An Essay on Word Painting1
Any meaningful attempt to appreciate a piece of vocal music must begin with its text, if it has one.2 Listener, analyst, critic and performer must all take the text as their starting point for the simple and obvious reason that that was where the composer almost invariably began. This dictum applies with equal force to solo and choral compositions, old music and new, nothing diminished by a few notable exceptions. It is now taken for granted that the analysis of a vocal work ought to take at least some account of the text that provided its raison d'être and generated much of the music's structure and surface. However, such accounts are rarely more than cursory. In the subjective fog that usually obscures them, we have lost sight of some central issues. First, text-influence has never been taken seriously as one of the basic responsibilities of any genuinely complete analysis.3 Second, until we have written the history of text influences on music, we cannot write a truly comprehensive history of musical style. Moreover, we will never arrive at a genuine understanding of the vast vocal literature stretching from Josquin to Bach without a thorough grasp of one of the more important classes of text-influence: namely, word painting.
The variety of ways in which words may influence music can be formed into a surprisingly symmetrical array (Figure 1).
Figure 1. Classification of Text Influences
In this classification, word painting represents only an option of one of the eight primary surface effects (rhythmic, phonetic, formal, rhetorical, dramatic, lyrical, symbolic, pictorial). These effects rarely occur in isolation, but interact in complex and often changing relationships necessarily suppressed in the diagram. Although that regularly branching array may seem peculiarly persuasive, it does not map some inevitable fact of nature but merely one way of looking at a set of flexible compositional practices. The systematic character of the array is inherent, rather, in our habits of mind and language. For that reason, it seems reasonable to clear the air by looking at some of our assumptions.
Noise is unwanted sound—let's get that out of the way from the start. Music is humanly organized sound, but organized with intent into a recognizable aesthetic entity. Sound consists of no more than transient disturbances of the circumambient air: rapid, minute alterations of pressure whose sum is zero and which contribute only to the entropy of the universe. From this one might easily argue that music is really nothing, leaving us to protest that some nothings are more desirable than others.
Nobody takes this sort of paradox seriously, of course; but, paradox serves rightfully as our point of departure. Notwithstanding the effort expended here to treat this subject systematically, the reader who looks for invariable principles from which to draw unassailable conclusions will find little satisfaction in what follows. Everything in music argues that such rigor is illusory, and only a musician with very limited experience would seriously pursue it. Yet, musical scholarship requires at least a degree of consistency as one of its necessary conditions. I shall attempt to reconcile that requirement with our often hopelessly paradoxical subject.
Let us pursue paradox a little further. We define word painting as the representation, through purely musical (sonorous) means, of some object, activity, or idea that lies outside the domain of music itself. But music itself (as we have "shown") is next to nothing physically; therefore, it can truly represent nothing. Music exists primarily on an aesthetic plane. It represents itself. If it communicates anything more than compulsive twitches and primal screams, it does so by some cultural agreement, by some convention less precise perhaps than that of language; but, like language, ethnocentric and ever evolving.
Language has even absorbed one musical convention so completely that we have to remind ourselves that it is purely artificial; there's nothing in it. Speak, write, reason as we will about "high" and "low" notes in music, nothing justifies those adjectives. All notes emitted by the same source, whether high or low, issue from the same point (for all practical purposes) and radiate into the same air. They do not differ in their placements in physical space—unless we choose to quibble that the piano's higher strings lie to the right of the lower ones. Notes have no altitude. Our stubborn belief in "high notes" and "low notes" as natural and rational expressions in the face of reality, merely emphasizes the ethnocentricity of our convention. The musicians of Classical Greece—a culture probably more obsessed with rationality than any other in recorded history—looked at things the other way around. Our "high" notes were their "low" notes; their scale ran in the opposite direction to ours.4 And with good reason unpolluted by subjective equivocation. After all, when one played the lyre, did not the string with the larger frequency (our higher note) lie closer to the ground? The modern guitarist who plays with the same configuration may well wonder if the Greeks did not have a better word for it after all.
