Three Approaches to Defining Jazz

This article appeared in THE MUSICAL QUARTERLY, 1989, Volume 73, Issue 4, pages 513-531, and was reprinted in JAZZ: A CENTURY OF CHANGE: READINGS AND NEW ESSAYS, Edited by Lewis Porter (Schirmer, 1997).

It was developed from the “What is Jazz?” chapter of the Jazz Styles book by Mark Gridley (Prentice-Hall, 1978), An Outline of Jazz (copyright 1973 by Gridley), and a paper presented to the Allegheny chapter of The American Musicological Society on October 26, 1985 by Robert Maxham and Mark Gridley.

Notes about the Authors:

Mark Gridley, Ph.D., is a professor at Heidelberg College, Tiffin, Ohio 44883. He developed the strict definition of jazz, collected the approaches to defining swing feeling, and researched the documentation and illustration of arguments revolving around all three approaches to defining jazz.  Gridley also edited 17 different drafts of this article, as well as soliciting and incorporating criticism of each draft from jazz journalists, musicians, and musicologists.

Robert Maxham holds a Ph.D. from Eastman School of Music and has a long-standing interest in philosophy, especially Ludwig Wittgenstein’s approaches to the problems of definition.  He developed the family resemblances approach as well as criticizing and editing the strict definition and the continuum approach.

Robert Hoff is a professor at Mercyhurst College in Erie, Pennsylvania and hosts “New Music Now” and “All That Jazz” for National Public Radio (WQLN)in Erie.  As well as criticizing and editing the strict definition and the family resemblances approaches to jazz, he proposed the approach that views jazz as a continuum.

Three Approaches to Defining Jazz

Mark Gridley, Robert Maxham and Robert Hoff
Copyright 1989 by Oxford University Press

We will examine three approaches to defining jazz.  None of them is entirely satisfactory.  But each has virtues that none of the others has.  All specify somewhat different, though overlapping, bodies of music that are tagged with the same label.  These clarifications are needed because too few speakers, and far too few authors, take the time or thought to explain what they mean by the term before they use it. Endless problems result when they assume that listeners and readers share their frame of reference.

Any scholarly analysis requires clear definition of the terms to be used, and the term jazz has been particularly problematical.  The term has been used in widely disparate ways.  This has caused endless controversy, much of which is probably needless.  Numerous books about jazz do not even offer a definition of the term (FOOTNOTE:  recent examples include JAZZ: AMERICA’S CLASSICAL MUSIC by Grover Sales, Prentice-Hall, 1984, JAZZ: A LISTENER’S GUIDE by James McCalla, Prentice-Hall, 1982, and THE ORIGIN AND DEVELOPMENT OF JAZZ by Otto Werner, Kendall-Hunt, 1984).  And efforts by writers who do attempt to define jazz have been confused.  Compare, for example, the following definitions that have been offered by various journalists and teachers:

Sigmund Spaeth: “Jazz is not a musical form; it is a method of treatment.  It is possible to take any conventional piece of music, and ‘jazz it.’  The actual process is one of distorting, of rebellion against normalcy.” (“Jazz Is Not Music” THE FORUM, August, 1928)

Chambers’ Encyclopedia:  “Jazz–dance music, generally syncopated, played by a band eccentrically composed.  The jazz drummer, a sort of one-man band, provides the characteristic feature of jazz, which is noise…”

Virgil Thompson:  “Jazz, in brief, is a compound of (a) the fox trot rhythm (a four measure, alla breve, with a double accent), and (b) a syncopated melody over this rhythm.” (August, 1924, AMERICAN MERCURY)

Henry Osgood:  “It is the spirit of the music, not the mechanics of its frame or the characteristics of the superstructure built upon that frame, that determine whether or not it is jazz.” (SO THIS IS JAZZ, 1926, Boston:  Little, Brown & Co., page 26)

George Antheil:  “The works of Vincent Youmans are clear, and extremely beautiful examples of jazz that is a pure music.” (In a letter to the editor of THE FORUM, published December, 1928)

Wilder Hobson:  “To some it means the whole cocktail-swilling deportment of the post-War era.  To others it suggests loud and rowdy dance music.  Many people go so far as to divide all music into ‘jazz’ and ‘classical’.  By ‘classical,’ they mean any music which sounds reasonably serious, be it ‘Hearts and Flowers’ or Bach’s ‘B-Minor Mass,’ while their use of ‘jazz’ includes both Duke Ellington’s Afric brass and Rudy Vallee crooning ‘I’m a Dreamer, Aren’t We All?’…But Duke Ellington bears just about as much relation to Vallee as the ‘B-Minor Mass’ to ‘Hearts And Flowers’…Ellington’s music is jazz.”  (FORTUNE, August 1933)

Willie Ruff:  “More than anything else, jazz is a feeling, a way of playing music, a way a musician feels at any given time.  It isn’t written…we want to give musicians room to improvise…Another important ingredient in jazz is syncopation.”  (1958 Young Audiences lecture/demonstration that was recorded and released as JAZZ FOR JUNIORS, Roulette R5205)

Leonard Bernstein:  “Jazz is a very big word; it covers a multitude of sounds, all the way from the earliest Blues to Dixieland bands, to Charleston bands, to Swing bands, to Boogie Woogie, to crazy Bop, to cool Bop, to Mambo–and much more.  It is all jazz…it is an original kind of emotional expression, in that it is never wholly sad or wholly happy…Rhythm is the first thing you associate with the word ‘jazz.’…But jazz could not be jazz without its special tonal colors, the actual sound values you hear…A popular song doesn’t become jazz until it is improvised on, and there you have the real core of all jazz: improvisation. (“The World of Jazz” from CBS Omnibus series, telecast October 16, 1955)

HARVARD DICTIONARY OF MUSIC, 2nd Edition: “A kind of indigenous American music of the 20th century, originally identified with social dancing, featuring rhythmic patterns peculiar to the ‘jazz beat’.” (Cambridge, Mass: Harvard U. Press, 1972, page 317)

