustic properties of the different places where the recordings are made.  So ultimately you'd need to take all your performers into the same studio in order to objectively measure their loudness.  But the studio situation might well inhibit them to the point that their style is altered and no longer authentic.  The bottom line:  there is in fact no way to objectively measure loudness unless one is prepared to totally control every aspect of each recording, which is neither practical nor meaningful.  And what is loudness anyhow?  Is it really something we experience objectively?  Clearly no.  If it were, then listeners would always need to be exactly the same distance from the performer or speaker system.  It is and always has been the result of a subjective response, based on the listener's psychologically determined sense of what the performer intends, which is how it's treated in Cantometrics.

Why not use a metronome to measure tempo?  A metronome is used to measure the frequency of a beat.  But it isn't always clear how the beat relates to perceived speed.  For example in a waltz one might hear a dotted half note as the beat rather than a quarter note.  Or one might not be sure which one to choose.  Choosing one instead of the other will change the 'tempo' by a factor of three.  What we learn from this is that the metronome marking is objective and accurate only when coordinated with a note value in a written score.  It is not really a measure of perceived tempo, which, again, is basically a subjective judgment.  So we ask the rater to code simply on the basis of how 'fast' or 'slow' the music appears to be going.

How is it possible to determine vocal width (i.e., vocal tension) simply by listening, rather than observing the singer's throat muscles or  consulting an x-ray?  Vocal width is an especially important parameter for many reasons:  it can be a highly distinctive and recognizable aspect of style; it would appear to be the most direct indicator of emotion; and it was central to Lomax's theories regarding the relation between vocal tension and sexual restrictions (more on that below).  Unfortunately it was not the easiest parameter to learn and it took me a while before I became relatively confident rating it.  And it's certainly been the most controversial of all our parameters.

At one point Alan invited one of the leading laryngologists in New York to consult with us on this matter.  He could diagnose vocal tension from listening and was able to give us some valuable tips.  Alan also invited a group of linguists to assist him in evaluating voice quality from sonograms, but results were inconclusive.  More work has been done with sonogram analysis of vocal timbre in recent years, so there's hope that our ratings can be made more reliable in future.  As things now stand, I feel confident in the great majority of cases that I can reliably code vocal tension.  But some voices are more difficult to 'hear' than others.

As far as accuracy in general is concerned, consensus tests were done by an independent investigator, Norman Markel, at the University of Florida (see Folk Song Style and Culture, Chapter 5).  While some lines were more consistent than others, consensus for the system as a whole was 82%.  ('Factors . . .' op. cit., p. 32).  As expected, the consensus for 'Vocal Width' was not so high, but nevertheless respectable, at 65%.

Regarding the comment about our results reflecting the 'researcher's culture rather than that of the performers,' I would think that would be a much more serious problem for traditional research than Cantometrics, where the same standards, however subjective, are applied across the board, based on clearly defined criteria.  Of course no research by any investigator using any method could ever be free of culture bias.

Finally, the links between vocal traditions and culture are not as straightforward as Lomax seems to think.

To explain this latter point, let me discuss the example which first let the cat out of the cantometrics bag.3.  'Folk Song Style and Culture', Introduction.3  During a major collecting exercise in Spain in 1953, Lomax noticed that the intensity of local sexual prohibition was reflected in the vocal tones adopted by his singers.  Thin, harsh, high pitched and piercing, with intense emotional delivery, in the proscriptive south.  Broader, softer, lower pitched and more relaxed in the comparatively permissive north.  A recording trip to Italy, two years later, found the pattern repeated.

Not only vocal timbre, but the tendency to sing together in groups, tonal cohesiveness and the likelihood of polyphonic singing were all associated with fewer restrictions on women.

