Blues lyric poetry: A Concordance


Michael Taft

(New York. Garland. 1984. 3 vol.)



The history of this concordance is a long one. The present volumes have grown­ - perhaps "evolved" is a better word - out of fifteen years of studying and thinking about the lyrics of commercial "race record" blues. Over the years of listening to and reading blues lyrics, I came to realize that the blues singers employed a type of formulaic structure in the composition of their songs which was somewhat similar to that of epic singers far removed in space and time from these Afro-American artists. While I understood this formulaic system in an intuitive way, I found it difficult to describe in a concrete, quantitative fashion the extent and nature of the blues for­mulaic system. My problem was that in order to discover those linguistic forms which were semantically or syntactically close to each other, I had to re-order or "decon­struct” lines and phrases in the blues texts. While my auditory sense sparked my intuitive understanding of the blues, I had to be able to visualize the songs in a form different from that printed in anthologies or transcribed from records if I was to find a concrete basis for my intuition.


My first attempt at deconstruction was manual. I laboriously wrote down phrases from the songs - one phrase to a separate sheet of paper. After two hundred songs, I amassed a collection of sheets which filled seven loose-leaf binders. I had succeeded in deconstructing a small portion of the corpus, but it was apparent that I had not succeeded in making these re-ordered texts accessible or "visible" for the kind of analysis in which I was interested. However, in 1974 I met Michael Preston when he passed through Newfoundland. The result of our discussions convinced me that the only way to re-order a massive corpus of texts was to use a computer. This revelation could not have come at a better time, for I was in danger of suffering the same fate as poor Alexander Cruden, the eighteenth-century compiler of a concordance to the Bible, who worked on his magnum opus between bouts in the madhouse (see the life of Cruden in Cruden, pp. 5-10).


Thus it was that I began preparing blues texts for concording; a process which involved keypunching each line of each song onto a card, labelling that card with a code for the singer and song, and shipping the cards to Preston at the Center for Computer Research in the Humanities at the University of Colorado. This process was a long one, and one which was to be repeated in various ways through the years as I added more and more information to the corpus and as I made more and more corrections to each "generation" of the concordance. For those who think that com­puter work is a simple matter of feeding information into a terminal, I would like to point out that part of this process involved driving 40,000 computer cards from Saskatchewan to Colorado because of mismatched computer systems between the University of Saskatchewan and the University of Colorado. Working with computers in the humanities requires strength and endurance.


The word "generations" accurately reflects the growth of this project. The very first blues concordance was an ugly child: printed on large, cumbersome, green­striped computer sheets, all in capitals, full of symbols which should have been sup­pressed in the print-out, and of course full of typographical errors which I never bothered to correct in my keypunching frenzy. In addition, this first generation was split into three unequal parts, since the concordance-generating program, at that time, could not handle more than one thousand texts and my corpus contained over two thousand separate songs.


As ugly as this child was, it was usable, and led to the completion of a formulaic analysis of blues lyrics (Taft, "Lyrics"). But further, more useful generations followed. The fascinating thing about working with computers in the humanities is that as one works, the computer hardware and computer programs become more and more sophisticated. For example, during the course of this project, the keypunch machine, once so ubiquitous on college campuses, became a technological dinosaur (I worked with the last one to be found on the University of Saskatchewan campus) as it was replaced by the terminal. The concordance program also evolved from a difficult, jury-rigged affair to a sophisticated, multifaceted and flexible system for the re­ordering of texts. Thus, I have had the opportunity to race along on a parallel path with a technology whose advancement must be timed in months rather than decades.


But at some point one must cross paths with that technology and decide that the state of the art has reached a stage which comes up to one's expectations. This work, then, is the once-ugly child of years past, now matured and ready for public scrutiny. In other words, the present state of computer hardware and software allows texts to be re-ordered so that structures once invisible can be clearly seen and understood even by a reader who has had no experience with print-outs, terminals or computer language.


The present work re-orders over two thousand commercially recorded blues lyrics of the "race record" era; that is, blues sung by Afro-Americans and produced in special series by the major recording companies in the United States between 1920 and 1942. These special series were intended for the Afro-American record-buying public, rather than for the wider North American market. These two thousand texts are the same ones presented in their more common transcribed, poetic form in the companion volume to this work (Taft, Blues). But for all its maturity, this concordance is still a child in another respect: it is the first published concordance to a body of oral texts and thus the first truly folkloristic concordance. Although it is related to literary concordances in its use of computer technology and in its final format, this concor­dance has grown out of a set of methodologies and theories which reflect folkloristic rather than literary problems.


