The Essential Idea of the Blues Formula: A Response to David Evans

I would like to thank the editors of the Journal of American Folklore for inviting me to respond to David Evans’s “Formulaic Composition in the Blues: A View from the Field” (2007), and I also thank Evans for taking the time and energy to write his critique of my work. In the following pages, I will answer Evans’s major arguments, many of which stem from his view of the blues formula—a view at odds with the one I describe in my book (Taft 2006).

I agree with his assessment of my transcriptions: there are many errors and quite a few howlers. But I transcribed what I heard, making use of question marks where I did not feel confident enough to guess, and bracketing words and phrases in asterisks where I was unsure of my transcription. While I obviously would have preferred my transcriptions to be entirely accurate, Evans’s corrections to my transcriptions are, as he admits, “what I heard on the original recordings,” since the singers left no “written versions of their lyrics” (2007:497 n. 2). I am sure that Evans has a better ear than I, and his transcriptions of particular words and phrases may well make more sense than mine. Transcribing for the purposes of formulaic analysis, however, presents a problem: one is likely to hear formulas or formulaic language where they might not exist, because one is predisposed to looking for formulas—a situation that would work against any objective analysis of the blues formula. For this reason, I tried not to prejudge the lyrics and erred on the side of some mishearings.

One of my strategies for countering inaccuracies in the transcriptions was quantitative. By transcribing a corpus of over two thousand blues songs, my percentage of inaccuracies—perhaps 10 percent of the words in any given song—would become statistically less significant than if I had analyzed only a few hundred lyrics. My inaccuracies, in fact, tended to make my corpus seem less formulaic than it actually was, since the words and phrases that I misheard often excluded those parts of a song from identification with any particular formula. For example, considering my transcription of Bo Weavil Jackson’s 1926 recording “Poor Boy Blues,” Evans corrects “Thinking about the wire that my brown had sent” to “Thinking about the words that my brown had said.” My phrase is found nowhere else in the corpus, while Evans’s transcription links it to manifestations of a formula found there: in a line in Jackson’s 1926 recording “You Can’t Keep No Brown” (“Thinking about the kind words that my mama had said”) and also in a line from Jazz Gillum’s 1938 recording “I’ll Get Along Somehow” (“Thinking about the words baby that you have said”). Evans corrects three of my transcriptions to manifestations of two paired formulas, “the blues ain’t here” and “the blues are rambling everywhere.” More accurate transcriptions would have, if anything, added support to the contention in my book that the blues is highly formulaic.

In transcribing sung lyrics, separated as we are by time, culture, and dialect from the singers, and relying as we must on poorly recorded and preserved commercial 78rpm discs, disagreements over the words of a song are inevitable. A more fundamental disagreement between Evans and me, however, lies in the definition of the formula. Evans’s critique of my definition reflects a divide in formulaic analysis between the syntax school and the semantics school, a debate that dates back more than forty years (see my discussion, Taft 2006:31–2) and that stems from two interpretations of Milman Parry’s eighty-year-old definition: “the formula can be defined as an expression regularly used, under the same metrical conditions, to express an essential idea. What is essential in an idea is what remains after all stylistic superfluity has been taken from it” (Parry [1928] 1971:11). Parry’s student, Albert Lord, and Evans (Lord’s student) have interpreted Parry’s “expression” as a “repeated group of words” (Evans 2007:485) and have interpreted Parry’s “essential idea” as the substantive words in a phrase. This interpretation places Lord and Evans in the syntax school.

But “an essential idea,” so stressed by Parry, would seem to imply something beyond syntax and lexis (words and phrases) in defining the formula. For this reason, one should not read Parry without also reading H. L. Rogers’s caveat that Parry’s definition is “crypto-psychological,” especially if used to justify a syntactic and lexical definition of the formula (Rogers 1966). The problem is that syntax and lexis are the end products of language generation and lead to the idea of formulas as set phrases or building blocks for composition (a metaphor that I reject early in my book). A formula must, of course, be manifested in a group of words, but these words are not the formula itself. If we were to consider formulas as nothing more than stock phrases, then blues lyrics would be a concatenation of such phrases, composed in a rather mechanical way, not too different from “desktopblues,” the spoof blues music and vocals generator available on the Internet (Noisegames.com 2007).

Of course, those in the syntax school do not define a formula as an exact repetition of substantive words, but the flexibility they show in this area reveals the semantic underpinnings of the “essential idea” that any formulaic analysis requires. Take, for example, the blues formulas “I woke up this morning / I got up this morning” or “I went to the station / I went to the depot.” Most analysts would at least entertain the idea that these pairs are the same formula, despite the lexical variations in the substantive words, “woke up/got up” and “station/depot.” But in accepting the possibility that these pairs are the same formula, the analyst necessarily delves below the level of lexis to examine the essential ideas of “emerging from sleep” for the first pair and “moving toward a place where trains stop” for the second pair. The analysis of these pairs is a semantic one. The essential idea of the Parry definition is the “essential idea.”