For better or worse, our musical traditions are not those of Classical Greece, but those of the Northern barbarians. The most momentous single phenomenon in the whole history of Western Music took form, probably in the 9th century, when Frankish musical scribes began to distinguish between the sounds they were already calling "high" and "low" by some kind of vertical distinction between their alignments on the page.5 From this seed, already evident in some of the expedients of the anonymous Musica enchiriadis,6 sprang the notations of plainchant and the Ninth Symphony. And all because Charlemagne, an illiterate, wanted all his subjects to march to the same drummer (or at least the same musical liturgy). He was unable to understand why scribes could write down his words, but not the music he heard.7
The leap from Carolingian innovations to the English madrigal may seem chronologically enormous, but it's not so great conceptually. Thus, when John Bennet, in his popular Weep, O mine eyes,8 reaches the lines, "O when, O when begin you / to swell so high that I may drown me in you?" all the parts reach their local high-point on the word "high" (their highest notes in the piece for all but the soprano), and then dip downward to the word "drown." Bennet's writing seems so simply and naturally expressive that no one would condemn his treatment as a mannerism or cliché. Yet this is one of two devices that come most readily to mind at the mention of the comprehensive term "madrigalism." Throughout the 16th, 17th, and 18th centuries, the application of higher notes to words alluding to height or high places (sky, heaven, etc.) and lower notes to contrary meanings, became almost inescapable in mass, motet, madrigal, anthem, lute song, and opera. Almost unlimited numbers of examples testify to the prevalence of this habit which we may label the Directional Convention.
It was surely no Baroque composer who could compose the children's round, "Row, row, row your boat gently down the stream" to an ascending melody! But, our musical instincts about what seems natural and rational suffer another blow when we look for early examples of this presumably inevitable directional convention. It seems to go back no farther than the progressive composers of the Josquin generation. In masses by earlier or more conservative composers, as often as not, ascendit meanders downward, and descendit leaps upward, while caeli et terra have their "altitudes" reversed with a regularity bordering on perversity. The word "perversity" may seem a bit harsh when we remember that the 15th century understood "high music" to mean loud, outdoor music, and "low music" to mean the softer music of the chamber. Nevertheless, the seeming indifference of such important composers as Dufay, Ockeghem, and Obrecht, to so "natural and rational" a convention brings us face to face with a serious problem in the study of word painting: How can we recognize true word painting? That is, how can we distinguish between the purposed musical realization of a text and our own fanciful reading imposed upon fortuitously conformable music? We cannot do so with complete certainty, but we must try to resolve such questions beyond a reasonable doubt, if we hope ever to write the history of one of the chief expressive and organizing features of Renaissance and Baroque style.9 Hence it seems desirable to take account of some present dangers that beset any study of word painting.
The first lies within ourselves as analysts. Even though we know better, we look for rules, laws, unbending principles in the hope that they may lead us to explanations of musical phenomena, while we tend to reject those casual rules of thumb that have the least hint of inconsistency about them. Whether or not consistent laws exist for other musical practices, they do not exist for word painting. It does not depend upon single-valued conventions for which we may compile a vocabulary of definitions. Even the most conventional gesture may have many possible interpretations, or none at all. For example, chromaticism (the other device most often associated with the term madrigalism) often expresses intensely painful emotions. Yet, no matter how widespread that use of chromaticism may have been, the various intensely chromatic keyboard pieces called Consonanze stravaganti by Giovanni de Macque,10 Gioanpietro del Buono,11 and Giovanni Maria Trabacci12 certainly intended nothing beyond the exuberance of chromatic experiment.13 Word painting, then, is not some kind of "code" that "explains" a piece of music. At best, where we happen to understand the conventions the composer has chosen to employ, it allows us a glimpse of his reading of his text—no small gain given the cultural gap that separates us from his style.