But what are these problems? The first questions which the compiler of a folkloristic concordance must ask are what is the definition of the text and, of all possible folklore texts, which should be included in the corpus? The literary concor­dance-maker can more simply demarcate his textual boundaries by stating that “this concordance will include all works by author X" or "this concordance will include a sub-set of works (poems, essays, a single novel) by author X." There might be certain problems with texts of "questionable authorship" or "attributed authorship" in literary concordances, but for the most part the corpus is well defined by its author and literary form.


The folklorist, however, has a much harder time defining his corpus. His texts will most likely have been performed by a number of people, perhaps from several generations or from different cultures and geographical areas (imagine, for example, a concordance to ghost legends). In addition, folk-literary forms such as ballad, legend and proverb defy clear definition. Indeed, over the last century, one of the major preoccupations among folklorists has been the definition and redefinition of terms, and the battle is far from over. Inevitably, the folklorist must rely on what Utley called "operational definitions”; that is, the folklorist must set rather arbitrary criteria for inclusion or exclusion of a text from his concordance. One hopes, of course, that the arbitrariness is not random, but carefully considered and based upon a clear under­standing of the folk-literary tradition from which the texts have been culled. In short, although the folklorist must make fundamentally etic decisions in defining his corpus, we must be aware of emic classifications (see Dundes).


Therefore, the texts included in this concordance represent a specific under­standing of the term "blues." From at least 1850, the word has been used to define many different forms of popular song (Taft, "Lyrics," pp. 60-61). I have already explained some of the narrowing criteria which I have chosen for this study: Afro-­American songs, commercially recorded between the years 1920 and 1942. As well, the songs must be of a certain form which seems to define "the blues" from the point of view of the singers themselves. Because I have already discussed the form and structure of the blues in some detail in the companion volume to this study (Taft, Blues), I will only summarize the operational definition here: the blues is a secular song composed of rhyming couplets in which one or both lines of the couplet may be repeated one or more times and in which the couplet itself might be embel­lished with refrains. These refrains may take either the blues-couplet form or any other form, but it is the blues couplet itself which is the defining feature of this song form.


If a song fits all of the above criteria- then it is eligible for inclusion in the concordance. But the problem of criteria raises another question which plagues the folklorist more than the literary concordance-maker: how large should the corpus be? The literary scholar has the luxury of a limited "bookshelf" - the complete works of

an author are finite and knowable. But this cannot be said of folk literature, for the folklorist's bookshelf is either infinite or only partially accessible. How many ballads have been sung? How many jokes have been told? Even if one limits the folkloristic bookshelf through an operational definition - texts performed by certain people at a certain time and having a certain form - chances are that much of this limited book­shelf would still remain inaccessible.


This is certainly the case with the blues. My operational definition limits the number of texts which I might choose for the concordance. But despite the fact that I have eliminated the thousands of non-commercial blues, white blues, post-war blues, and songs called "blues" which do not conform to the couplet structure outlined

above, I have still left myself with a huge corpus of texts - perhaps in the order of ten thousand songs. Although ideally this concordance should include all pre-war, Afro-­American commercial blues, only about one fifth of  the "bookshelf'” was available to me. Because race records are a part of the ephemera of American popular culture, many have completely vanished, leaving only a trace in the ledger books of recording companies. Other texts survive on only a handful of discs which may be too scratchy to reissue or which belong to collectors who have not been canvassed by reissue record­ing company officials.


Because I have relied upon long-playing albums of the blues reissue trade (see the Discography) rather than on the original 78rpm discs for my research, my corpus is not only limited to those songs available on modern phonodiscs but is somewhat distorted by the particular biases of these reissue recording companies. In short, some blues singers have been reissued many times over while others have been ignored. Historically, reissue companies have favored the more "traditional" blues over the more "sophisticated" big-city and vaudeville blues in their reissue schedules. Of course, I purposely searched for albums which contained the less-known (or less popular) blues singers in order to counterbalance this unwanted overlay on my crite­ria for the corpus, but despite my search, there are no examples from such prolific artists as Leothus Green, Viola McCoy or Lucille Hegamin. Other singers, such as Buddy Moss, Mamie Smith or Charlie Spand, are not represented in proper propor­tion to the number of songs they recorded. Still others, such as Tommy Johnson and John Hurt, are probably over-represented (if that is possible) because of their great popularity in the reissue market.


