This modest volume contains the collected essays on Western music of the most distinguished musicologist this country has yet produced. The topics in it range from medieval music to Verdi, and it is to be followed soon by a second book, only slightly larger, which will assemble Strunk’s articles on Byzantine chant. Together with the introduction to a collection of facsimiles of Byzantine notation and the footnotes and brief introductions to the selections in Source Readings in Music History (an anthology of writings about music from the early Greeks to Wagner), these writings represent the total published work of Oliver Strunk.

This meager harvest is—to some extent, at least—deeply misleading. The footnotes to Source Readings betray a wealth of original research which almost alone would justify a lifetime. Strunk’s influence as a teacher and as a collaborator has been as remarkable as it has been self-effacing. Indeed, the references to him in the work of other historians are, as often as not, acknowledgments of suggestions made, notes and transcriptions lent, even whole pages written for them, rather than page references to his own publications.

Now seventy-three years old, Strunk retired in 1966 after thirty years of teaching at Princeton, but he has continued his researches in Byzantine music while living in Italy, taking up residence near a church where Byzantine forms can still be heard. It is thanks to his work, above all, that we can now decipher the early notation of Byzantine chant with a reasonable assurance of getting the right notes. When Strunk began to teach, musicology barely existed in this country. He helped to create it, and has done more than anyone else to defend the quality of research.

In his writing, he has never been concerned to set down a sweeping panoramic view of a period or to elaborate a large general theory: his restraint makes his work seem bloodless to some of his colleagues. What he has consented to publish is generally the irrefutable establishment of a single fact or of a new method of classifying a given material, both the fact and the method being of seminal importance. But the importance is often left to the reader’s intelligence to discover: it is perhaps the one serious limitation on Strunk’s work that the less one knows about a subject the less one can learn from one of Strunk’s papers on it. He takes an aristocratic view of his profession.

The style of the essays, too, is extraordinarily concise.1 The facts are presented with a dry elegance: Strunk is reluctant to draw from them any conclusions except the most brief and obvious ones, although many of the shortest essays have had far-reaching effects and opened up new fields of research. When there is anything beyond a simple statement of the facts, it is generally to call attention to those aspects of the research which are tentative, provisional, and limited. He does not—in print—so much speculate as block incautious conclusions. The papers are intended to be exemplary—and they are: almost unassailable results are presented with a minimum of provocation (even when they contradict established wisdom) and a maximum of reserve. Each essay appears as solid and self-contained; its radiation is invisible.

Strunk’s approach, his discipline in more than one sense of that word, may be called “philological,” for lack of a better term. “Philology” has been given vague and even contradictory meanings, but throughout the nineteenth century it generally meant the study of old texts and their humane critical interpretation in the light of the culture that produced them. The texts were documents of civilization, to be deciphered, purified of the corruptions of time and incompetent scribes, and interpreted by the historical understanding. In this sense archaeology, history, and literary criticism were necessary parts of philology. “Philologist” may also mean “student of language,” to be distinguished from “linguist.” The philologist is interested not in language in general but in its evidence of a culture that has disappeared. The philologist, unlike the antiquarian, treasures a sense of distance from the past: he does not wish to revive it or transform it for modern use, but to comprehend its continuity with the present.

To approach music in the spirit of philology, as Strunk has done, poses peculiarly difficult problems. Music history as a subject of research is in great part a product of the Romantic revival of the Middle Ages. Since we do not know very much about classical Greek music and almost none of it has come down to us, medieval music has taken the central place in music that classical studies traditionally held in the study of literature. But as in classical studies, one of the chief problems of music history—as distinct from music theory, ethnology, and criticism—was that of deciphering a dead language, or decoding notations whose meanings we no longer understand. For a good deal of medieval and Renaissance music, we are not always sure about the pitches that were sung, and we are uncertain about the rhythms. Nor do we know enough about how the music was performed and with what instruments.


The humanistic side of musical philology has been as important as textual studies, and they have gone hand in hand. For example merely to understand the medieval theorists who help us to comprehend the composition and performance of Gregorian chant, we must enter into the conflicts between rival religious orders, which are reflected in their different aesthetic ideals.

