Online Appendices to Hayes and MacEachern “Folk Verse Formin EnglishAppendices”

1、hayes and maceachernfolk verse form in english: appendicesp. 13online appendices to hayes and maceachern, “folk verse form in english”to appear in languagethe material here would not fit into our article “folk verse form in english,” but seems relevant and also perhaps lays to rest issues that might

2、 call the main analysis into question. we therefore have chosen to make it electronically accessible.appendix a: on the reliability of grid transcriptionfolk songs existed largely in an oral tradition, and were sung almost exclusively by individuals who knew no musical notation. the field workers wh

3、o set the folksongs down on paper listened to the songs and relied on their rhythmic intuitions and formal musical training to establish an appropriate graphic depiction of the songs rhythm. in our own work, we examined the musical notation of the field workers, and converted it into the grids on wh

4、ich our analysis is based. the question addressed here is: should the resulting grids be counted as legitimate data, or are they merely a large pile of capricious, subjective interpretations imposed by ourselves and by the fieldworkers?first, with regard to the conversion of musical notation to grid

5、s, the opportunities for any subjective judgments on our part to influence the theoretical outcome are low. the only parts of the conversion task that are potentially subjective are the establishment of grid structure above the levels indicated by musical bar lines (usually, nothing depends on what

6、this structure is), and the division of the temporally continuous grid into lines. in our experience, any subtle judgment calls that arise in the latter task do not affect the classification of the lines by rhythmic cadence, and thus would not affect our results.second, we must consider the harder t

7、ask faced by the musical field worker; that of establishing the rhythmic structure from purely auditory impressions. in general, we think that transcription as carried out by skilled musicians is sufficiently reliable for our purposes. the crucial fact is that the transcription process shows high in

8、tersubjective reliability: when the same song is reduced to notation by independent observers, they usually obtain the same results, differing only in nonessential detail.to elaborate on this point, we wish to describe a quite rare occurrence, one in which different musicians really did come to a di

9、fferent conclusion from the same auditory input.in 1917, cecil sharp, conducting field work in kentucky with maud karpeles, took down the music of “nottamun town” from una ritchie and her cousin sabrina ritchie. in his field notes (published in karpeles 1932, #191a), sharp used a modified version of

10、 9/8 time, later adjusted to straight 9/8 in a version he published for public use (sharp 1923). in 1965, jean ritchie, unas younger sister, published a book of ritchie family songs. she employed two amanuenses, melinda zacuto and jerry silverman, to prepare the musical notation. in notating the lar

11、gely identical version of “nottamun town” that jean ritchie sang, zacuto and silverman used 6/8 time rather than 9/8. this difference implies radically different grid and structural analyses. we translate a couple of lines of the song into grids implied by sharps transcription and by zacuto and silv

12、ermans, abstracting away from some minor details to facilitate comparison:(1)a. sharp: 9/8xxxxxxxxxxxxx|innot-ta-muntown,xxxxxxxxxxxxx|notasoul wouldlookup,xxxxxxxxxxxxx|notasoul wouldlookup,xxxxxxxxxxxxx|notasoul wouldlookdownb. zacuto and silverman: 6/8xxxxxxxxxxxxxxxxxx|infairnot-ta-muntown,nota

13、soulxxxxxxxxxxxxxxxxxx|wouldlookup,nota soulwouldlookup,xxxxxxxxxxxxxxxxxx|nota soulwouldlookdown,our opinion is that these two interpretations are not equally legitimate structural renderings of the same input material; rather, zacuto and silverman committed a gross (though isolated) transcription

14、error. looking at the grids, one can see that the zacuto/silverman transcription places many strong beats where no syllable occurs, and repeatedly slides in and out of parallel with the lines of the song. in contrast, sharps transcription closely aligns musical strong beats with strongly stressed sy

15、llables, and employs a grid with the correct periodicity. zacuto and silvermans error is established vividly if one attempts to sing or chant the words while tapping on a table to what zacuto and silverman claim to be the strongest beats.this, then, is what a genuinely serious transcription error lo

16、oks like. we present it as the exception that proves the rule: in our experience errors of this magnitude are quite rare.the process of transcription, at least as we do it, consists of listening for a sequence of strong beats at evenly spaced intervals, letting this serve as the basis for locating b

17、ar lines (or, for grids, the highest grid levels), then recursively subdividing longer intervals into shorter ones until an appropriate structural location has been found for every syllable or note. assuming that zacuto and silverman did pretty much the same thing, we may conjecture that they erred

18、at the first step of this process, which is what establishes the 6/8 vs. 9/8 distinction. there are two plausible sources for their error. first, 9/8 rhythm is unusual in music, especially in folk songs (we have never seen a 9/8 folksong other than “nottamun town”). second, the song itself contains