The only solution to this unwanted selectional restriction was to include enough songs in the corpus so that the inclusion or exclusion of any one song or any one artist for that matter would not be significant. Over two thousand songs sung by about 350 singers representing the wide range of race record blues from vaudeville blues to downhome blues ensures that this corpus is, if not complete, then at least representa­tive My decision to limit the corpus to two thousand blues reissued on albums was also based on certain practical considerations. A reissued song is, at least ideally, an acces­sible song; the inclusion in the concordance of some rare, unreissued recording would not help the scholar who wished to check my printed text against his own copy in his collection of blues albums. The time involved in tracking down and transcribing unreissued recordings from record companies and private collectors would have pro­longed a project which has already stretched beyond a decade. Then too, the addition of one or two thousand extra songs would have made this concordance even more cumbersome than it already is. At some point the very size of a reference work becomes a hindrance to its usefulness.


A further problem confronts the folkloristic, concordance-maker: In what form should the oral texts be transcribed and entered into the computer data bank? Just as the boundaries and definitions of folklore genres are vague, so, too, are the bound­aries of any given text. What is part of the oral text and what is "peripheral”, where does an oral text begin and end, and how many performers are actually presenting the oral text? These folkloristic questions have no set or agreed-upon answers (see, for example, Georges). Even after one has decided upon the boundaries of the given texts, one must decide how these oral texts are to be presented on the printed page and how they are to be electronically stored. The ethnopoetic struggles of scholars such as Dennis Tedlock, who has tried to reproduce Zuni texts in a form which captures their expressiveness, is indicative of the folklorist's problem. By contrast, the literary concordance-maker has his texts already printed, his boundaries already de­marcated by the limits of the printed page. The author he is analyzing has chosen the form of the text for him. I do not mean to suggest that the literary scholar has no problems in this area (see Bender's discussion), but his problems are more mechanical and less philosophical than those of the folklorist.


My choices in this matter have again been based upon my operational definition of the blues and upon practical matters. As I stated above, the essence of the blues is the blues couplet. Indeed, the nature of this type of song is such that one might very well define the genre as one big blues composed of a large but finite number of couplets, lines and formulaic phrases; each individual text is but a sub-set of these couplets. Therefore, this concordance, like its companion anthology, analyzed "stripped-down" texts (see Taft, Blues) in which spoken asides, interjections by other performers and parts of the songs which do not conform to the blues couplet structure have been excluded from analysis.  Like its companion work, this concor­dance is more a study of the blues couplet than of the blues song. I would be the first to admit that these stripped-down versions do not represent the true nature of the performed text but in order to visualize the compositional-structural components of the blues couplet - the original intent of this work - such a redefinition of the bound­aries of the text was necessary. Indeed, these texts have been stripped down in another way: despite the number of times a line has been repeated within the couplet, I have analyzed only one singing of that line. In this way, the concordance reveals formulaic and linguistic repetitions in the corpus but not the stylistic repetitions of lines within couplets. From a practical point of view, to have included every repetition of every line would have burdened the concordance with duplicate entries and ballooned its size, thereby hiding the more subtle details of blues lyric structure.


The method of transcription which I used for this concordance reveals the dif­ference between literary and folkloristic problems in representing the text. The con­cordance-maker of printed texts is bound by the orthography and spelling variations used by the author under analysis; this is especially troublesome for analyses of medieval and Renaissance works where there might be several spellings for the same word (see the problem discussed in Preston and Pfleiderer, p. xvii). The concordance­maker of oral texts however is free to standardize orthography and spelling for the sake of consistency and clarity in his analysis. I have previously outlined my philoso­phy of transcription (Taft, Blues); in essence, I have chosen to preserve the morphology of the blues but not its phonology. In this respect, my transcription philosophy is not too different from that recently proposed by Dennis Preston. Thus, Uncle Remus dialect spellings will not be found in this work.