The semantic view of the formula involves fuzzier, less certain distinctions than allowed by the syntax school, because it lacks the clear outlines of syntactic or lexical sameness. Yet it represents more accurately the complexities of blues composition. Blues singers and their audiences traded in manifestations of ideas, and those manifestations, more often than not, were a limited set of all the possibilities available. Blues singers were far more likely to express the emergence from sleep as “woke up this morning” or “got up this morning” than as “left my bed this morning” or “escaped from the arms of Morpheus this morning.” This self-imposed limitation, as well as the never-ending discourse in the blues on love, movement, and anxiety, are two reasons why the blues is so formulaic. But it is still impossible to say with certainty that “woke up” and “got up” are the same formula—only that they are two manifestations of the same essential idea. One singer (or audience member, including the blues scholar) might have seen these two phrases as the same, while another might have seen important semantic distinctions (one can wake up without necessarily getting up). I make this point quite clearly in my book when I write, “the blues formula can be both a theoretical construction and a concrete reality of poetic structure” (2006:55).

Evans believes that I am hedging my bets here (2007:486), but I am not betting at all. Rather, I am expressing the inherent uncertainty of formulaic analysis. Evans’s assessment of my analysis of the “woke up/got up/rose this morning” complex is that these phrases are “clearly variants of a single formula,” because they “combine the ideas of waking up and morning” (2007:487). I think he is right, but I would not use the word “clearly” to describe the surface-level variations on the semantic predication, “emergence from sleep at the beginning of the day.” There is enough semantic wiggle room, as I indicated above, between “wake up” and “get up” and “rose” to create some “crypto-psychological” doubt. Notice, by the way, that Evans links these variants through shared “ideas”—a mark of semantic rather than syntactic analysis.

Evans cautions that applying Chomskyan theories to the study of blues formulas “leads us down a blind alley” (2007:486). Some of Chomsky’s concepts, however, work quite well for describing the formulaic qualities of the blues. The basic principle of transformational grammar, in which deep linguistic structures generate surface-level language, addresses the process of turning an idea or thought (“emergence from sleep”) into a phrase (“woke up this morning”). I have used this Chomskyan principle in combination with Geoffrey Leech’s generative semantics (1974) as a descriptive tool rather than as a full-blown linguistic analysis, and instead of being a blind alley, my approach has yielded a well-lit path toward understanding the blues formula.

What helps to light this path is another Chomskyan principle: the relationship of competence to performance. At the level of competence, which involves an understanding of blues semantic structures, blues singers manipulated a soupy mixture of similar ideas, such as the ideas inherent in the predication, “movement towards some place.” Evans takes me to task for labeling this particular idea as a blues formula; however, I write that this formula is the most diffuse of my examples (2006:113), and I subdivide this formula into more exacting predications (e.g., “go to a place name,” “go to a compass direction”) to show the possibility of more than one formula at work in this complex. As this example demonstrates, at the level of competence the blues formula is difficult to grasp, which is why it is so important to examine the blues formula at the level of performance as well. Evans has no complaint with my formulaic analysis of the singer Garfield Akers; this singer’s performance of highly formulaic blues is evident and illustrates the transformational properties of blues performance in general.

Evans claims that I misrepresent his views on the blues formula and that he has a long history of analyzing this structural element. Yet his stance has always been, in my opinion, overly cautious and tentative. He quotes his own work: “we must be cautious about applying the oral formulaic theory to the folk blues in more than a general manner” (2007:488, quoting Evans 1982:315). He writes that “no one had felt the need to offer a precise, separate [from Parry and Lord] definition of a blues formula” (2007:486), and that he “did not want to overuse the term ‘formula’ in writing about the blues, since it had become so well established in the study of oral epic through the work of Parry and Lord” (2007:489). His approach has led him to a vague notion of the blues formula, which in his view might be a half-line, a whole line, or a stanza. Yet, while shying away from the application to the blues of the pioneering theories of Parry and Lord, he still considers himself a formulaic scholar in the Parry-Lord mode. This tentative approach, however, has resulted in a lack of rigor in Evans’s study of blues formulaic structure.