Strunk’s “philological” approach may be seen at its most impressive in a brief paper of 1939, less than six pages long, called “Some Motet-Types of the Sixteenth Century.” (That useful work of reference, the Harvard Dictionary of Music, defines the motet as follows: “As a rule, a motet is an unaccompanied choral composition based on a Latin sacred text and designed to be performed in the Roman Catholic service.”) The opening sentence of Strunk’s essay has a characteristically modest precision:

This paper aims to take a systematic and comprehensive but necessarily superficial look at the motet of the 16th century with a view to defining the general character and extent of the relationship between liturgical situation and musical style.

It has long been recognized that the musical style of a particular piece of Gregorian chant depends less on its text than on its place in the church service, and Strunk asks whether one can find a similar relation for the polyphonic motets of the Renaissance which replaced the chants and were often based on them. During the sixteenth century, he observes, a leveling process took place, and distinctions between kinds of motets—and, consequently, correspondences between the style of a given motet and its liturgical function—become harder to discover. Strunk concentrates on motets written toward the middle of the century, above all those of Palestrina, and demonstrates that it is still possible to observe a typology connected with the place of the motet in the service, and that genuinely different musical styles continue to characterize varieties of motets classified according to their liturgical situation.

In his preface to these essays, Lewis Lockwood remarks about this study of motets that Strunk’s

way of looking at the motet literature inevitably stressed, yet without insisting upon it, the partial preservation of earlier modes of thought and practice within a particular branch of post-medieval music. The paper thus implicitly conveyed a sense of the historical continuity between two eras while yet maintaining a sense of their distance from one another. The result is not simply a means of classification that is heuristically valuable; the essay adds to this a quality of historical richness that gives insight into the musical thought of the period and gains in force from the perception that the musicians of the period themselves had a consciousness of the past that can yet be recaptured by historians.

This sense of the continuity of the past, conveyed through the critical examination and classification of documents, is the tradition of classical philology at its most effective. Lockwood observes further of this paper:

it shapes a concept that could lead to the emergence of a whole field of research. I put that in the conditional since even now, thirty-five years after the publication of the essay, its implications have yet to be taken up with care and insight by specialists in the subject.

Strunk’s lead has not been followed because his approach has become uncongenial to most of us today. To some, the focus appears too narrow: Strunk brings a wide range of cultural as well as specifically musical experience to bear on what appears to be at first sight the solution of a minor problem. Still, the concentration on minor problems is surely as seductive in academic circles today as it has always been. What is unfashionable about Strunk’s thought is its combination of intellectual rigor with a skeptical refusal to commit himself to a particular system. These qualities are closely related: the rigor is itself a manifestation of the skepticism.

If the rigor precludes an imaginative indulgence in the large satisfying generalizations that sweeten the teaching of history, the lack of system is even more disconcerting. It appears unscientific. Musicologists today often apply very restrictive methods to the examination of the past. They are generally very seriously committed to one system or another—to the statistical (and sometimes computerized) analysis of stylistic characteristics, to Schenkerian theories of musical form, or to psychological theories of musical perception. Such commitments are often the principal buttress of a pretension to scientific method. Basic to many of the new programs of research is the attempt to develop methods that can always be successfully carried out, whatever the intelligence of the researcher. This inevitably produces a certain proportion of uninteresting work; but one must add that the old humanistic and more intuitive methods produced their quota of foolishness as well.


Strunk’s use of any system of analysis is narrow but not restrictive; his attitude is best seen in the conclusion of the essay on motets:

To deal effectively with any large body of evidence the historian must begin by putting it in order. In many instances, as in this one, several means of ordering will present themselves. Which one he begins with may in the end make very little difference; before he has finished he will probably have to use them all. In stating the case for this means [i.e., the ordering of the motets according to liturgical function] I do not question the value, indeed great value, of others. But I am persuaded that the use of this means, a means in keeping with the spirit and intention of the works themselves, is the logical first step and, in any event, an essential one.

The italics are mine, and they emphasize a radical view of historical analysis, camouflaged by the reasonable tone of the style. A means of ordering is a category of analysis, a system of understanding, and Strunk’s sense that all means of ordering may be necessary implies a skepticism toward each one in isolation as well as a hospitable acceptance of different modes of thought.