Some might argue that a computer concordance does violence to the text, that its deconstruction distorts literature and dehumanizes the humanities. But as computers become more common tools in humanistic research, these worries will lessen. How­ever, the violence and deconstruction of texts should not be taken lightly. I have already stated that the purpose of a concordance is to re-order a text so that the analyst might visualize it in a new way, but in the case of folklore this jumbling of the text also reveals the way the singer and his audience see the text. Once again, the etic and emic properties of this kind of study come to the fore. Because of the formulaic nature of the blues, there is every likelihood that when a singer sings a phrase or line, both he and his audience recognize that particular part of  the song. Perhaps semi­consciously , they compare this specific singing of the phrase with other singings of that phrase and phrases similar to it. In an instant, the singer and his audience compare the way the sung phrase is juxtaposed with others, both in the song being sung and in other songs which include that phrase. Thus every phrase in the blues has the potential of literary richness far beyond its specific usage in one song. Pete Welding has been one of the few to discuss this property of the blues lyric:


The blues is most accurately seen as a music of re-composition. That is, the creative bluesman is the one who imaginatively handles traditional elements and who, by his realignment of commonplace elements, shocks us with the familiar. He makes the old newly meaningful to us. His art is more properly viewed as one of providing the listener with what critic Edmund Wilson described as "the shock of recognition" a pretty accu­rate description, I believe, of the process of re-shaping and re-focusing of traditional forms in which the blues artist engages.


If one were to illustrate how the audience undergoes this "shock of recognition," how the mental processes of the listener bring about this shock, one would construct something like a concordance. Each word and each phrase would be lined up against all other words and phrases which are similar to it in all the songs in which the phrase occurred. By looking down a page in the concordance one sees in an instant what must occur for the listener at the moment of "shock." Both the singer and his audience automatically re-order and deconstruct the text as it is being sung; that constitutes their method of appreciation and the basis of their understand­ing of the blues.


That the blues concordance is fundamentally emic in its format should come as no surprise. I stated earlier that after listening to and reading thousands of blues lyrics, I gained an intuitive understanding of the structure and meaning of blues lyrics; I could sense the formulaic structure even if  I could not explain it. Uncon­sciously, I was constructing a mental concordance of the lyrics I had heard – adding more and more texts to that concordance as I listened to more and more blues - so that I began to sense the same "shock of recognition" that the traditional blues au­dience must have felt. The computer concordance is simply a concrete representation of this intuitive process.


I suspect that literary concordances are similarly emic. After all, Edmund Wilson was not referring to folk literature when he wrote of "shock." Current theories of reader-response critics would seem to indicate that concordances in general reveal something of how we read. For example, Stanley Fish asks of an utterance, "what does it do?" (p. 75). One thing it does is evoke a series of contexts for that utterance and usages of that utterance which extend beyond the reader's encounter with it in a specific context. Fish sees literature as existing in the temporal flow of the reader's experience:


In short, something other than itself [the written passage], something existing outside its frame of reference, must be modulating the reader's experience of the sequence. In my method of analysis, the temporal flow is monitored and structured by everything the reader brings with him, by his competences; and it is by taking these into account as they interact with the temporal left to right reception of the verbal string, that I am able to chart and project the developing response. (p. 85)


A concordance would help Fish to chart this response, for it reveals not only the one-dimensional "left to right" context of a verbal string, but the three-dimensional place of that string within a body of literature. Of course, for proper reader-response analysis, one would need a concordance of all the linguistic experiences of a given reader in order to chart his response. But even the limited concordances of specific authors and works show the mental processes at work of author and reader.


It is interesting that the computer-stored concordance of which this work is only one manifestation is even more emic than the printed concordance. Just as the blues listener expands his understanding of the song form with every new blues he hears, and just as the listener is capable of several different re-orderings and comparisons at the same time, so, too, the computer concordance has the same capability. The pres­ent study is only one possible way of re-ordering a set number of texts. But the data in computer storage can be expanded by the inclusion of yet more texts, just as a new blues text will be stored in the memory of the listener. As well, old texts in storage can be emended to clear up transcription errors. But more importantly, the concordance-­generating system allows many different kinds of re-orderings. For example, the words under analysis can be listed according to their sequential order in the corpus or according to the alphabetical order of the context which follows the words; they can be presented randomly down the page as they appear in their line contexts or they can be centered and aligned so that they appear in a column. The words can be presented without any context (a simple word index), within the context of their line, or in a more extended context which makes use of as much space on the page as possible.