Joachim Berendt:  “Jazz differs from European music in three basic elements:  1. a special relationship to time, defined as ‘swing’ 2. a spontaneity and vitality of musical production in which improvisation plays a role  3. a sonority and manner of phrasing which mirror the individuality of the performing jazz musician.  These three basic characteristics create a novel climate of tension, in which the emphasis no longer is on great arcs of tension, as in European music, but on a wealth of tension-creating elements, which continuously rise and fall.  The various styles and stages of development through which jazz has passed since its origin around the turn of the century are largely characterized by the fact that the three basic elements of jazz temporarily achieve varying degrees of importance, and that the relationship between them is constantly changing.” (THE STORY OF JAZZ, Englewood Cliffs, NJ:  Prentice-Hall, 1978, page 7)

Henry Martin: “…jazz is a twentieth century music originated in America by black Americans and characterized by improvisation and a strong projection of rhythm.” (ENJOYING JAZZ, New York: Schirmer, 1986, page 4)

Max Harrison:  “Attempts at a concise–even a coherent definition of jazz have invariably failed.  Initial efforts to separate it from related forms of music resulted in a false primacy of certain aspects such as improvisation, which is neither unique nor essential to jazz or swing (the quality of rhythmic momentum resulting from small departures from the regular pulse), which is absent from much jazz, early and late.” (THE NEW GROVE DICTIONARY OF MUSIC AND MUSICIANS, London: Macmillan, vol. 9, 1980, page 561)

Lay uses of the term have been so confused and inconsistent that the following popular issues remain unresolved, pending a clear and consistently applied definition for jazz:

Is jazz art music?  How can we answer this question unless we have clear definitions of jazz and of art music?  And then what definition of jazz would apply to the music of Al Jolson from his 1928 film “The Jazz Singer” or the music of the vocal group called Peter, Paul & Mary, that won first place in the vocal division of “jazz” popularity polls run by PLAYBOY MAGAZINE in 1964, 65 & 66?

Was jazz popular during any given era of twentieth century music?  This question becomes problematic, especially for music of the 1930s and 1970s, because a number of best-selling records during those periods contain improvisations, but the improvisations were so brief that they might fail to qualify their performances for the “jazz” designation because they might not be making some critical minimum contribution to the overall performance.

How should we interpret discographer Brian Rust’s statement about Glenn Miller’s output, “…the majority of the Bluebird and Victor sides are of little or no jazz interest…”? (JAZZ RECORDS, Vol. 2, 1970, page 1125; Storyville Publications, Chigwell, England)

To attack the dilemma that underlies solutions for these problems, we will look at three different kinds of definition and test their fit on styles that have proven especially difficult to classify.  We will examine, first, a strict definition, one that requires improvisation and swing feeling, second, a much different approach, in which styles are tied together only by family resemblances, and third, an “essence approach”, in which the relative presence of certain components determines the relative “jazzness” of a performance.

Framing A Strict Definition

Let us first evaluate a strict definition, one that requires a performance to be improvised and to swing in the jazz sense (FOOTNOTE: as opposed to swinging in the general sense, a quality ascribed to any successful performance of music that has steady tempo and lively execution; see Mark C. Gridley’s JAZZ STYLES, Prentice-Hall, 1978, pages 14-16).  For the sake of the preliminary parts of this discussion we shall say that to improvise is to simultaneously compose and perform, and to swing is to perform in a way that succeeds in projecting rhythmic properties that tend to elicit from the listener the perception of a lilting, buoyant feeling that is peculiar to jazz. (Such circularity in this characterization is unavoidable here.)  Traditionally, swing feeling and improvisation have been considered essential to a definition of jazz.  In fact, these two traits are prominently featured in all fifteen texts that are currently used in the U.S. for courses called introduction to jazz/jazz history/jazz survey. (FOOTNOTE: UNDERSTANDING JAZZ by Leroy Ostransky (Prentice-Hall, 1977), JAZZ: A HISTORY by Frank Tirro (Norton, 1977), LISTENING TO JAZZ by Jerry Coker (Prentice-Hall, 1978), THE MAKING OF JAZZ by James Lincoln Collier (Houghton-Mifflin, 1978), THE JAZZ TEXT by Charles Nanry (Van Nostrand, 1979), THE JAZZ BOOK by Joachim Berendt (Lawrence Hill, 1982), INTRODUCTION TO JAZZ HISTORY by Donald Megill & Richard Demory (Prentice-Hall, 1984), A STUDY OF JAZZ, 5th Edition, by Paul Tanner & Mauric Gerow (William C. Brown, 1984), JAZZ: A LISTENER’S GUIDE by James McCalla (Prentice-Hall, 1982), ENJOYING JAZZ by Henry Martin (Schirmer, 1986), JAZZ: AMERICA’S CLASSICAL MUSIC by Grover Sales (Prentice-Hall, 1984), WRITINGS IN JAZZ by Nathan Davis (Gorsuch Scarisbrick, 1978), THE ORIGIN AND DEVELOPMENT OF JAZZ by Otto Werner (Kendall-Hunt, 1984), THE JAZZ EXPERIENCE: A GUIDE TO APPRECIATION by Joseph Levey (Prentice-Hall, 1983), and JAZZ STYLES: HISTORY AND ANALYSIS, 2nd Edition by Mark Gridley (Prentice-Hall, 1985)).  Musicians and music publishers also allude to their primacy.  For example, in music publishers’ brochures describing big band arrangements, a note is frequently included to the effect that, “only the first tenor saxophone part requires jazz,” meaning that the first chair tenor saxophonist is the only member of the band who must IMPROVISE during performances of this particular arrangement.  Similarly, a musicians’ contractor might phone a player requesting he fill the “jazz trumpet” chair in a band, meaning that the player will have the responsibility of improvising solos in addition to the ordinary responsibilities of doing “section work” (playing ensemble parts).  So we see that the interchangeability of the terms “jazz” and “improvisation” is common in the music business.  Regarding the attention ordinarily paid to swing feeling in jazz, a brief perusal of record and concert reviews or a little eavesdropping among musicians will reveal that the quality and extent of swing feeling is frequently the topic of discussion among listeners who are describing a jazz performance.  This is also a popular topic among teachers and students who are involved in teaching and learning the skills of jazz performance.