At first the argument seems logical and well substantiated.  The rules of a harsh, restrictive society finding expression in a harsh, restricted vocal style, with the commercial issues of Lomax's fieldwork confirming the geographical spread for both countries.4.  The Spanish fieldwork is surveyed on a fine set of LP recordings from the long-defunct Westminster label, 'Songs and Dances of Spain' WF 12001 onwards (11 volumes).  The Italian recordings can be sampled on Everest / Tradition TLP 1030, 'Music and Song of Italy'.  See also the relevant volumes for both countries in the 'Columbia World Library of Folk and Primitive Music'.4  But to what extent can we draw general conclusions from such findings?  The recent splendid Topic reissue of Tangent's Music in the World of Islam has reminded me afresh of the magnificent singing of Dunya Yunis, a woman from a mountain village in the Lebanon. 5.  'Music in the World of Islam' (Topic TSCD 901-3).5  How readily she equates with those singers from southern Spain!  Harsh, narrow toned and tormented, with a passion that could carve its imprint in granite, and long extended mordant passages to underline the tension.  It stings like a cry of anguish in the teeth of a society which continues to harbour medieval attitudes against women.  Yet, when her song lapses into something akin to recitative, I am reminded of similar field recordings from further east in Rajasthan.6.  The example in mind is on Harmonia Mundi HM959, 'Songs of Love and Devotion of Rajasthan'.  See also Ocora OCR 81, King KICC 5117, or Saydisc SDL 401.6  There, if what's available to western listeners is anything to go by, singers tend towards a terse vocal projection.  But, when compared with our other examples, they appear broader and softer toned and lower pitched.  According to cantometrics theory this should signify some measure of sexual liberation, but I'd be very surprised if the sexual mores of Rajasthan are any more easygoing than those of Spain, Italy or the Lebanon.

Cantometrics is a statistical method.  Its patterns and correlations are based on multiple instances, not just one.  There will always be exceptions, which is why statistical methods are generally preferable for comparative studies than the close examination of individual examples.

What's wrong here, I think, is that the style has been viewed against too small a number of cultural elements.

Now who's being subjective?  Why are you assuming that your personal response to the performances of a single vocalist constitutes a valid exception to the Cantometric results, which encapsulate the styles of many?

If we take the term culture to mean a socially acquired system of shared meanings, attitudes and values and the symbolic forms through which they are expressed, then this can only be considered in its entirety.7.  Adapted from Peter Burke, 'Popular Culture in Early Modern Europe' (Temple Smith, London: 1978).7  For, while culture moulds our attitudes to life and gives meaning to the world which surrounds us, it does so as a complete and interactive system of ideas, beliefs and moral codes.  Therefore, if it is to act as a means of expression, singing style must reflect the whole of culture, not just part of it.

This seems a rather restrictive notion of the role of song.  Why shouldn't certain aspects of music reflect certain aspects of culture?  I'm sure you'd agree that certain instruments have strong sexual overtones, why not certain modes of vocalizing?  What Alan observed in Italy and Spain was in my opinion a remarkably astute insight, possibly a profound one, with many ramifications well worth exploring.  His theory was apparently borne out by correlations obtained years later, from a much larger sampling of world music, compared with data regarding restrictions on women from the Murdock Ethnographic Atlas (see Folk Song Style . . . pp. 163-169, especially figures 51 and 52, comparing tonal cohesiveness, relaxed voices and polyphony with 'productive complementarity').  The results look very convincing, so much so that I'm strongly inclined to see this as a vindication of Alan's original insight.  It should be noted, moreover, that the correlations in question involve not only sexuality but what is now called 'gender.'  While many studies of 'women's music' have been done since, few if any have paid much attention to the affect that the role of women has on performance style generally, among men as well as women.