Phrase concordances analyze repeated strings of three, four or five words. Re­verse concordances analyze words backwards, producing lists of rhyme-words. Letter concordances deconstruct the words themselves and re-order the corpus by frequency of letters. The corpus itself can be altered so that single singers or specific groups of singers can be pulled from the data bank for smaller comparative concordances.




Sample page I of the Concordance on PC



The first sample page shows one possibility: a phrase concordance which lists all four-word strings in the corpus. This sample also shows the option of listing the phrases in the order they appear in the corpus as well as placing these strings in a center-column alignment. This sample shows the strings within the context of their poetic lines, rather than in the more extended context of as much of the song as would fit on a line of the concordance page.



Sample page II  of the Concordance on PC




The second sample page chooses some of the same options as the first - an unextended, center-aligned context -  but it is a selected concordance of the repertoire of only one singer in the corpus, Blind Lemon Jefferson. In addition, the words under analysis are listed alphabetically according to the context which follows the words. (Preston and Pfleiderer give further possibilities for re-ordering texts on pp. 409 -  423.)


Each type of concordance allows one to visualize the texts in a new way. The present work is only one manifestation, only one possibility, among many. In deciding which manifestation to use for this work, I chose one which most clearly shows the structure of blues lyrics. Rather than being a form of poetry in which innovative words and phrases are the norm, the blues relies on formulas, idioms and well-recognized semantic units to convey its meaning and artistry. An "extended KWIC (key word in context) concordance" such as the one chosen for this work presents the blues most clearly as the type of poetry which "schocks" us with repeated and recoignized phrasing. My choice of this type of format was undoubtedly based on the same principles as Duggan, who chose a similar KWIC concordance in his exploration of the formulaic nature of the Chanson de Roland.


The reader will notice that the extended KWIC format comprises a list of cap­italized "head-words" running down the left-hand margin of the page under which are listed specific contexts for this head-word. The word under analysis appears in the center of the page, preceded and succeeded by as much poetic context as will fit on the line of the page. The instances of the analyzed word are listed in alphabetical order according to the succeeding context of the song in which it is found. Thus, on the third sample page hollering about is followed by hollering and; hollering and crying is followed by hollering and screaming.






Sample page III  of the Concordance on PC






Sample page III  of the Concordance on the  Web



Clicking on the second instance of  hollering and crying  brings up the complete transcription of the lyrics, in this case Blind Blake´s Depression´s Gone from Me Blues. 































Sample page III  of the Concordance on the  Web, with the  Concordance-frame scrolled to the right



Every word under analysis is identified according to its place in the corpus. Thus, to the right in the Concordance-frame of the second instance of  hollering and crying appears:


the name of the singer:                  Blind Blake

the title of the song:                      Depression Gone from Me Blues

recording place:                              Grafton, Wis.

date:                                                 c. June 1932                                                                 

record number:                               (L-1476-2) Pm-13137 Bio BLP-12023


On the bottom frame the lyrics for this song are shown.


In general, the word-forms in this concordance correspond to single words as found in a standard American dictionary. The numbers following the head-words give the reader the number of occurrences of that word in the corpus; thus, there are forty-three occurrences of hollering (or 0,017 % of all occurances), but only two occurrences of hollers. Hyphenated word-forms are listed under their individual components; for example, a-hollering­, will be found under a in  the concordance.


For the sake of a clear, uncluttered page I have included very little punctuation.

The colon (:) indicates the approximate place of the half-line caesura, which is characteristic of the blues form (see Taft, Blues, for a more detailed explanation of the caesura). This mark also helps the reader's eye to catch the first and second halves of lines. Asterisks (*) enclose parts of the transcriptions which are questionable: that is, passages which are only educated guesses at what is actually being sung. These hypothetical passages may be one word, as in Them Smoky Hollow women:sure put a *method* on you or phrases, as in *Hollering for a good long-legged man*. Passages which I have not been able to decipher at all are marked by three question marks (???) regardless of how long or short these passages might be; see, for example, hollering (??? you fall). On occasion, I can only decipher a part of a word, which I then mark with three question marks and that part of the word which I can decipher; see, for example, hollering don't you murder me/I'm down in the bottom ???ing for Johnny Rye.