The requirement that a performance contain improvisation allows us to differentiate jazz from other types of music that swing, such as popular music that bears swing feeling and frequently can be heard on American radio and television broadcasts (i.e., Andre Kostelanetz, Ray Coniff, Percy Faith, Broadway show music of the 1950s, much music that might be jazzy or jazz-influenced though it is not ordinarily called jazz).  This requirement also allows us to differentiate types of music that sound quite similar, such as that of Herb Alpert and Chuck Mangione, Alpert’s usually being un-improvised, therefore differentiable from Mangione’s, which usually contains enough improvisation to qualify it for the “jazz” label.  For those people who tend to say, “Jazz is a FEELING more than anything else,” a musician’s playing needs to elicit swing feeling from his listeners in order to qualify the player as a jazz musician.  And the requirement that a performance swing in the way peculiar to jazz allows us to distinguish jazz from rock music, Indian music and other nonjazz idioms that employ improvisation.

Problems become evident when we observe that either of these two criteria, when taken INDIVIDUALLY, applies to a larger number of styles than when both are taken TOGETHER.  The union of the two terms into a SINGLE definition creates difficulties.  For instance, some performances that have been called jazz DO SWING but contain little, if any, improvisation.  Some of Duke Ellington’s and Stan Kenton’s concert pieces fit this category. (All but the brief trombone solo in Ellington’s “Work Song” from BLACK, BROWN & BEIGE; Kenton’s recording of Robert Graettinger’s “Dance Before The Mirror” from CITY OF GLASS, Capitol W736.)  Other performances have been called jazz and DO CONTAIN IMPROVISATION, but don’t SWING.  The Johnny Hodges sax solo in “Come Sunday” from Ellington’s BLACK, BROWN & BEIGE as well as some performances of The Art Ensemble of Chicago (PEOPLE IN SORROW, Nessa n-3) and Cecil Taylor (SILENT TONGUES, Arista Al 1005) fit this category. (FOOTNOTE:  On April 22, 1982, the students of a jazz appreciation class at John Carroll University ranked the Taylor improvisation as “doesn’t swing at all.”  On October 26, 1985, the participants at the meeting of the American Musicological Association, Allegheny Chapter, agreed that the Hodges solo did not swing.  On February 22, 1986, panelists and audience in Denver, at the Colorado Endowment for the Humanities “What Makes It Jazz” session, concurred that the Art Ensemble selection did not project swing feeling.)

Despite this difficulty of excluding so many performances that are ordinarily classified as jazz, a strict definition of jazz requires the presence of both swing feeling AND improvisation.  So if we want to include as many styles that have been called jazz as we possibly can, the best approach to framing a strict definition might lie neither in loosening the term “jazz” itself, so that it INCLUDES styles that are called “jazz” merely by association with the jazz tradition, nor in EXCLUDING EITHER swing feeling or improvisation.  It might lie instead in adopting flexible definitions for jazz swing feeling and jazz improvisation.

In the process of framing flexible definitions for jazz improvisation, we must confront several knotty questions.  For example, what constitutes improvisation in the jazz sense?  How extensive must it be, and how fresh? (See Robert Brown’s “How Improvised Is Jazz Improvisation” in Charles Brown (Ed.), PROCEEDINGS OF NAJE RESEARCH, 1981, pages 22-32.)  Can it be no more than variations of timing and the alternation of ornaments? (FOOTNOTE: Listen to the Johnny Hodges solo cited earlier, as well as various versions of his solo on “Things Ain’t What They Used To Be”, compare for instance his July 30, 1945 Victor recording with his 1959 Columbia recording.)  Or must it be full-blown melodies and accompaniments? (FOOTNOTE:  Compare the Paul Gonsalves sax solos on these two versions of “Cop-Out”:  March 13, 1957, available on a Columbia sampler called JAZZ OMNIBUS Columbia CL 1020 and September 8, 1959 on Duke Ellington’s FESTIVAL SESSION Columbia CL 1400.)  Is it sufficient that a musician REPEAT patterns acquired either through his own creative effort or through his familiarity with the works of others? (FOOTNOTE:  Charlie Parker has occasionally reused some of the phrases first recorded by him on “Now’s The Time” November 26, 1945 for Savoy, and Miles Davis has frequently quoted his own phrases originally recorded by him April 22, 1959, available on Miles Davis KIND OF BLUE, Columbia CL 1355.  These are not exceptions.  They constitute an accepted practice.  In fact, many jazz musicians can be identified by the astute listener who detects the player’s favorite phrases in a new recording.  Also jazz musicians frequently quote each others’ favorite phrases, sometimes an entire chorus being recreated.  Gridley once attended a night club performance in which Sonny Stitt inserted note-for-note Lester Young solos into his tenor sax solos and note-for-note Charlie Parker solos into his alto sax solos.  Jon Faddis has crafted a jazz trumpet style by quoting Dizzy Gillespie phrases, and Wynton Marsalis has done similarly with Miles Davis phrases.)

If we are to be flexible, perhaps it behooves us to realize that no poet or novelist can be expected to write in a language entirely original to him, purged of all idioms and cliches.  Likewise, it would be unfair to expect that jazz musicians create not only fresh “paragraphs” and “sentences,” but even the “phrases” and “words” they use.  So the frequent recurrence of standard patterns should not by itself DISQUALIFY a passage as improvisation.  In crafting a strict definition for jazz, it might be useful to allow that an improvisation, then, CAN be constructed from pre-existing elements, only IF these elements are REORGANIZED, and they are reorganized at the very moment they are performed.

To answer the question of how extensive improvisation must be in order to qualify a performance as jazz, we ought to ask, “What minimum amount of improvisation should be necessary in a performance before we can call that performance ‘jazz’?”  Some of a jazz performance certainly needs to be improvised, such that SOME significant aspects of each performance are NOT PRESET.  For a flexible application of the improvisation requirement, let us suggest that the amount of NONPRESET material be just LOW ENOUGH to include performances in which improvisations are the outstanding feature and just HIGH ENOUGH to EXCLUDE works that do NOT exhibit improvisation as a prominent feature.  Note, however, that this would result in removing the jazz label from much of the big band dance music of the 1930s and 40s, thereby removing from the jazz category a large body of music that many usually consider to be jazz.