19、a metrical irregularity near the end, which under any analysis forces a deformation of the grid. sharp correctly treated this as a local irregularity, whereas zacuto and silverman made it the basis for the entire analysis.appendix b: beyond quatrainsnone of the principles on which our analysis depen

20、ds are intrinsic to quatrains. in principle, then, the same formal patterns we have seen in quatrains should also occur at higher and lower levels. in this section we explore cases of this sort.b.1. higher levelsnot all stanzas are quatrains. apart from the obvious case of two-quatrain stanzas, ther

21、e are stanza types that interpolate extra material or abandon the binary principle in various ways.we first consider large scale long-last effects. these seem to be to some extent an art verse phenomenon; the examples we give either are art verse (3), or look suspiciously like it (1), (2); see reeve

22、s 1958, 21-22). all are english; the appalachian folksong tradition from which we take most of our examples existed at a much greater distance from contaminating literary influences.the first example, which represents a fairly common type, is the two-quatrain stanza in which the first quatrain invol

23、ves two salient couplets, while the second quatrain is salient as a unit (cf. malof 1970:123):(1)4i am a sailor stout and bold,g long time ive ploughed the ocean,4i fight for my king and country too,gfor honor and promotion.4i said: my brother sailor i bid you adieu,4no more to the sea to go along w

24、ith you4 i travelled the country through and throughgand still be a rambling sailor.karpeles 1974, #298ait can be seen that such an arrangement manifests long-last on a large scale: its structure is 4g4g444g, analogous to the gg4g quatrain.another stanza from an english song achieves long-last by a

25、kind of interpolation: a couplet is placed in the penultimate position of a quatrain, lengthening it:(2)4bold nelsons praise im going to sing,4 not forgetting our glorious king,4he always did good tidings bringgfor he was a good commander.4theres sydney smith and duncan too,4lord home and all the jo

26、vial crew,4 and they were the men that were true blue,4full of care, yet i swear,4none with nelson could compare,gnot even alexander.karpeles 1974, #282similar, even more substantial long-last effects occur in written art verse. maceachern 1993 discusses such an effect in macauleys “horatius” (from

27、lays of ancient rome), which is an art-verse poem heavily influenced by the folk tradition. stanza 40 of “horatius” shows a three-line interpolation: (3)3fherminius smote down aruns:3lartius laid ocnus low,4right to the heart of lausulus3horatius sent a blow.3f“l(fā)ie there,” he cried, “fell pirate!3no

28、 more, aghast and pale, 4from ostias walls the crowd shall mark4the track of thy destroying bark. 4no more campanias hinds shall fly4to woods and caverns when they spy3thy thrice accursed sail.”here, we see the erstwhile penultimate line of the stanza (“from ostias walls.”) expanded into a complete

29、quatrain. as a result of this interpolation, the second section obeys long-last on a truly impressive scale (3f344443). in all cases of large-scale long-last constructions known to us, the nonfinal cadences of final unit are 4, conforming to the total long-last cohesiveness constraint.the long-last

30、principle is perhaps also the force behind another major stanza type, the five-line stanza consisting of a quatrain plus a repetition of its final line. what is striking about such cases, which form 5.6% of our folksong data corpus, are the stratagems employed for “gluing” the fifth line into the pr

31、eceding constituent, thus creating a long-last construction:(4)4lord lovel he stood at his castle gate3a combing his milk-white steed, 4and along came lady nancy bell4to wish lord lovel good speed, good speed,3to wish lord lovel good speed. karpeles 1932, #21bthe maximal salient constituents here ar

32、e 43443, forming a long-last construction. the “glue” in question is the reduplication of good speed shown in boldface, which extends what would otherwise be a 3 line (yielding 43433) to a more acceptable 4.the glue that creates a long-last construction is most often reduplication, but it can take o

33、n different forms as well, including an empty epithet or other expression (75a), or occasionally an early start on the next line (75b): further discussion of these line boundary bracketing mismatches, which are relatively rare, may be found in hayes and maceachern (1996; this web site).(5)a.4john ha

34、rdy was a brave and desperated man,3he carried his gun every day. 4he killed him a man in the shunny camps,4this day hes condemned to be hung, i do know,3this day hes condemned to be hung. karpeles 1932, #87b.4well met, well met, my own true love,3well met, well met, says he. 4ive just returned from

35、 the old salt seagand its all for the sake of thee, / and3its all for the sake of thee. karpeles 1932, #35unote that in (75b), the “gluing” effect is limited to the creation of a g line, yielding the less perfect (because less cohesive) long-last construction 434g3. this is true as well in a number

36、of quatrains in which only a single syllable is reduplicated:(6)4rise up, rise up, little matthy groves,3and mens clothing put on. 4it never shall be said in the old scotlandgi slewed a naked man,man,3i slewed a naked man. karpeles 1932, #23afinally, it must be admitted that in a minority of five-li