Both the skepticism and the hospitality to new methods are shown with greater force in Strunk’s use of statistical analysis in the essay “Relative Sonority as a Factor in Style-Critical Analysis,” written as long ago as 1949, in what Strunk has called the Stone Age of computing. Starting from the observation of the general and gradual increase in the use of full sonority (of complete triads) in music from the fifteenth to the sixteenth centuries, Strunk asks whether the measurement of relative sonority (the frequency of the complete triad) can “yield trustworthy indications as to date, provenance, and authorship.” He then gives tables of the frequencies for a number of works ranged chronologically from 1460 to 1559: the results bear out the gradual increase but show considerable variation even in works written close together in time. A further table compares the frequencies of three sections of a mass by Josquin, one of which has been generally considered an interpolation by a later composer: the section considered spurious does, indeed, reveal a significantly greater use of the full triad than the other two.

The statistical sample that Strunk uses is small, but the variations in it are significant enough to justify his conclusions:

To attempt now to answer the several questions raised at the outset, it may be said, I think, that with works of unknown authorship, the measurement of “relative sonority” can be used only to confirm conclusions reached by other means. With works of questioned authorship, it can be used effectively to separate the spurious from the genuine, although even here it cannot be expected to be of much help when the actual author of the questioned work is, as often happens, a member of the supposed author’s immediate circle. But with works of known authorship it should prove a useful and reasonably dependable means for establishing a rough chronological sequence.

These distinctions have an importance beyond the limited field of Renaissance music, and bear upon stylistic studies in general—including those in literature as well. It is generally conceded that the study of one isolated trait tends to be unreliable, except when used in conjunction with other knowledge, but it is rare that the value and limitation of a specific statistical study is defined with such precision.

It is, in addition, by no means certain that all our stylistic perception can be quantified or even properly systematized. Much of it depends on an appreciation of what an artist will do when working with an element that is outside his usual norms—the setting of an unusual text, the choice of a theme unlike those he habitually employs. Without Strunk’s understanding of the limitation of a given “means of ordering” the musical material, these aspects of music (and they are among the most significant) are likely to puzzle us unreasonably. Checking one element against another statistically may therefore not be of much use: on the contrary, we should expect that a work eccentric to its period or to its composer’s style in one respect will be eccentric in many others as well. For this reason, we may not presume that more high-powered and sophisticated use of statistics and computers will invalidate Strunk’s sober conclusions. In assessing statistics, we must be prepared (as Strunk remarks in this paper) to take into account the specifically musical problems of each individual work, as well as environmental and individual differences between composers.

The reserve in Strunk’s use of new techniques is in accord with his austere view of the profession of music historian. The modesty of his presentation and the reasonableness of his approach arise from his refusal to take the short cuts, the one-sided views that both the dilettante and the fanatic require in equal measure. His tough reasonableness is seen at its most ironic and most affable in an intervention (not printed in this volume) during a symposium on instrumental music in 1957.2 As in the paper on relative sonority, the subject was the stylistic character of a work of music, the “internal evidence” which enables us to date a work and attribute it to a composer.

The problem raised was the troublesome one of those dubious anonymous works attributed to a composer like Haydn because they resemble some of his music, or are found in a collection that contains some of his authenticated work, or because he has copied out a work for use in performances or study. H.C. Robbins Landon, one of the distinguished Haydn scholars of our time, cited a symphony that appeared stylistically to be clearly by Haydn at the beginning of his career, and that was discovered more or less by accident to be by a very minor figure like Carlos d’Ordoñez. Robbins Landon concluded that the internal evidence of the music itself was unreliable: what was needed in every case was external evidence—documents, manuscripts dated and signed, contemporary catalogues with the composers clearly identified. He called for the compilation of a giant locator catalogue of all the eighteenth-century works available in manuscript or printed sources in libraries, monasteries, publishers’ lists, etc., that would give the opening measures for each work and thereby facilitate the attribution of doubtful works. This would, of course, be an ideal system for identification by computers.

Strunk, whose work on Haydn was done mostly in the early 1930s when he was chief of the music division of the Library of Congress, emphasized at length the dangers of trusting to documents, that is, to external evidence alone, and remarked:

The way in which Mr. Landon proposes to deal with these questioned works is, of course, the best one. There is nothing more satisfactory than to assign them to the man who actually wrote them: that ends the whole argument. But it is a very expensive way of doing it. It costs a lot of money; it costs a lot of time; and it costs a lot of effort. And, one wonders sometimes if the most spectacular results it achieves—as in the case of the “Jena” Symphony3—are really worth it….