Nevertheless, and this represents a difference of opinion between Lomax and myself, I see Cantometrics as an essentially heuristic method, the results of which must always be regarded as provisional.  There are many reasons for this, among them the possibility of rater error or bias, the very real difficulties in obtaining truly representative samples, and the difficulty of interpreting very broadly defined correlations.  For these results to be accepted as representing the situation for all societies everywhere, further studies would have to be done, on a culture by culture basis (e.g., will the same correlations hold true among Amerindians or Northern Europeans?), and also with a sample corrected for various types of possible distortion (e.g., the result could be distorted by the fact that a very large part of our sample is from Africa and Polynesia where the correlation does seem to hold - how would things look if these samples were removed?).   All such testing would need to involve the basics of scientific method, random samples, double blind encoding and replicability.  Even if the correlation turns out, as I suspect it will, to be valid only for certain societies but not all, I feel sure nevertheless that it would be a highly significant result, well worth following up on.  That is not something I'd say for all the correlations Alan found, though most seem reasonable.  But each and every one, no matter how questionable, deserves further research and testing.

Yet, whatever the validity of the experiment, I feel that Lomax, in discerning an axiom of folk art, was intuitively more right than wrong.  It is not just that, by the time of those 1953 observations, he had already spent more time in the field than any collector before him.  It is that few of his predecessors developed such a close empathy with their subjects, or possessed equal opportunity to understand and appreciate how singers use their skills to express their emotions.

Exactly.  What made Alan so remarkable as a researcher was his ability to take in both the large view and the small, the grand overview and the telling detail.  Both are necessary ingredients of effective and meaningful research.  Alan's very special ability to foster a close rapport with those whose music he recorded and studied, made it possible for him to see through the surface to the social and psychological conflicts lying beneath.

For wherever people are trapped in the hardship and squalor of stark poverty and ruthless exploitation, singing represents one of the few avenues available for giving vent to what they feel.  Now, since culture defines our view of the world, it also defines our emotional responses and tells us how these should be expressed.  It is therefore a crucial factor in deciding the form and context of a people's songs, as well as having a strong say in how these will be performed.

Yes, that's an excellent point.

The problem with doing things cantometrically, however, is that the method relies on analysing performances 'typical' of their tradition.

Not exactly.  It involves sorting out the performances that are truly representative from those which are not.  Thus Lomax had no interest in recordings made by most professional 'folk' singers, but had no problem with traditional singers whose styles were unusual or in some way 'atypical.'  In some cases, performances were included specifically because they were atypical and thus of special interest.

But what constitutes typical song performance?  The examples of Sam Larner and Harry Cox have often been quoted, and with good reason.  For although less than 15 miles separated their respective Norfolk villages, Larner's engaging, conversational manner couldn't be further removed from Cox's dry and diffident approach.  Similar contrasts can be found wherever you look.8.  Look no further than the review section of 'Musical Traditions 12'.  Peta Webb's review of 'An Hour of Song' (ITSC 002) contrasts the styles of Maggie McGee and Dan McGonigle, who live a few miles apart on the Inishowen peninsula, Donegal.8  Which, for example, represents the 'true' style of County Tyrone, Northern Ireland?  Robert Cinnamond's thin, stark projection, with the high notes held and extended as though time had lost all meaning, or Geordie Hanna's lower-pitched, gapped, ornamented and altogether more emphatic delivery?  All four of these, by the way, together with the rest of the singers who make these isles such a rich and varied tapestry, are bracketed within a general European culture area.  So far reaching is this grouping that it finds a unity of style stretching 'from the Caucasus to the Atlantic'.9.  Lomax (ed.), 'Folk Song Style and Culture.'9  Of late, I've been taking a close look at one isolated parish on the fringes of this region, in Gaelic-speaking Ireland.  There are around 12 singers in this parish and the styles in that one small neck of the woods are anything but unified.

There can be no absolute measure of diversity.  To someone from a totally different culture, all these singers might sound 'exactly the same.'  To someone like yourself, steeped in these traditions, there is astonishing diversity.  What Cantometrics contributes is a set of clearly defined measures by which diversity can be evaluated across the board by exactly the same standards.  The standards are admittedly crude but such crudeness makes it possible to compare different traditions all over the world in a reasonably uniform manner.  Since  they are intended for very broad based comparison, they will seem to 'run roughshod' over all sorts of things the specialist would find important.  I see no problem with that.  Each approach has its place.  In my opinion the specialist and the generalist need each other and should work together.