The reader will notice some words and phrases in brackets ([ ]). These passages occur in one repetition of a line but not in another. For example, the phrase, I'm like a [drunk] man  represents the following repeated line:


Most times when I get hungry: I'm like a drunk man acting a clown

Most times when I get hungry: I'm like a man acting a clown


In other instances, the brackets enclose two or more passages separated by commas, as in hollering: People is [raving, hollering] about hard times. Here the singer has substituted one word for another in the repetition of the line:


People is raving about hard times:tell me what it's about

People is hollering about hard times:tell me what it's all about


The value of this new way of visualizing a text is revealed not so much in the use of the concordance to look up a specific word or phrase, but in browsing through the work. By browsing, by randomly flipping through the pages and letting one's eye "be caught" by a particular pattern, one makes discoveries. And these discoveries are all the more significant because they do not grow out of preconceived notions about the texts. The concordance forces one to see what could not be seen before. For this reason, I have not followed the practice of some concordance-makers who omit certain overly common words (a, the, in, I) from analysis. By browsing through those parts of the concordance which analyze these less substantive words, one often finds especially interesting linguistic patterns and congruencies; for example, prepositions such as in and to reveal patterns of  phrasing common in the blues.


For each word in the concordance the percentage of  the total number of  its occurances in the corpus is shown. As one might expect from lyric poetry, the most common word is I, which occurs 9.887 times in the corpus and makes up a total of 4.229 % of the 233.775 words in the entire corpus.

This may appeal mostly to the statistician, but it does tell something of the kind of corpus under analysis. That the blues song is highly repetitive (or formulaic) and that the blues singer chose to limit the themes of his song to a relatively few and use only a limited, idiomatic vocabulary in his compositions is evident from these statistics. The recent concordance to Meredith (Hogan, Sawin and Merrill), by comparison, reveals a corpus which is less structured and repetitive: 188.440 words compared with the blues corpus’s 233.775, yet 17.967 head-words compared to only 6.422 for the blues corpus. (These statistics for Meredith will not be found in the printed concordance but are available from the electronically-stored Meredith concordance at the Center for Com­puter Research in the Humanities at the University of Colorado.) In other words, even though the corpus of Meredith´s collected poetry is 25 % smaller than the blues corpus under analysis it contains almost three times  as many word forms. These statistical comparisons may well indicate some fundamental differences between elite and folk literature.


The information for each song in the Anthology and Concordance is structured as follows:


Name of singer                               Alexander, Texas

Title                                                  Long Lonesome Days Blues

Place and date                                New York, 11 Aug. 1927

Record numbers                             (81213-A) OK-8511 Rt RL-315


Name of singer

Where a song is attributed to a group by Godrich and Dixon, such as Memphis Jug Band, I have listed the song under the lead vocalist for the group, again as cited in Godrich and Dixon. Where a singer does not appear under his own name in Godrich and Dixon, but under the name of a group or another singer, this informa­tion appears in parentheses after the singer's name. For example, see Big Bill Broonzy, song # 20 (State Street Boys).


The title of the song  as given in Godrich and Dixon. Where Godrich and Dixon give more than one title for a song, I have noted only the first title which they list.

Place and date

Information on place and date of recording also comes from Godrich and Dixon unless more accurate information was available from record notes.


Record numbers

The line marked "record numbers" contains four pieces of information. First the matrix or master number of the recording is given in parentheses. This number pinpoints the location and sequence in the daily recording sessions of the record companies and is important in identifying the relationship of the song to the entire output of the race record era. In the above example, the master number is 81213. Within the parentheses following this number is the "take" number or letter which indicates which version of the song sung in the recording studio has been transcribed.

Thus, this was the first (and perhaps only) version of the song recorded by Alexander, since it is marked "take A." Note, however, that the first and second song by Luke Jordan are different takes of the same song sung by Luke Jordan and recorded in succession; they have the same master number but different take numbers (1 and 2). All master and take numbers are from Godrich and Dixon unless more accurate information was available from record notes (as was the case with the first song by Luke Jordan).

The next information on this line is the original catalogue number for the 78rpm recording of the song. The letters before the dash are an abbreviation for the record company or label on which the song was recorded (see Abbreviations for Race Record Labels) and the alphanumerical designation after the dash is the catalogue number. In the above example, OK-8511 indicates record number 8511 in Okeh Record Company catalogue. Where a song was recorded on two or more race record labels, I have indicated only the first label and catalogue number listed in Godrich and Dixon. In some cases the recording was never issued, but remained a test pressing or a master in the possession of the record company. The word "unissued" replaces the non-existent catalogue number in these cases.