From the Romantic era we inherit yet another internal danger. The Romantics turned music into a near sacred vocation and composers into demigods who justified their elevation gloriously, whereas their musical forebears had achieved no less glory as dedicated craftsmen. The Romantics gave us the notion that only purely inspired musical creation—whatever that may be—is worthy of the admiration of the elect; that in comparison with abstractly conceived compositions, literal word painting must seem somehow inferior. Under this intimidating influence Berlioz grew ashamed of the program of his Symphonie fantastique and became the first voice of that chorus that continues to urge us to close our ears to half of its emotion. The program and the music are not incompatible; they nourish each other. Even if a Renaissance composer should have set out to "translate" his text into music, his success rested on the musical virtues of his work, not on the literalness of his translation. Literalness need not—indeed, rarely does—interfere with the musicality of the results. Just as a fine sonata must always surpass mere textbook commonplaces, a successful musical discourse must always transcend those rhetorical ingenuities which may form its substructure or animate its surface. The literalness of some of those details awakens a lurking uneasiness among some who fear it, without foundation, as a usurper of genuine musical values. Nevertheless, we who come after would be ill-advised to spurn its clues to the better understanding of a work of the past. Such scruples do not worry the world of rock, where pretentious musical illiterates peddle the "musical message" of some number according to its words, its costumes, its amps, and the staged frenzy of its performers. However remunerative such writing may be, it serves here merely to remind us that, in the search for authentic musical values, we must avoid looking for the wrong ones. We must accept word painting as a factor that contributes to musical value, but does not and cannot substitute for it.
Another caution: nothing obliges a composer to be consistent. Then as now, nothing required him to treat the same topos in the same fashion from piece to piece, or even within the same piece if it happened to recur. Still, we may not anachronistically imagine him tiring of some particular usage. ("Oh, I did that once; it bores me to do it again!") The composer-craftsman of the past had a more practical respect for the techniques he had been taught or had learned for himself. We may see how he applied some of his techniques, but since we do not yet know the basis on which he chose among his options, we may actually admire his work for the wrong reasons. What we perceive as aptness, his contemporaries may have taken as conventional; where we admire inventiveness, his equally admiring colleagues may have approved his characteristic response to the implications of the text. We cannot hope to appreciate his work fully until we understand what choices he had available to him. Some decisions were made for him by the external history of the piece: the occasion, the preferences of his patron, the forces available, etc. Within these or similar constraints, he had to look at his text not very differently from the way an opera composer looks at a libretto: as a fundamental part of the composition, but always susceptible to some degree of adjustment. The composer may modify the structure of his text in several ways: he may repeat units, he may combine separate text units into a single musical unit, he may break up a unit of text into two or more musical units. As he proceeds, he must decide at every point which particular combination of voices he will assign to each word and phrase of text, with what degree of prominence and emphasis, what texture, what balance, and what particular combination of musical techniques—of which word painting is but one. Melody, rhythm, and harmony, the surface features of the music, must emerge out of this complex background of pre-conditions—or perhaps better: co-existent ideas—often rendered more complicated by the pre-existent materials of cantus firmus, paraphrase, or parody. Word-painting is one of the prominent text-music relationships that we may reasonably expect to help us make sense out of that welter of pre-conditions.