Instead of viewing the singer as the passive vessel of external forces, it is far more realistic to see him or her as a creative artist working within the constraints of culture, just as a painter works within the constraints of paint, canvas, subject and imagination.  Here it's as well to emphasise that I'm not trying to advance abstract theory for its own sake.  As listener and performer I'm constantly aware that I'm digging in soil where I have no roots of my own and that my life's experience and outlook are vastly different from those of the people whose songs I've acquired.  Understanding the culture of tradition-bearers enables us to look through their eyes at what moves them when they sing.  By so doing we can comprehend what should be moving us.

The relationship between tradition, or as you say, 'external forces,' and individual creativity is a very old question which Cantometrics can never settle.  As I'm sure you realize, Lomax had a great respect for creativity which, as he clearly realized, can be at its greatest when constrained by even the most conservative and strictly defined traditions.  Cantometrics was designed to look past the creativity to the traditions beneath it, as well as the social, historical and psychological forces which lie beneath those traditions.  Only specialists like yourself can do justice to the creativity of each individual artist, chafing against those rigid bounds.

If the record is silent concerning variances between near neighbours, there are valiant efforts to explain similarities which arise over far flung and seemingly unconnected traditions.  Lomax makes much of Eurasia, a huge region which embraces a great deal of North Africa, Southern Europe and Continental Asia.  In his eyes the movements of migration, conquest and trade were responsible for spreading across this area a type of society characterised by rigid hierarchies, deferential attitudes, strongly defined sex roles and absolute despotism.  There is obviously a lot in this, but far too much emphasis seems to be placed on culture as it originally developed, and not enough on culture as it now is.

I believe the idea was to understand how the former has influenced the latter, how the same basic themes and conflicts keep recurring.

For if music expresses our view of the world, then it will change as that world changes.  In any event, Lomax tells us, there followed in the wake of this autocracy, a sorrowful, repining kind of song.  It was heavily ornamented, melodically indeterminate, with narrow intervals and rhythmically free.  This he calls the high lonesome complaint.  It spread to the Orient in one direction, to the empires of North Africa in the other and down to the Wolofs of Senegal, before being carried from there into the mouths of Mississippi convicts and mule skinners.  A staggering traverse by any standards, but the ball doesn't stop rolling even then.  'There are', he tells us, 'more than hints of this ancient song type in the so called big songs - the highly-ornamented and complex bardic melodies - that entertained the kings of Ireland in ancient times.'10.  Alan Lomax, 'The Land Where the Blues Began' (Methuen, London: 1993).10  I do not know where Lomax finds the evidence to support this statement, for the ravages of Irish history were so far reaching that we have next to no idea what the orally transmitted bardic music of ancient Ireland sounded like.  Hardly anything earlier than the seventeenth century survived to be written down.11.  Donal O'Sullivan, 'Songs of the Irish' (Browne & Nolan, Dublin: 1960).11

Maybe here you have come to the heart of the matter.  To what extent can the far ranging associations revealed by Cantometrics be carried into our understanding of what music is and what it means?  One of the things I always admired most about Alan was that he wasn't just a scientist or scholar, but many other things, among them a documenter, a social reformer, a philosopher, but also a poet - and visionary.  Just like the bards of old he'd be inspired by what he was seeing around him - or learning from his research - to expand a bit on some of the thorniest matters, not without some exaggeration, but always with an eye to a deeper truth.  The large-scale patterns revealed by Cantometrics are convincing, at least to me, but it takes a special sort of eye - and ear - to read beyond the statistics to the vast historical, social and human dramas which produced them.

Perhaps it is too late in the day to wonder what the well heeled kings of Ancient Ireland would have found to listen to in the high lonesome complaints of their downtrodden subjects.