The final information on this line is the label and catalogue number of the long-playing album from which the song was transcribed. The label appears as either a two­ or three-letter abbreviation followed by an alphanumerical catalogue designation, or where the catalogue designation contains no letters, as an abbreviation attached by a dash to the catalogue number. In the above example, Rt is the label, Roots, and RL­315 is the catalogue number for the album. For the code to long-playing album abbre­viations, see the Discography and Abbreviations for Long-Playing Album Labels in the preface to the Anthology.




The Center for Computer Research in the Humanities and the Computing Center at the University of Colorado in Boulder as well as the computing centers at Memorial University of Newfoundland and the University of Saskatchewan played important parts in the compilation of this concordance, and I thank them. I would also like to thank the various administrators at the University of Colorado who have supported inter-university and inter-disciplinary scholarship of the kind represented by this project. I am indebted to Samuel S. Coleman, whom I have never met but whose programming skills are evident in this and many other concordances. I must also thank Cathy Preston for good conversations and helpful ideas. I owe my greatest debt to Mike Preston whose care and enthusiasm for this project kept me going, and whose generosity and expertise turned an idea into a finished work.





Bender, Todd K.

"Literary Texts in Electronic Storage: The Editorial Potential", Computers in the Humanities, 10 (1976), 193-99.


Cruden, Alexander.

A Complete Concordance to the Holy Scriptures of the Old and New Testament. New York: Dodd, Mead, n.d. [First ed., 1737]


Duggan, Joseph J.

A Concordance of the Chanson de Roland. Columbus: Ohio State Univ. Press, 1969.


  -    -

The Song of Roland: Formulaic Style and Poetic Craft. Berkeley: Univ. of California Press, 1973.


Dundes, Alan.

"From Etic to Emic Units in the Structural Study of Folktales", Journal of American Folklore, 75 (1962), 95-105.


Fish, Stanley E.

"Literature in the Reader: Affective Sylistics." New Literary History (1970). Rpt. Reader-Response Criticism: From Formalism to Post-Structuralism. Ed. Jane P. Tomkins. Baltimore: Johns Hopkins Univ. Press, 1980, 70-100.


Georges, Robert A

"Do Narrators Really Disgress? A Reconsideration of `Audience Asides' in Narrating." Western Folklore, 40 (1981), 245-52.


Godrich, John, and
Robert M.W. Dixon.

Blues & Gospel Records 1902-1942. Rev. Ed. London: Storyville, 1969.



Hogan, Rebecca S.,
Lewis Sawin and
Lynn L. Merrill, eds.

A Concordance to the Poetry of George Meredith. New York: Garland, 1982




Preston, Dennis R.

"`Ritin' Fowklower Daun 'Rong: Folklorists' Failures in Phonology." Journal of American Folklore, 95 (1982), 304-26.


Preston, Michael J., and
Jean D. Pfleiderer.

A KWIC Concordance to the Plays of theWakefield Master. Contextual Concordances. New York: Garland, 1982.


Taft, Michael.

Blues Lyric Poetry: An Anthology. New York: Garland, 1983.


  -    -

The Lyrics of Race Record Blues, 1920-1942: A Semantic Approach to the Structural Analysis of a Formulaic System. Diss. Memorial Univ. of Newfoundland, 1977.


Tedlock, Dennis, trans

Finding the Center: Narrative Poetry of the Zuni Indians. Lincoln: Univ. of Nebraska Press, 1978.


Utley, Francis L.

"Folk Literature: An Operational Definition." Journal of American Folklore (1961). Rpt. The Study of Folklore. Ed. Alan Dundes. Englewood Cliffs, N.J.: Prentice-Hall, 1965, 7-24.


Welding, Peter J.

"Big Joe and Sonny Boy: The Shock of Recognition." Record notes to Big Joe Williams and Sonny Boy Williamson. Blues Classics BC-21. One twelve-inch 33 1/3rpm phonodisc. Berkeley: Blues Classics, n.d.


Wilson, Edmund, ed.

The Shock of Recognition: The Development of Literature in the United States Recorded by the Men Who Made It. Second ed. New York: Farrar, Strauss and Cudahy, 1955.