We also have to classify cases in which a performer arranges or simply repeats someone else’s improvisations.  The music of Supersax and Manhattan Transfer are examples. (FOOTNOTE: On SUPERSAX PLAYS BIRD Capitol ST 11177, Buddy Clark has arranged phrases from Charlie Parker improvisations and scored them for five saxes.  Manhattan Transfer is a vocal group that specializes in recreations.  Their version of Joe Zawinul’s “Birdland,” for instance, takes not only the ensemble parts from the original recording by the band called Weather Report, but also the sax and piano improvisations and sets lyrics to them for vocal performance.)  The strict definition would EXCLUDE these cases (perhaps justifiably because the premium that many jazz musicians place on spontaneity is well worth preserving, and it’s dangerous to confuse this jazz kind of spontaneity with another form of musical expression that is essentially an interpreter’s art).  Therefore, in defining jazz improvisation, let us rule out improvised alterations of timing and ornaments when such alternations constitute the only changes that occur from performance to performance, changes that ordinarily are classed with “interpretation” rather than with “improvisation.”

If we adopt a strict definition of jazz, then what can be said about performances in which the music doesn’t seem to be improvised?  Some of Duke Ellington’s concert music, as well as the nth repetition of his band’s hits, falls into this category. (For instance, many recordings of his “Cottontail” differ very little from each other).  If we define jazz rigidly so that SOME improvisation is ALWAYS required, then such performances canNOT be jazz.  This may seem to be a ludicrous conclusion, but it is both logical and useful.  In fact, a term for such works already exists:  SWINGING CONCERT MUSIC.  And this term not only saves the marriage of jazz and improvisation, but it also creates a category for a musical tradition that, though DERIVING from jazz, might better be kept SEPARATE.

As in trying to frame a flexible definition for jazz improvisation, we also confront several knotty questions when we try to frame a flexible definition for jazz swing feeling.  Ordinarily the feeling is described as a quality of experience elicited by music, and it is frequently characterized as a buoyant, lilting feeling that pulls the listener along.  Scholars and music psychologists remain unclear about precisely what causes listeners to report that a given performance swings, but a few ideas have received wide discussion.  One is the presence of “swing eighths,” patterns in which alternate eighth notes are slightly lengthened or shortened, yielding a ratio of about 60:40, rather than the straight 50:50.  (This may remind music historians of the phenomenon of improvised “notes inegales” in the music of the Baroque era.)   The 60:40 pattern is significant because it is a discernable and quantifiable component in the music that has traditionally been considered “swinging.” (See Richard Rose’s “Computer-Assisted Swing” in Jazz Educators Journal, 1985, 17, 3, 14-15; and Mark Ellis’ “An Analysis of ‘swing’ subdivisions and asynchronization in three jazz saxophonists” in Perceptual and Motor Skills, 1991, 73, 3, 707-713.)  A second, objectively quantifiable predictor of swing feeling is steady tempo, and a third, somewhat less easily measurable predictor, is a balanced proportion of syncopated rhythmic figures.

The prediction of swing feeling via the presence of swing eighths, steady tempo, and syncopation provides some of the methodology necessary to categorize any performance as swinging, since those features can be verified by acoustic analysis.  But the mere presence of any of these is not sufficient to allow us to reliably predict the listener’s judgment of a performance’s swingfulness.  The mere presence of syncopation, for instance, is not sufficient to evoke swing feeling in the listener.  In fact, there is a good possibility that the AMOUNT of syncopation in a performance corresponds in a curvilinear fashion to perceived swing feeling.  In other words, if amount of perceived swing feeling were plotted against amount of syncopations, the resulting graph would take the shape of an inverted U (see Andre Hodeir’s JAZZ:  ITS EVOLUTION AND ESSENCE, 1956, Grove Press). For example, too MUCH syncopation might be part of the reason that some of Cecil Taylor’s playing does not swing.

A fourth suggested cause of jazz swing feeling is the alternation of tension and relaxation that has been described by Hodeir in 1956, Pekar in 1974, and Gridley in 1985.  Gridley said that the alternation in density of activity might be an objective correlate of the subjective impression of tension/relaxation (Mark C. Gridley, JAZZ STYLES:  HISTORY AND ANALYSIS, 2nd Edition, 1985, Prentice-Hall, pages 7-8).  Hodeir described the ride rhythm (quarter note, dotted eighth note, sixteenth note, quarter note) as possessing the same properties.  Gridley believes that not only does this occur in the typical jazz drummer’s work, but that it also occurs in the rhythms of an improvised line.  In other words, on the suggestion of Peter Manuel of Columbia University, Gridley thinks that the syncopation might constitute density of activity, to function as an anacrusis, and that the on-the-beat playing constitute diffuseness.  To alternate such characteristics is to alternate the subjective impression of tension with that of relaxation.

As a fifth suggestion, there is also a relationship between THE DETECTABILITY of the tempo and the degree of perceived swing feeling.  In other words, even in pieces that WERE constant in tempo, naive listeners have ranked music LESS SWINGING if they had, also and independently, rated the beat as “difficult to detect” (Gridley, unpublished study, presented to Society for Ethnomusicology, April 24, 1982, Kent, Ohio).  Not entirely consistent with this finding is a sixth possible cause for jazz swing feeling: the spontaneous, slight expansions and contractions of the duration for each measure. (This means that absolutely constant tempo is incompatible with swing feeling.)  Related to this component is a seventh element, the rhythmic displacement practice common among jazz musicians (toying with the beat within the improvised line), perhaps ala Chopin’s “tempo rubato”.  And do not overlook the possibility that an interaction of elements exists, such that we cannot measure the relative contribution of each individual element.  We canNOT even realistically ESTIMATE their contributions.