As far as Ireland is concerned, there is, as you know, a tradition of highly ornamented solo singing.  What is not so well known, and what Cantometrics reveals, is how unusual that is for Britain and Western Europe as a whole, where embellishment is either absent or restrained.  Lomax was tempted to connect the Irish style with the larger 'Bardic' traditions of Eastern Europe, Asia and North Africa, but there are differences which make such a connection problematic (for one thing, Irish singers tend to have more relaxed voices) and there would seem to be no clear historical connections.  As I see it, this remains a mystery, but a fascinating mystery, which only broad based comparative study could have revealed.

Nevertheless, it is probably true that Cromwellian dispossession of the Catholic hierarchy, and consequent absorption of its file into the ranks of the peasantry, had a hand in shaping the elaborate melodies and handsome texts of modern Gaelic tradition.  True also that the passionate droning, nasal intonation and ornamentation associated with singers from Connemara finds echoes all over Eurasia, and that Connemara singers seem to share the Eurasian fondness for wandering melodies, narrow in compass and indeterminate in structure.  But vocal styles and melody structures vary throughout Gaelic Ireland.  Over generalisation is an easy trap to fall into, but the province of Munster has thrown up melodies no less complex than those of Connemara, but frequently they are much more wide ranging more sharply defined and arguably more handsome.

Cantometrics was conceived as an ongoing process, with ever expanding frontiers.  It would certainly be interesting to have samplings from all the provinces of Ireland and hopefully that will someday be the case.  It's also important to understand that Cantometrics is, in principle, expandable, so additional parameters could be added to capture certain features characteristic of a given region.  That said, it's important to understand that the specialist and the generalist must of necessity calibrate their observations very differently.

Frequently too, and with the exception of certain singers from West Cork, one comes across performers from Munster who use relatively little ornamentation and who project their voices in a clearer, more open throated manner.  Melodically and stylistically we can place them in a different camp to Connemara and Eurasia.12.  With the exception of Connemara, the world of Irish Gaelic song is woefully under-represented on commercial record.  However, see Amhrain ar an Sean Nos. RTE CD 185 for a fascinating survey of the riches of the idiom - and a bewildering array of singing styles.12  Yet, while both parts of Ireland share many songs and are culturally very close, they differ in one important respect.  Munster, generally, is good farming country; Connemara, on the other hand, is a barren wilderness of thin, stony, rock strewn soil and savage Atlantic storms.  Ireland, as with peasant societies generally, once possessed severe modes of sexual conduct,13.  K H Connell, 'Irish Peasant Society' (Oxford University Press, Oxford: 1968).13 and it's people knew everything which tyranny, in the form of rack renting absentee landlordism, artificially induced famine, mass evictions and emigration could throw at them.  The miseries of foreign exploitation undoubtedly played their part in shaping the way people sang, yet comparison between all three areas suggests that the high lonesome complaint may be the subject of a further hypothesis; that harsh impassioned singing, whether from Connemara bog, or Eurasian desert, can also be attributed to the rigours of being forced to wrest a living from a harsh unyielding environment.

Or, possibly, due simply to the force of tradition itself, which can be no less relentless.  Neither Cantometrics nor Lomax has, in my opinion, said the last word regarding the relationship between song style and culture.

Cross-cultural parallels account for many of the song world's stylistic similarities, but there are others which I suspect would be less readily explained.  What for instance, of singers from the Spanish Asturias, such as Manuel Gandoy, who use tone and voice projection in a way that we'd usually associate with the Balkans?14.  Manuel Gandoy, 'En Gaita Traigo a Asturias' (Spanish Fontana 64-29-066).14  What of singers from the Tatar region of the former USSR, who not infrequently sound like their confrères from the Hebrides?15.  For example, compare 'Music of the Tatar People' (Tangent TGM 129) with Scottish Tradition 2: 'Music From the Western Isles' (Tangent TNGM 110).15  Am I alone in hearing parallels between the raw emotion of flamenco and the tortured singing and strident guitar playing of Fred McDowell, Bukka White and Blind Willie Johnson?  Apparently not, for Lomax himself refers to the blues as 'America's Cante Jondo'.16.  Lomax, 'The Land Where the Blues Began', p. 9.16  Yet what social similarities exist between black Americans and Spanish gypsies?  I would not question Lomax's assertion that song expresses the culture of which it is a part.  But how that expression is realised is far more complex, multifaceted and indeterminate than the cantometrics method will allow.