The foregoing remarks imply yet another caution: even in the work of the most doctrinaire madrigalist, word painting is not omnipresent. The presence of a text invites a composer to respond to it, but the composer—whether Ockeghem or Stravinsky—may choose to ignore it. We need to know what elements in a text induce a composer to bend his music to it, but we need to know, also, the nature of the musical situations that lead him to overrule the appeal of the text. Such matters will require years of attention from many scholars.14
The context in which we find a device of word painting raises another question. In a text that presents the words "high" and "low," we would ordinarily expect to hear them sung on relatively higher and lower notes respectively. This usually happens in the normal conduct of word painting, but not always. Let us construct a hypothetical example using the text, "he ran about looking high, and ran about looking low," sung all on one repeated note, except for the word "looking," which sounds a third higher than the repeated note in the first half of the line, and a fourth lower than the repeated note at its second occurrence. If the style of the rest of the piece did not contradict us, we would deduce (probably correctly) that the composer had a textual motivation for the shape of his line. The proximity of the upward and downward leaps to words of parallel imagery seems to associate them with those meanings. But what happens if we increase the distance by placing the leaps on "ran" instead of "looking"? Or, pushing the matter further, suppose each half of the line were set to but a single note, the first a fifth higher than the second? a third higher? a step higher? We may find no difficulty in recognizing pictorial intent in every one of these examples, depending on the context in which we find them. This high-low prejudice eventually becomes so strong that, even where a composer avoids word painting in order to shift emphasis to another point in the text, I believe we may often detect a reluctance to set words like caeli et terra in reversed positions. Terra may occur only a step lower, or even on the same step, but only rarely higher than caeli.
Let us form another artificial example out of the text, "I climb to heaven, only to fall to earth"—a routine job for the practiced word-painter. He would probably look for appropriate melodies rising on "climb" to peak at "heaven," and descending on "fall" to reach a low-point on "earth." But suppose the line had read: "to heaven I climb, to earth I fall"? An attempt to parallel the meanings of the four key words, although not impossible, could lead the composer into awkward constructions. Faced with such a problem, the composer could set this new version of our text to the same sort of music as our first version, even though that would mean leaving "heaven" at the lower end of his upward run, and "earth" at the upper end of his downward run. Here again, the context allows reason enough to read pictorial intent in the passage: upward motion still decorates the upward image, downward motion sets the downward image, and the local discrepancies in the positions of "heaven" and "earth" lose their importance to the larger context. Obviously, one might push such interpretations beyond reasonable bounds; for example, when Ockeghem sets each noun of caeli et terra with both upward and downward motions (as in the Sanctus of his Missa Mi-mi),15 we have no right to label the conformable motions as pictorial and simply dismiss the inconvenient ones. Moreover, Ockeghem's masses taken together do not offer consistent evidence of pictorial concern. Even allowing for the uncertainties in the underlay of the sources, a preponderance of evidence argues against pictorial interpretation.
As one last example of the role of context, let us imagine a poem of unrequited love which places the word "pain" at a verbally significant position. Suppose that position does not suit the composer musically (for whatever reason), and he decides not to apply the usual chromatic motion (or chromatic harmony) directly on the painful word. If he places the chromaticism somewhere in the same musical phrase, we may readily understand it (just as in our preceding example, regardless of what words it sets) as conventional word painting. But, let us suppose he chooses to distribute chromaticism liberally everywhere throughout the piece, but not in the key phrase. Could he possibly have intended the omission as a kind of noema, emphasizing the idea by understatement? But what of all the other chromaticisms in the piece? Do we take them as word painting in the narrow sense, as symbols of pain applied locally, or as a response to the prevailing idea of the pain of unrequited love? It takes a thoughtful reading of both poem and music if one hopes to arrive at a responsible opinion of such treatments. Since word painting can function at many levels,16 an analysis must account for all of them. It is quite possible to miss the forest for the trees; to pass over larger pictorial implications while successfully cataloguing all the local devices.
Underlay presents us with another danger. The sources that preserve older music usually fail to align the syllables of a text with their proper notes, leaving us to guess, most of the time, at what the composer may have intended. We must review all underlay critically, because whether analyzing a composition or making any judgment at all about its word painting, we do so on the assumption that the words and the music that symbolizes them are in proper alignment. As every transcriber of early music knows (or ought to know), underlay is largely an analytical problem. The theorists tell us less than we need to know, while the sources often disagree with them and with each other. This observation takes nothing away from the authority of either the theorists or sources, but it reminds us that we need all the help we can get. We have every reason to suppose that a better understanding of word painting practices would offer some help. If we knew more about the conventions appropriate to the style of the piece being transcribed, we could recognize certain fixed points in the text as sites to which a composer might apply one of the appropriate devices. If one of the common devices of that style should happen to stand not far from appropriate words, we might have reason to underlay it with them—thus reducing the areas of uncertain text placement to the distance from one fixed point to the next. Small help, perhaps, but not to be disdained.