Even the availability of a methodology for predicting the listener’s response to the objective determinants of swingfulness, however, is not sufficient to settle all doubts about what jazz swing feeling is.  Musicians and critics tend to agree that swing feeling constitutes a dimension of musical perception.  This means that some performances swing more than others.  We also need to note that listeners sometimes hint that not only is swing feeling a quantifiable dimension, but that it is also a multifaceted group of QUALITIES, amenable to description in terms of distinctive varieties.  For example, few listeners would argue with the statement that pianist Count Basie swings in a way that is QUALITATIVELY different from that of pianist Duke Ellington and that pianist Bill Evans swings in still a third way. (And some listeners claim that Evans fails to swing at all.)  Additionally, many suggest that the swing feeling typifying dance bands of the 1930s differs qualitatively from the swing feeling typifying jazz/rock bands of the 1970s.  And, as if those distinctions are not already confusing enough, frequently one listener will report that a given performance does not swing AT ALL, whereas another listener will report that the same performance swings a lot (though both might agree that the same performance was highly spirited).  The most flexible view is best exemplified in a classic article (“Swing As An Element of Jazz”, CODA, August/September, 1974, pages 10-12; updated in Jazzis, June, 1996), in which Harvey Pekar made clear distinctions among the earliest jazz musicians in terms of the extent of swing feeling that each projects in his playing, and he made clear distinctions between the swingfulness of Sonny Stitt and that of Ornette Coleman.  Yet at the end of his work, Pekar encouraged us to remain open to NONSWINGING jazz, music that comes from the jazz tradition but fails to swing.  He also asked whether swing feeling would continue to be as important in jazz of the future as it was in jazz of the past.  Before responding to Pekar’s suggestions, let us propose that, for a strict definition of jazz that requires both improvisation and swing feeling, the most flexible definition of swing feeling would have to include as swinging any performance that anyone ever reported to have elicited swing feeling in them.  (Note that swing feeling here is a purely rhythmic phenomenon.  A band can play out of tune yet still swing.)

Another problem is that, although listeners may agree that certain performances are rhythmically similar and other performances are rhythmically different, they disagree about whether those same performances swing, because such listeners use the term “swing” to denote their own emotional response to a given sound, and some listeners are pleased by the same sound that fails to move other listeners (I have interviewed jazz musicians who felt that John Coltrane’s playing on INTERSTELLAR SPACE-Impulse ASD 9277 swings, and others who felt that it did not swing at all).  It is important to note that these listeners are not HEARING differently.  They are just RESPONDING differently to what they hear.  We are dealing here with situations in which the percept is the same but the affect is different. (See Morris Holbrook and Stephen Bertges’ “Perceptual Veridicality in Esthetic Communication” COMMUNICATION RESEARCH, 1981, 8, 4, 387-423 and Holbrook and Joel Huber’s “Separating Perceptual Dimensions From Affective Overtones” JOURNAL OF CONSUMER RESEARCH, 1979, 5, 272-283.)  At the earliest stages of processing the sounds, all listeners agree.  This is the recognition stage.  At later stages in processing those same sounds, the listeners disagree.  This is the result of differences in what happens at a deeper level, the associative stage, because, after the sounds are first perceived, the perceptions interact with the listeners’ own idiosyncratic ways of thinking and feeling, ultimately leading to a response (“That’s really swinging! or “These guys just don’t swing!”)

Harvey Pekar’s suggestions, especially because they reflect so much historical perspective, lead us to question whether a strict definition for jazz is truly workable, especially if it requires that a performance swing. Pekar’s observations force us to consider reclassifying the music of players who DON’T swing, but whose music is improvised and stems from the jazz tradition.  Perhaps it should instead be called IMPROVISED CONCERT MUSIC.  Pekar’s suggestions also cause us to remove so many performances that have loosely been called jazz, that we question the utility of framing a strict definition.  However, we must remember that if the swing requirement is applied with some flexibility, defining jazz by requiring it can INCLUDE most of the music that has traditionally been called jazz.  On the other hand, the more flexible form of the definition is still useful enough to EXCLUDE music, like “Rhapsody in Blue,” that does not swing, though the definition also excludes some music that most listeners call “jazz” (many of Cecil Taylor’s post-1960 works, because they don’t swing, “Come Sunday” from Duke Ellington’s “Black, Brown and Beige,” because it is neither swinging nor largely improvised, Manhattan Transfer’s vocal recreations of instrumental jazz, because Manhattan Transfer is not improvising, and so on).

A “Family Resemblances” Approach

Because practical application of the strict definition requires an impossible-to-attain degree of precision in specifying swing feeling, and a considerable loosening in the way many of us ordinarily define improvisation, and because, even with such loosening, it still excludes many styles that have traditionally been accorded the jazz designation, perhaps we should take an entirely different approach.  Perhaps we should not seek common elements in all the styles that have ever been called jazz, but instead we should seek ways in which some styles resemble and differ from others, and in which those others resemble and differ from still others.  Instead of searching for a single fiber that continues throughout the entire thread of jazz history, perhaps we should just satisfy ourselves with finding links between adjacent styles, and thereby appeal to Ludwig Wittgenstein’s metaphor that “the strength of the thread does not reside in the fact that some one fibre runs through its whole length, but in the overlapping of many fibres” (PHILOSOPHICAL INVESTIGATIONS, Macmillan, 1953, page 32a).  In other words, there may be no hidden unity in the diverse forms called “jazz.”

Ludwig Wittgenstein believed that, for many words, the USE of the word WAS its meaning.  He suggested that some words simply resist strict definition, and we must content ourselves with observing their use instead of trying to frame definitions.  It would be inconsistent with Wittgenstein’s approach to force a change in the way a speaker uses the word “jazz.”  To the person who is using the word, jazz is whatever he thinks it is, whatever he is using it to describe.

Acceptance of the above perspective leads us to consider that perhaps some music is called jazz simply because it has similarities with some of the musical events that were previously called jazz.  We know, for instance, that during the 1920s almost any lively popular music could be called jazz. (Some music that today we call vaudeville even came under the jazz umbrella.  Witness the title of the Al Jolson movie about the Jewish vaudeville singer:  “The Jazz Singer.”)  In fact, according to former Lombardo trombonist George West, when Guy Lombardo first came to America, he felt that the music he played was jazz. (In this way of thinking, of all the elements that have ever been felt to help qualify a sound as jazz–syncopation, improvisation, saxophones, drums, blue notes, etc.– at least ONE element must be present for any performance to be called jazz, but no one PARTICULAR element must ALWAYS be present, i.e., no SINGLE element is necessary and no SINGLE element is sufficient.)  By adopting such a stand, we would be able to include ANY style that has ever been called jazz, thereby placating everyone who would be outraged at the suggestion that certain works of Cecil Taylor are not jazz or that certain movements of Duke Ellington’s “Black, Brown, and Beige” are not jazz or that some of Glenn Miller’s biggest hits are not jazz.  We could even include most of the lively popular instrumental music of the 1920s, much of which is routinely excluded by jazz scholars on the grounds that it is not improvised, though it is included by most laymen because it is lively and stems from what F. Scott Fitzgerald dubbed “The Jazz Age.”  We could include “Rhapsody In Blue” because it is bluesy and syncopated, even though it is not improvised or swinging.  And we could include Manhattan Transfer’s vocal recreations of jazz instrumentals because the recreations sometimes swing, even though they lack improvisation.