The Cantometrics method must be considered separately from the interpretation of its results.  And there is more than one method.  I've done searches of our entire world database selecting for certain clusters of traits only, an approach designed to reveal exactly the sort of parallels you mention.  Such parallels, cutting across the usual boundaries, can be extremely suggestive, as Lomax was well aware.  What they might mean is another story.  As I see it, Cantometrics opens doors to all sorts of possibilities, and there are still a great many to be considered.

Mention of such icons brings me back to The Land Where the Blues Began, and the premise propounded throughout the book that the idiom was born in Mississippi of African parentage around 1900.  Few observers nowadays would dispute the date,17.  Paul Oliver, 'Songsters and Saints': Vocal Traditions on Race Records (Cambridge University Press, Cambridge: 1984), Ch. 9.17 but I thought the question of location was a canard which had starved to death years ago for want of substantiation.  The blues could well have had its genesis in Mississippi, for it found its apotheosis there in the raw and harrowing singing of a litany of Delta bluesmen like Son House, Robert Johnson and Big Joe Williams.  But apotheosis does not necessarily equal origin and the requirements for incubation were a feature of black social conditions all over the South.

This is a very complex issue with many ramifications, historical, cultural, social, economic, not to mention (lest we forget) aesthetic.  As far as the blues is concerned, I can't think of anyone in a better position to assess all of these than Alan Lomax, so I would think his premise would need to be taken very seriously, Cantometrics or no Cantometrics.  Since the blues is a solo idiom with a fair amount of vocal embellishment (rare in Africa), and in some other respects not all that typical for Africa Cantometrically speaking, Lomax's conclusions are based on other aspects, notably rhythm, voice quality, instrumental technique, movement style (see his work on Choreometrics), etc.  One can trace the continuities quite clearly, I'd say, from certain one string instruments in Africa, to the one-string Diddley-bow, to the single chord guitar accompaniments of early blues, with their very 'African' driving rhythms, not to mention the characteristically African pentatonic melodies superimposed over standard European chords to produce the so-called 'blue-note' clashes we know so well.  The African elements are there.  But the style as a whole was clearly produced as a response to historical events and social pressures as well.  It is definitely the product of the African American culture of a certain time and place, but with roots in certain practices common in Africa.

So many writers have sought to explain the blues in terms of African survivals and adaptations without appearing to give due weight to the socio-historical implications of their arguments: that if African music reflects African culture, then we would expect that music to be transformed as the culture was transformed.  Also, if the blues did not come into being until almost two generations after the abolition of slavery, then it must be the product of social forces which followed in the wake of slavery.  Plantation experience and African heritage moulded social attitudes and musical background, but neither can directly explain the idiom.

The relation between song style and social forces may not be as straightforward as one might think from reading some of Lomax's commentaries.  In my opinion, the force of tradition itself must also be reckoned with.  Certain African traditions may have survived simply through a force of their own, regardless of cultural transformations.

Cantometrics is rather better at discovering cultural patterns than it is at explaining them ...

It's important, as I see it, to make a clear distinction between Cantometrics as a method and the way in which its results have been interpreted.  While many of Lomax's explanations are convincing, others are not.  As I see it, this methodology is designed to open the doors to many possibilities, only a few of which have been considered to date.

and it generally supports traditional arguments about the African nature of Black American music.  It does, however, throw up a thought-provoking - and disturbing - variation to this theme.  Pointing to the wide melodic intervals and leaderless choruses of African music,

African 'Bantu' choruses typically involve an antiphonal relation between a 'leader' and a group.  I wouldn't call them leaderless.  The complex, interlocked group vocalizations of the Pygmies and Bushmen, however, are often leaderless.