37、ne stanzas, there is nothing in the text to induce a long-last structure, aside perhaps from repetition itself. thus some 4343 quatrains are extended to 43433, and some 4444 to 44444:(7)a.4o how do you like my fine feather bed?3and how do you like my sheet? 4and how do you like my pretty little wife

38、3that lies in your arms asleep? 3that lies in your arms asleep? karpeles 1932, #23fb.4come all you alabama girls and listen to my noise,4and dont you marry the arkansas boys. 4if you do, your portion shall be4johnny cake and venson is all youll see,4johnny cake and venson is all youll see.karpeles 1

39、932, #75bin many of these, there appears to be a purely musical device to achieve the xxxxx cohesion: the fourth line ends in a non-tonic note (which is less cadential), the fifth on the tonic. but it seems likely that repetition itself, for reasons as yet unknown to us, sometimes provides the neces

40、sary cohesiveness. possibly this might serve as the basis of another fairly common stanza type, in which a normal quatrain is followed by a repetition of its second couplet:(i) 4 o mother, o mother, come dig my grave, 3 come dig it wide and deep; 4 and place fair ellender by my side 3 and the brown

41、girl at my feet; 4 and place fair ellender by my side 3 and the brown girl at my feet; karpeles 1932, #19amore generally, we do not understand the principles that govern repetition in folk stanzas. empirically, the generalizations seem fairly clear: (a) repetition in cadential positions, such as xxi

42、xxi is allowed only for refrain material; (b) non-refrain material may be repeated either in non-cadential position, such as xixxix or xixixix; or it may involve repetition of quatrain-final material with post-quatrain material, as in the stanzas of (74)-(77) or (i) above. just why this distribution

43、 of repetition should be respected is a topic we must reserve for future research.b.2. micro-cadential patterns it was observed by stein and gil (1980) that many chants consisting of just a single line (such as sports or political cheers) have a particular rhythmic structure. the examples below are

44、taken from their article. (8)a. x x x x x x x x x x x x | | | | | one, two, three four five, once i caught a fish alive. b. ho, ho, ho, chi minh! c. sir, sir, ya sadat hail, hail, o sadat (arabic) d. el, el yisrael el, el, israel (modern hebrew) page: 8this is also noticed by attridge, p. 95, who fi

45、rst suggested the parallelas stein and gil point out, other logically possible renditions are considerably less satisfactory: (9) a. x x x x x x x x x x x x | | | | | *one two three, four, five, b. *ho ho ho, chi, minh! c. *ho chi minh, minh, minh! these cases can be understood within our theory wit

46、h a fairly minimal extension: suppose that, just as the quatrain is grouped in binary fashion into couplets and lines, the line is itself grouped into a hierarchy. typically, the line is divided into two hemistichs, which are divided into two dipods or tripods, which are divided into two or three fe

47、et. such a proposal finds support from the same kind of evidence that supports higher level constituency: agreement of metrical constituency with phonological phrasing, and occasionally rhyme (especially rhyming hemistichs). naturally, the agreement of metrical and phonological groupings is more sub

48、tle and less regular at lower levels, but the patterning is definitely present on a statistical basis (guron 1974; kiparsky 1977; napoli 1978; zwicky 1986; hayes and kaun 1996).to make this idea concrete, we give below a folk song verse in which the agreement of phonological bracketing and metrical

49、bracketing is quite salient. the grid has been simplified in noncritical ways for expository purposes.(10) x x line level x x x x hemistich levelx x x x x x x x x x x x tripod level | | | | | | | | | | |one mor- ning, one mor- ning, one mor- ning in may, x x line level x x x x hemistich levelx x x x

50、 x x x x x x x x tripod level | | | | | | | | | | | i spied a fair cou- ple a- mak- ing their way; x x line level x x x x hemistich levelx x x x x x x x x x x x tripod level | | | | | | | | | | |the one was a la- dy, a la- dy so gay, x x line level x x x x hemistich levelx x x x x x x x x x x x trip

51、od level | | | | | | | | | | |the oth- er was a sol- dier, and a brave one i say.karpeles 1932, #145cin a full version of the grid, there would be one lower level to accommodate rapid sequences like was a and and a in the fourth line, and for consistency we would assign this level constituency as we

52、ll, as the foot level. the crucial thing to notice is that at every line-internal hemistich break, the words and syntax have been chosen to place a fair-sized phonological break in alignment with the hemistich break. intuitively, the quatrain reads like (81):(11)one morning, one morning, / one morning in may,i spied a fair couple / a-making their waythe one was a lady, / a lady so gay,the other was a soldier, / and a brave one i say.even tripod breaks are