A fine instance of what happens with external evidence was suggested to me the other day by Mr. Hertzmann: the Bach Lukaspassion. This is a work that is unquestionably in Bach’s autograph. The watermarks are of no help, and the paper is absolutely correct: it is a Bach autograph from the 1730’s. There is only one thing peculiar about the external evidence: it does not say on the title page that it is by Bach, and I gather from Mr. Mendel that this would be a rather unusual thing in a Bach autograph. The internal evidence, however, will not fit at all. Not only is it inconceivable that Bach could have written it in the 1730’s, but it is also hard to believe that he could have written it at any time….

We have many other illustrations of this kind: almost every composer has written out in his own hand, on his own music paper, compositions by other people. Mr. Hertzmann told me just the other day of a curious case he had heard of recently in Italy: madrigals by Verdi—real Verdi autographs. It turns out, of course, that they are simply copies of madrigals by Palestrina. It seems that one has to use both methods, external and internal, without relying wholly on either one.

The overkill in stylistic studies has increased so much since 1957 that Strunk’s final comment gains in force today:

It is precisely when you are looking for something else that you find out the most interesting things. When you look for a needle in a haystack, perhaps you may turn it up by system, but often you find it simply by kicking the straws around.

It is not the drudgery to which Strunk objected; he had himself spent many years of painstaking work including the copying and transcribing of hundreds of manuscripts. What is at issue is the substitution of drudgery for reasoning—the combination of internal and external evidence and their interaction. As he commented:

Now it would seem to me that if the study of the internal evidence almost invariably leads us to the wrong conclusions, something is wrong with our methods. It puts us in the position of the man who says, “I smoke only Camel cigarettes,” but who cannot tell a Camel from a Chesterfield unless he looks at the brand on the package.

Strunk emphasizes with considerable tact that the experts were largely not fooled by spurious works, and did not need to discover the name of the real composer to know that they were not by Beethoven or Haydn. I do not think Strunk was very much interested in the experts who were fooled, except in one special case for which he laid down an admirable rule:

The internal evidence is, I think, least valuable, when one tries to use it in the conventional stage of the development of a style, before real technical virtuosity and well-defined technical manners have developed. It is unreliable, too, when you work with the very beginning of a composer’s career.

One might add that for a conventional stage4 or for a very young composer with no style of his own who largely imitates many of his contemporaries it does not much matter if we get all the attributions right, which is perhaps why Strunk was unenthusiastic about the project of a gigantic locator catalogue. We ought to know much more than we do about the music of the 1750s, a period that still rests in the shadow of what comes before and after, but the attributions are not the most pressing matter: nothing much in the history of music—or of Haydn scholarship—would have been changed if it had turned out that Haydn had, after all, written that symphony by Ordoñez.

This is perhaps a hard saying for those music historians who think of their task almost entirely as the compilation of a composer’s oeuvre in the exact chronological sequence, and for whom attribution is never to be taken lightly. The sense of what is susceptible of stylistic definition enables Strunk, in the one popular piece in his volume of essays, a brief account of Haydn, to start with the last decades of Haydn’s life and only then to return to the earliest years in which Haydn’s work, even when thoroughly individual, lacks the coherence of the later period.

Strunk’s publications are clearly only a small part of his work and his influence. He has abandoned the results of his researches with reckless generosity to students and friends, not only by his suggestions, but by turning over extensive notes on a variety of subjects as well as transcriptions of medieval and Renaissance scores. He trained many of the most important musicologists now working in a great variety of fields: Lewis Lockwood (in the Renaissance and Beethoven), Joseph Kerman (in the Elizabethan madrigal and eighteenth—and nineteenth-century opera and chamber music), Don Randel and Leo Treitler (in medieval studies), and Kenneth Levy (in Byzantine chant) among many others. I was never really a student of Strunk’s although I knew him when I was at Princeton, but I once wrote several pages on the way Haydn’s use of repeated notes in the upbeats of rondo themes enables him to create magnificent surprises in his finales, thinking this to be all my own idea. I have only remembered while writing this review that the basis for these pages came from a lecture of Strunk’s heard many years ago.

This Issue

February 6, 1975