Lomax cites these features as evidence of well-integrated, free, egalitarian societies.18.  Lomax, 'The Homogeneity of African-Afro-American Musical Style', in N E Whitten and J F Szwed (eds.) Afro-American Anthropology (Free Press, New York: 1970).18

The clearest examples can be found among the most typical African hunters, i.e., the Pygmies and Bushmen, whose musical performances do seem very clearly to reflect the organization of their egalitarian, acephalous societies.  The picture for African farmers and herders is somewhat more complex.

The picture is exaggerated, but it highlights the dramatic contrast between African freedom and American slavery.  Turning to the blues, however, he does not look to slavery to explain the idiom's flattened notes and expressions of discontent.  Instead he points to the post-abolition campaign of terror and destabilisation waged on black people by an insecure white population.19.  Lomax, 'The Land Where the Blues Began.'19

I am not in a position to evaluate this argument, but I find it extremely persuasive.  The plantation system was dehumanising, degrading and horribly brutal.  Yet an enclosed institution could well have generated a measure of communal solidarity among its inmates, to help counter the worst excesses.  But a free black population turned loose on the world, may have been perceived by the whites as more of a threat than ever they were when suppressed by slavery.  If so, would the whites not have tried to destroy that threat via the disorientation of lynch mob and Ku Klux Klan?20.  Those parts of the Caribbean where African cultural retentions are stronger than in the USA, show markedly stronger African musical retentions also. See, for instance, 'Drums of Defiance': Maroon Music From the Earliest Free Black Communities of Jamaica (Smithsonian Folkways SFCD 40412), 'From Slavery to Freedom:' The Saramaka Maroons of Surinam ( Lyrichord LLST 7354) and 'Early Afro-Cuban Songs' (Albatros VC 4932).20  In breaking the back of whatever bonds of community kept black people going through the years of slavery, did they not also sow the seeds of social alienation so characteristic of the blues?  If so then it is here we should look to explain this crystalisation of pre-existing song forms, so negative in content, so disaffected in outlook and so completely different in emotional temperament to the songster tradition which preceded it.21.  Oliver, 'Songsters and Saints'.21  The blues makes sense, not as a survival of West African tradition, but as the artistic expression of the transition from slavery to nominally free, downtrodden citizenship.

The above strikes me as a very sensible analysis, with which, I believe, Lomax would agree.  The blues certainly does seem to have been a response to conditions imposed during the reconstruction era, but that's no reason to assume it lacks African roots.  But why focus so much attention only on the blues?  There were other traditions, such as Spiritual and Gospel, even more clearly African in origin.

The music of black America was born out of the cultural experience of black Americans.  If they could do nothing to change their situation, their musical lingua franca at least gave them a voice through which to channel their feelings - a capacity shared by the natural inheritors of folk tradition the world over.  Today, while much of that capacity has been displaced by the intertwining of mass media and mass society, what's left of the world's musical traditions is increasingly subsumed beneath a plethora of sanitised, soulful singers and gentle, tinkling soft-rock bands.

Criticism of certain aspects of Alan Lomax's work should not be taken as criticism of the man's achievements generally.  I feel a considerable personal debt, not just for the way he has liberated so much of the world's music, but because most of the intellectual and ideological baggage I carry around with me stems from his influence.

Here we can fully agree!

There are those of us - and Lomax would surely count himself among the coterie - who believe that unburdening of emotions through song is important, both for the community of performers and for the community of listeners.  In Folk Song Style and Culture he referred to Telstar rising 'balefully on the western horizon'.  A quarter of a century later we are bombarded with satellite videos, computer games, cable networks and the whole paraphernalia of a technology designed to entertain, not to express.  At the same time the march of progress destroys native cultures almost as fast as it destroys rain forests.  Before we are deafened completely by a soundscape of stultifying blandness, is it yet too late to reopen the cantometrics debate?

Even after all this time, I hope not.  Let the debate continue, sez I.

Victor Grauer - 24.7.05

Article MT159

Top of page
Articles
Home Page
Reviews
News
Editorial

Site designed and maintained by Musical Traditions Web Services  Updated: 24.7.05