The question of underlay raises yet another hobgoblin: the question of contrafactum which can afflict both plainsong and polyphony. The Renaissance and Baroque habit of substituting a new text for the original one in an already completed composition (in order to make it suitable for another occasion) can make a mockery of a word painting study. However, the case is not so clear-cut. Since the counterfeiter knew the conventions of his time better than we do, he could well have chosen the piece or adjusted the text in such a way as to preserve some word painting of the original—or even to make some appear where the original had none at all. Some of Bach's adaptations of earlier cantatas seem to hold out this possibility. After the English Reformation, some formerly Latin motets acquired amazingly felicitous English texts. Strophic compositions, too, are a kind of contrafactum; in the absence of other evidence, we usually assume that the composer set only the first stanza; that assumption remains untested. Whatever the problem, the validity of our judgments about a text-music relationship hinges—as does every other kind of analysis—on our having the authentic text before us.
The perils arising from such pitfalls lead many to mistrust the subjective element that must affect this sort of study for a long time to come. However, whether in our hypotheses or in our practice, the judicious exercise of subjectivity cannot be avoided in any kind of musical analysis. It becomes abhorrent only when it masquerades as objective science.
What creative value can word painting possibly have had in the process of composition? Its meaning for us depends on that. It must have had considerable value for so many composers of demonstrable genius to have employed it for so many generations. To make the issues a little clearer, let us leave aside the more glamorous or dramatic applications of it in motet and madrigal, and consider a lesson we may draw from the repertory of masses.
A composer would have been likely to set the words of the Ordinary of the Mass many times in the course of his career, and the works of, say a Josquin or a Palestrina, give us reason to believe that a dedicated craftsman would always seek fresh and interesting ways to fulfill that recurrent commission—but within the vocabulary of received contemporary practices. The structure of the Mass text formed a natural basis for the structure of his music, and the conventions that might apply to each segment of text provided him, at each moment, with a set of possibilities which always included the possibility of evading the conventions. Like the sets of standard cadences and the rules of counterpoint, word painting was but one of the options conventionally available to him. If he declined them all—that is, if he embarked upon a path of "free" composition—he followed a most uncommon practice. Composition was rarely completely free in any of the arts. Most often, it made use of pre-existent models or materials in one form or another. A composition formed upon no prior melody or polyphony, and which also disregarded all conventional rhetoric would have been rare indeed. Such compositions usually spring from some constructivist principle, from canon, soggetto cavato dalle vocali, and the like. The idea of total freedom had little aesthetic meaning for a composer. It had no basis in his training or his social outlook before Romanticism reshaped the creative world. Therefore, when we find ourselves tempted to call a passage or a composition "free," it may mean only that we don't know what the composer is doing.
When a composer writing a mass chose some of the possibilities available to him and rejected others, he created emphasis by both his choices and his omissions from the vocabulary of standard conventions. In effect he gave his discourse an individual stamp that had meaning for his contemporaries. Today, many listeners find one 16th-century mass much like another except that the tunes of one may please them more than another. That attitude would have seemed naive in the 16th century, when paraphrase and parody techniques generated whole tangles of masses based on the same tunes. The 16th-century connoisseur undoubtedly enjoyed the tunes in such masses as well as we do, but he gave his critical attention to the fertility of imagination with which the composer renewed and invigorated the given materials, musical and textual. Just as the pre-existent musical materials could be re-worked with endless variety, so could the text of the Mass. A mass was a mass; what mattered was how a composer said his Mass (or motet, or madrigal, etc.). For us to ignore this challenge of musical aesthetics, to ignore the way a composer said his text, is as willfully ignorant and as marginally musical as listening to Schubert Lieder or Verdi operas simply to enjoy the tunes—or worse, just the voices. The fact that unlettered listeners with just such tastes make up the preponderant majority of the musical public does not discredit the assertion. If anything, it bolsters it. To take any other view would be tantamount to redefining the brilliant musical creations of our heritage according to the irresponsible whims of our own age. Of course, we cannot help being of our own time and society. But, if the art of the past has any message worth receiving, it must speak to us in its own voice, not ours. Great works of art do not reach us through movie versions of Moby Dick and Tom Jones or rock versions of the Pirates of Penzance and the Bible. Those are the voices of our culturally anaesthetized kinfolk speaking and listening only to themselves. Let us count our blessings.