If you feel that a definition requiring ALL jazz performances both to be improvised and to be perceived as swinging EXCLUDES too much to be useful, then you probably agree with Wittgenstein’s suggestion that some terms cannot be subjected to quasi-mathematical treatment.  The things to be defined are just too varied–like members of a family.  No single trait is necessarily common to each and every one of them.  Applied to the term “jazz,” Wittgenstein’s suggestion would allow a large number of styles to be sheltered under a very comprehensive umbrella.  It would not require that ALL of the styles exhibit even ONE common defining characteristic.  So two ENTIRELY dissimilar styles,  such as Guy Lombardo’s music and Woody Herman’s music, might be linked by a third style, such as big band music in general–or even by a chain of styles in which each pair of adjacent links would share a common feature but in which no single feature would characterize every style.  For instance, the music of the Modern Jazz Quartet does not resemble that of Guy Lombardo, but because of the Modern Jazz Quartet’s use of improvisation and swing feeling, its work does resemble that of the Woody Herman band, which, in turn, resembles the work of Guy Lombardo because both Lombardo and Herman have similar instrumentation.

This “family resemblances” way of characterizing jazz might have its appeals, but, like a strict definition, it carries serious problems with it.  Whatever gains we make in clarity and comprehensiveness, we seem to lose in simplicity and usefulness.  For instance, applying Wittgenstein’s idea of family resemblances to jazz means that two jazz styles need not share any predetermined set of defining characteristics.  Yet in doing this, we leave several questions unanswered.  For example, what are real jazz musicians DOING when they play jazz?  Do they HAVE to SWING?  No.  Is it ENOUGH that they play in bands with musicians who DO swing or improvise?  Maybe.  But remember that NO SINGLE association is sufficient.  Is it important that aspiring performers develop an improvisatory technique?  No, not necessarily.  Is it sufficient that they should COPY the genuine improvisations of others?  No.  There is no single element that is sufficient.

Considering such practical inadequacies, then why IS anyone attracted to a family resemblances approach for characterizing jazz?  Only with sufficient HISTORICAL perspective might a theorist find the family resemblances approach plausible.  After all, it DOES include all the styles that have ever been called “jazz.”  But the thinker who has such historical perspective might, in fact, lack REAL perspective.  An analogy should clarify this point.  Dividing the history of music into periods requires stepping back and seeing the gross DIFFERENCES rather than the long-term continuities.  For example, though Palestrina’s melodic writing is chant-like, his masses don’t belong in the Gregorian Chant bin of the record stores, even though both the chant and his masses derive from the same tradition.  Similarly with jazz, the theorist who is familiar with the ENTIRE span of jazz history can link the playing styles of Fats Waller and Cecil Taylor.  Yet, though both have roots in jazz, their products are so different as to demand separate categories.  This is a realization that the historically minded observer might overlook, especially if he is partial to a family resemblances approach to characterizing jazz.  Might it be more useful to SEPARATE things that are DIFFERENT?  Isn’t that part of what we do when we make a definition?  In other words, the family resemblances approach is good at letting us rule IN styles, but it provides very little help in showing how a given style can be ruled OUT of the jazz category.

Another attraction of the family resemblances approach might stem from observing a timbral similarity among many jazz-related styles.  Perhaps listeners have observed that much jazz is rough in tone and features drums and saxophones.  However, if we make too much of this similarity, we could be led to DEFINE jazz in terms of timbre, despite the fact that TIMBRE has never been sufficient to define a style, though during the late 1960s, the presence of horns in the popular group called Chicago is thought to have been part of the reason the group was classified as a “jazz-rock” group, a band whose music had little else in common with the jazz tradition.  (Furthermore, many jazz sounds are smooth in tone quality and lack drum or saxophone timbres.)

What about “blues tonality” as an element that links many jazz styles?  We might cite this in a family resemblances characterization of jazz.  But no matter what definition of “blues tonality” is used, this feature falls far short of characterizing a majority of jazz styles. (Some listeners call “bluesy” a preponderance of chromatically lowered thirds and fifths in melody lines.  Some call “bluesy” a preponderance of pitch bendings akin to vocal ornamentations common in Afro-American folk music.  Some call “bluesy” a combination of these properties.  See William Tallmadge’s “Blue Notes and Blue Tonality” in THE BLACK PERSPECTIVE IN MUSIC, 1984, 12, 2, 155-165.)  Even during the 1920s, jazz musicians did not choose a characteristically high percentage of chromatically lowered thirds and sevenths.  And the very existence of piano and vibraharp in jazz groups refutes the notion that “neutral thirds” (heptatonically equidistant tones, exemplified by the pitches “in between the cracks” that hornmen occasionally play) are essential for jazz.  When tuned in the standard well-tempered system, these instruments are INCAPABLE of PRODUCING such “blue” pitches.  And frequency-of-occurrence observations overturn the contention that many pianists favor “crushed tones” (simultaneously sounding the major third and minor third, thereby approximating the “blue pitch” sound).  “Crushed tones” are NOT very common in the work of jazz pianists.

So we see that, even though a family resemblances approach to characterizing jazz solves some of the problems that accompany a strict definition, the family resemblances approach CREATES at least as many problems as it solves.  Most of the problems accompanying a family resemblances approach center about the need for simplicity, the attempt to give all scholars a simple definition so that when they begin arguing the specifics of a question in jazz, at least they all know what is designated by the category name itself.