The rigor that I applied involved an analysis, with the help of a computer, of over two thousand lyrics. Poring through computer printouts and analyzing both the surface and deep structures of thousands of half-lines took a fair amount of sitzfleisch. Evans writes that I demonstrate “in greater length and quantity than anyone previously, the existence of half-line formulas in the blues and the range of variation of certain formulas” (2007:490). He adds that “Perhaps I [Evans] and others should have devoted more attention to half-line formulas in the blues, but this can be done from the armchair and I leave that subject to Michael Taft” (2007:491). Exactly. While I made considerable use of the fieldwork conducted by Evans and others over the last eighty years, I make no apologies for labeling my book an armchair study. The body of interviews conducted with blues singers yields insights into blues structure, but I discovered no rigorous analyses by blues singers of their formulaic patterns, nor would I expect to do so. Most fieldworkers have directed their questions toward other aspects of the blues, and most singers were either uninterested in explaining the intricacies of their choice of formulaic manifestations or were not able to articulate precisely how they put their songs together.

Evans and others have certainly examined formulaic language in the blues. They have traced the use of evocative formulas—or in most cases, formula clusters—by many singers, noting variations and possible influences (see, for example, Springer 2006). But this type of analysis owes more to the general literary method of exploring the multiple expressions of certain themes than to any detailed analysis of blues formulaic composition. As Evans points out, almost all scholars have noted that the major themes of the blues are love, movement, and anxiety (2007:496), and they have arrived at this conclusion through thematic analysis. Evans writes that “it is nice to see [this conclusion] affirmed through [Taft’s] study of formulas” (2007:496), but it is more than nice. Blues singers did not simply pepper their lyrics with these themes; they expressed these ideas in the very semantic structure of their poetry, giving the blues a visceral power not lost on contemporary audiences.

Evans’s structural analysis of the blues is most rigorous at the macrolevel of the entire blues song rather than at the microlevel of the half-line formula. As he points out, I am more skeptical about such analyses (2007:495). I am, for example, less willing to assign the kind of symmetrical structure that Evans discovers in Jefferson’s “Shuckin’ Sugar Blues” (2007:492–4). I do not believe, however, that Evans and I are particularly far apart in our views on blues song structure. We both believe in associative links from stanza to stanza, although he sees less free association than I do in blues composition. My view is that the mere fact of juxtaposition of any two stanzas in the blues leads to an association between them, regardless of the content of those two stanzas, and that, in general, blues singers are “cubist” in their treatment of subjects— that is, they present multiple viewpoints on love/motion/anxiety through a string of stanzas. Evans’s New Criticism-inspired close reading or explication de texte of the Jefferson song finds a kind of symmetry, based on associations from one stanza to the next, but his observation that the song would be more perfectly symmetrical if Jefferson had transposed two stanzas seems to work against his analysis: Jefferson did not, in fact, sing the song with Evans’s “perfect symmetry” (2007:494). Evans seems to object to my comment that one would have to psychoanalyze the singers to understand the stanzaic connections in their songs, yet he admits that the symmetry he sees is a result of “principles . . . used on an unconscious level, not articulated by the singers” (2007:491).

Evans detects a “major error” in my severing commercially recorded blues from the folk blues (2007:489). One of the main principles of modern folklore scholarship, however, is that context influences performance, and the context of the recording studio within the larger context of commercialization affected almost every aspect of the blues. The alchemy that turned a locally performed tradition into one of the most influential forms of American popular culture changed the blues singers’ strategies in composing their songs, just as it changed almost every aspect of the tradition. And commercialization, in turn, increasingly affected local performance in the years after 1920. Evans’s observation, for example, that the singing of most folk blues runs between two and four minutes (2007:490) is most likely a result of the influence of the two-hundred-second limitation found on a commercial 78rpm recording rather than an inherent property of folk performance. Clearly, the folk-popular interchange is no better seen than in the blues. Evans implies that the folk blues, because it was an oral tradition, was more formulaic than the commercial blues. That may be the case; further formulaic analysis would have to be carried out to confirm this suspicion. But Evans’s assumption is based on the long-held view that formulaic structure is a marker of orality. Yet through my description of the highly formulaic nature of the commercial blues, which was, with a few exceptions, written and rehearsed before being performed in the studio, I demonstrated that there is no relationship between formulaic structure and orality (see esp. Taft 2006:285–7).

Somehow, Evans interprets my link between commercialization and blues formulaic composition as a negative view (2007:496), but I am not at all negative in this respect. Prewar commercial blues singers produced one of the finest bodies of American poetry in the twentieth century and did so using a highly formulaic structure. I applaud that achievement. At the 1972 annual meeting of the American Folklore Society, when I first made public my understanding of the blues formula (Taft 1972), I was roundly criticized from the floor by the ballad scholar Albert Friedman—no friend to formulaic studies (see Friedman 1961)—because he thought that my analysis destroys the essential beauty and artistry of folksong. I take the opposite view: a quantitative, microlevel examination of blues poetry enhances our understanding of the aesthetic shared by singers and their audiences (even the disembodied audiences of popular culture) and brings out the beauty and artistry of these performances. I do not know that current blues scholars are interested in traveling down this road, but I would be gratified if my book does indeed help to reanimate the discussion of blues formulas.