Dedicated musicians and musical scholars know that we can never completely recover the past. That is not our goal, nor should it be. Our aim is to communicate with the past as well as we can in order to derive profit from its musical treasures so different from our own—riches we can only simulate. Therefore, since the composers of the past bent themselves to the task of discoursing upon a text to the best of their abilities, we must take pains to understand their responses to that text if we would truly hear their music.*
1The substance of this essay is drawn from the author's book in progress, Music and Words: A Systematic Examination of Madrigalism and Other Text-Influences on Music. It will provide tools for the solution of a number of the problems raised below.
2For some songs truly without words see Owen Jander, "Vocalise" in The New Grove Dictionary of Music and Musicians, edited by Stanley Sadie (London, 1980), 20, p. 51, to which we may add: Stravinsky's Pastorale (1907), and Henry Cowell's Vocalise (1937) and Toccanta (1938).
3As far as I know, the first comprehensive analytical program to allow text-influence a significant role in its scheme is Jan LaRue's Guidelines for Style Analysis (New York: Norton, 1970). LaRue makes a good beginning, although he is concerned chiefly with instrumental music.
4See almost any standard text book. To cite but one: Donald Jay Grout, A History of Western Music, 3rd edition (New York: Norton, 1980), pp. 28 ff.
5For an investigation of the early history of this phenomenon, see: Leo Treitler, "The Early History of Music Writing in the West," Journal of the American Musicological Society, XXXV (1982), 237-279.
6Diplomatic facsimiles of these expedients are conveniently available in the translation of Léonie Rosenstiel, Colorado College Music Press Translations, 7, Albert Seay, ed. (Colorado Springs, 1974), pp. 6 ff, et passim.
7For an account of Charlemagne's role in this development, see: Leo Treitler, "Homer and Gregory: The Transmission of Epic Poetry and Plainchant," The Musical Quarterly, LX (1974), 338 ff.
8Edmund H. Fellowes, The English Madrigal School, XXIII: John Bennet, Madrigals to Four Voices (1599) (London, 1922), pp. 60-62.
9See note 1.
10For a brief example see Willi Apel, The History of Keyboard Music to 1700 (Bloomington: Indiana University Press, 1972), pp. 425.
11For a brief example see Willi Apel, op. cit., p. 493.
12Luigi Torchi, L'arte musicale in Italia, III, p. 372, and also Roland Jackson, "On Frescobaldi's Chromaticism and Its Background," The Musical Quarterly, LVII (1971), 257.
13Although these are keyboard rather than vocal works the argument still holds, for there are many instrumental works with comparable text associations that cannot be challenged; see note 1.
14See note 1.
15Johannes Ockeghem: Collected Works, ed. Dragan Plamenac, 2nd edition (New York, 1966), II, p. 15, mm. 29-38.
16A preliminary version of this idea (refined in the book in progress) appeared in my dissertation: Guillaume Costeley (1531-1606): Life and Works (New York University, 1969), pp.178-185.
*Special thanks are herewith offered to Martin Bernstein, who in many ways influenced the preparation of this study. Celebremus: 1984. xii. 14.