Approaching Jazz As A Dimension

Aspects of both the strict approach and the family resemblances approach can be used in forming a third approach to defining jazz.  This approach hinges on the idea that, of those elements that have been previously associated with jazz, the more that are present, and the more clearly they can be heard, the more a particular performance qualifies as a jazz performance.  In other words, perhaps jazz is not an all-or-none event, but is instead a continuum, a dimension: jazzness.  It accepts the history of musical elements that have been associated with the term jazz, those elements that were found to recur in the links made between adjacent styles when we examined jazz by a family resemblances approach.  This approach acknowledges that different weights are attached perceptually to the various elements when a decision is made by the listener who is trying to determine whether a given performance is jazz.  However, this does not suggest a prioritization for elements, though many listeners attach considerable weight to improvisation and swing feeling, and less weight to instrumentation.  Furthermore, it does not deny the situations in which only one or two elements are present or in which the presence is barely detectable.  This approach simply says that little jazzness exists in such situations.  For example, it is likely that little jazzness would be perceived in Stan Kenton’s recording of “Reflections” from Robert Graettinger’s CITY OF GLASS, yet much jazzness would be evident in Charlie Parker’s “Ko-Ko” (Savoy recording made November 26, 1945).  “Rhapsody in Blue” would be perceived as jazzy, but not unless there were more steady tempo passages and improvisation would it qualify as bearing as much jazzness as Count Basie’s 1937 recording of “One O’Clock Jump.”  Manhattan Transfer’s first recording of Joe Zawinul’s “Birdland” would qualify as jazz because of its syncopations and swing feeling, but more jazzness would be evident in Weather Report’s original version of “Birdland” because of its use of saxophone and improvisations.

We suggest that, in the mind of the listener who is trying to categorize a performance, the jazzness approach derives from

  1. taking each element that has, in the past, been associated with jazz, and accepting it as a dimension of its own,
  2. then weighing the relative representation of each element and
  3. taking the number of elements together as the aggregate stimulus leading to  a listener’s decision of whether to call a given performance jazz.

In other words, for the jazzness approach, improvisation need not have a strict definition.  It can be present in greater or lesser degree.  Swing feeling need not have a strict definition.  It, too, can be present in greater or lesser degree.  Bluesy flavor need not have a strict definition.  Like improvisation and swing feeling, it can be present in greater or lesser degree.  Then a determination is made of the total number of elements present, each taken with respect to how conspicuous it is.  From that, a fix on the jazzness of the performance is made.




We are now in a position to return to the original problems, those issues that remained unsolved, pending an adequate definition of jazz:



Is jazz art music?  To address this question we must have definitions of jazz and art music.  And, depending on which definition of jazz chosen, some of these examples qualify, others don’t.  For instance, can we accept as “jazz” the music of the singing group Peter, Paul & Mary, that won first place in the vocal division of the “jazz” popularity polls run by PLAYBOY MAGAZINE in 1964, 65 & 66?  Should we include the music of Al Jolson from his 1928 film “The Jazz Singer”?  By a strict definition, requiring full-blown melodic improvisation and jazz swing feeling, none of the music of Peter, Paul & Mary qualifies as jazz. Therefore it would not matter how we defined art music.  We could not use those styles to support or refute the contention that jazz is art music.  But, if we followed Wittgenstein’s admonition to look at how the word is used, we would say anything that has ever been called jazz is jazz.  So the Al Jolson and the Peter, Paul & Mary music would be jazz, and we could match their music against art music criteria to argue whether jazz is art music.



Was jazz popular during any given era of twentieth century music? (See “Is Jazz Popular Music” by Mark Gridley, THE INSTRUMENTALIST, 1987, March, Vol. 41, Issue 8, pages 19-22, 25-26, 85.)  This question becomes problematic, especially for music of the 1930s and 1970s, because a number of best-selling records contain brief improvisations, though, because these improvisations might not exceed some critical minimum contribution to the overall character or duration of a performance designated as necessary to qualify that performance for the jazz label, they might fail to qualify their sources as “jazz” records.  Following Wittgenstein, any music that was ever called jazz still warrants the label “jazz.”  So music of the very popular Ted Lewis of the 1920s, Rudy Vallee of later periods, Paul Whiteman of any period, could be called jazz, and because they were popular, even though they employed little improvisation, and far less swing feeling than was projected by the music of Count Basie and Duke Ellington, we could say that jazz was popular during those periods.  During the 1980s, in many record stores that had no “new age” bins, pianist George Winston’s records appeared in the “jazz” bins.  So, by the reasoning associated with Wittgenstein, we also could apply “jazz” to the very popular recordings that Winston made, even though most jazz musicians agree that the music does not swing, and Winston himself refuses to label his music jazz.  And Manhattan Transfer’s tremendously popular vocal recreations of jazz instrumentals could qualify because, even though they are not improvised, they do sometimes swing.  By these kinds of reasoning, then we could say that jazz has been and continues to be popular.  However, if we use a strict definition of jazz, then we find that the music never has received widespread popularity in that, with a few exceptions such as Louis Armstrong, Benny Goodman and Miles Davis, even many of the critically acknowledged greats in jazz rarely sold many records or packed many nightclubs and theaters.



How should we interpret discographer Brian Rust’s statement about Glenn Miller’s output, “…the majority of the Bluebird and Victor sides are of little or no jazz interest…”? (JAZZ RECORDS, Vol. 2, 1970, page 1125; Storyville Publications, Chigwell, England.)  Rust is referring to the fact that improvised solos are infrequent, brief and often constitute inconsequential portions of a given Miller performance.  He is not implying that the band fails to swing, just that the band’s priorities are not on improvisation (personal communication to author, July 13, 1987).  He is also revealing his own adherence to a strict definition of jazz.


We must remember that there are problems with all three approaches.  The worst problem is that any time you define a word you must use other words.  So what plagues all three approaches to defining jazz is precisely what plagues definition itself.  Any attempt to define jazz requires reference to elements which themselves require definition.

The overwhelming strength of the strict definition is its simplicity.  And this is then the most telling weakness of the dimensional approach because the dimension approach does not allow you to make an unequivocal statement that a particular performance is jazz.  The family resemblances approach and the jazzness approach are more relativistic, and they lead to fewer exclusions, having a flexibility that accommodates jazz as an evolving art form, thereby excluding far fewer performances than the strict definition excludes.  The jazzness and the strict definition are both problematic because they both require us to determine the extent to which a given element is conspicuous.  For example, for both the jazzness and the strict definitions, we must determine the percentage of fresh material in every performance in order to address the improvisation requirement, though this is more severe in the jazzness approach than with the strict definition.  However, in the strict definition, the amount of improvisation must exceed a critical minimum before we can call the performance improvised, whereas in the jazzness approach the amount of improvisation affects only the degree to which the performance is likely to be called jazz.

As appealing as any approach might seem, it remains problematic, however, because an adequate definition of swing feeling remains to be crafted.  For example, how is a performance supposed to make you feel before you can call the performance “swinging”?  How is this different from the way a successful performance of any other kind of music should make you feel?  Can we legitimately (arbitrarily) take the rhythmic conception of one particular jazz style–bop perhaps–and adopt it as epitomizing what constitutes jazz swing feeling?  And in determining the contribution that improvisation makes to jazzness, what is the minimum extent of change and the minimum percentage of nonpreset material that qualifies a performance for the label of “improvised”?  And should the listener be able to determine this, just by listening?  Can any change in the practical meaning for these terms can be tolerated from decade to decade, so that we can still call the music jazz, despite how much it has changed from Louis Armstrong to Albert Ayler, from Jelly Roll Morton to Keith Jarrett?

In conclusion, we feel that the definition having the greatest utility for scholars is the strict definition.  Its simplicity allows us to determine what is not jazz, though it excludes much music that the public ordinarily calls jazz.  The reprioritization of elements that is possible with the jazzness approach might reflect most adequately the changing nature of what is regarded as jazz from decade to decade, improvisation being most important sometimes, swing feeling other times, bluesy flavor other times.  That having the greatest utility for fans is the family resemblances approach because it allows them to continue calling “jazz” anything they ever thought was jazz, though the definition by jazzness is probably the way the majority of fans are responding.


The following excerpt “But Is It Jazz?” from Mark Gridley’s Jazz Styles: History and Analysis, 7th Edition (Prentice-Hall, 2000, pages 362-363; Copyright by Prentice-Hall, 1999) illustrates the continuing usefulness of the distinctions made above. It applies these criteria to controversies regarding whether to include acid jazz and the music of Kenny G. It also adds another criterion that is frequently applied: people who do not like a particular style of jazz refuse to call it “jazz” because to use the term implies acceptance.


After fusion had arrived, many musicians and jazz fans did not consider it to be a form of jazz. Most of them eventually softened in their view, but some still consider fusion styles to be separate from the descendants of Dixieland, swing, and bop. Keep in mind, however, that before fusion, many reacted to the avant-garde of Ornette Coleman and Cecil Taylor by saying that it was not jazz, either. But despite the views of some musicians and purist jazz fans, both the avant-garde of Coleman and Taylor and the fusion of Miles Davis, Weather Report, and Kenny G continue to be found in the jazz bins of music stores, not in the classical or rock bins. This indicates that–at least to the outsider–these styles sound more like jazz than like anything else. Were the outsiders missing the distinctions? Or were the insiders missing the commonalities? Incidentally, big band swing had been classified as nonjazz by many Dixieland fans during the 1930s and 40s. They felt that it was not true jazz, partly because it represented a dilution of jazz traits. Collective improvisation did not occupy as much of each performance as it did in early jazz, for instance. Swing big bands emphasized refinement more than spontaneity.

Different people use different criteria for deciding whether a given performance is jazz. For those whose definition requires both improvisation and jazz swing feeling, much music by Weather Report fails to qualify because its rhythmic properties do not resemble the swing era or bop grooves, though they do achieve their own infectious groove. The piano music of New Age stylist George Winston would not qualify because it fails to swing. Music stores, on the other hand, use much looser criteria and display both Weather Report and George Winston in their jazz bins.

Let us examine three other labelling dilemmas. (1) For those who find jazzness in music whenever it uses saxophones and a particular accompaniment style, music by Kenny G qualifies as jazz, even though it would not qualify by a strict definition that requires swing era or bop rhythmic properties. Musicians might be more specific, though, and call music of Kenny Gjazzy pop.” (2) Classification procedures are so loose that jazz journalists once contemplated including rap in the jazz category. Perhaps this was suggested by the African American origins they share. However, (a) most rap does not have melody. It is poetry recited atop a repeating funk rhythm. Though rhythmically compelling, (b) most rap does not swing in the jazz sense, and (c) not much of it is freshly improvised for each performance. (3) If a personal perception of jazzness increases with the number and obviousness of aspects that remind a person of jazz, then acid jazz performances would qualify to the extent that they featured instruments and harmonies associated with jazz, even without improvisation or swing feeling. If any passages conveyed swing feeling, for instance, they would bear more jazzness.

There is another motive for saying something is not jazz despite its roots in the jazz tradition: rejection of pop jazz. To say a style is not jazz is merely another way that some jazz fans recognize that (1) improvisation does not occupy as much of each performance as in more serious jazz and/or that (2) improvisations in it are not as elaborate or (3) as well-crafted or (4) as rhythmically compelling as in other styles. Or it is a way these fans say (5) they do not like it and are refusing to give it their stamp of approval by calling it “jazz.” Some dislike it because of reasons 1 – 4. Some dislike it for other reasons. There has long been a reluctance among musicians and purist fans to include within the jazz category any watered-down variants of a style that derives from the jazz tradition. This was why distinctions were made between swing bands and sweet bands during the 1940s. Count Basie fit the former category, and Glenn Miller fit the latter, for instance. During the 1990s, the same distinctions could be made between saxophonists Michael Brecker and Kenny G. So you see that to be fair in classifying styles, we need to consider the actual characteristics of the music, not just the reactions of listeners. When one listener dislikes a style, this does not necessarily mean that another listener will also consider it bad music. We have also learned that if one listener does not consider a musical style to be jazz, this does not necessarily mean the style will not qualify as jazz for another listener.

Copyright 1999 Prentice-Hall, Upper Saddle River, NJ