"Ethnopaleography" and Recovered Performance: The Problematic Witnesses to "Eddic Song"

Harris, Joseph

Performance-John Foley's "enabling event of performance" (1995:xiv and passim)-is, more than composition or transmission, probably agreed to be the single most important stage in the life of oral poetry-at least if we take the older supply-side point of view and leave the tradition-savvy audience to others. But its importance does not make performance any easier to reconstruct for old and poorly attested oral literature. Nor is it clear in oral-derived poetry exactly how to construe the role of the written (re) composition, apparently endemic to chirographic transmission of such poetry, in its relation to "performance." The case for writing as a kind of performance in the early manuscript cultures of Anglo-Saxon England, as argued by Katherine O'Brien O'Keeffe (1990) and A. N. Doane (1991, 1994a, 1994b; cf. 1998), has been well received, and Doane has suggested a plan for editing Old English verse that resembles the expressive contemporary textual representations of Dennis Tedlock (for example 1971) and even more Dell Hymes's efforts at reconstituting the oral feel of earlier anthropological textualizations (for example 1977/1981; cf. 1994).1 Moreover, John Foley has tried systematically to develop ways of understanding such poetry even without a folkloristically approved living context (1995).

But medievalists and classicists would still give a lot to understand in a simpler and more direct sense exactly how the oral performances they are concerned with took place-to recover, to the degree possible, the viva vox-even though we know ourselves inevitably condemned to deal principally with black marks on parchment. That this "desire" is "driven" by "nostalgia" for unheard voices does not, for me, invalidate the "drive" itself.2 It goes without saying that scholars comb the older sources for descriptions of performance, but questions remain. We apply models from other traditions to fill in the gaps in our direct information, but every such application is unsatisfyingly metaphorical. Recently, however, the "ethnopaleographic" project of Dennis Tedlock has offered some analogy to our problem and a remarkably inventive new way of thinking it through.3 In this article I will attempt to take Tedlock's procedure as a model for a minor exercise devoted to Old Icelandic verse; the differences of situation and scope are, however, so great that my imitation of Tedlock will be pale indeed. I will focus on a single aspect of the performance of the poetry, its possible musical presentation, and on the reliability of witnesses to "eddic song."

In his capacity as translator of the Popol Vuh, Tedlock's work is closely analogous to that of a classicist or medievalist though probably considerably more difficult. The "Council Book" is preserved in an early eighteenth-century transcription (with parallel Spanish translation) of a sixteenth-century alphabetical version (written in the town of Quiche, Guatamala), which itself somehow goes back to hieroglyphic Mayan. Tedlock is eloquent on the difficulties associated with Mayan philology, but, he says, there is "one respect in which the translator of an ancient Quiche text has an advantage over Old World classicists, even with their magnificent dictionaries . . ." (1987:143). That advantage is the possibility of appealing to contemporary Quiche speakers. Their chronological proximity to the language of the old texts (comparable to our separation from Shakespeare) is enhanced by selecting as consultants Quiche speakers who master various more formal areas of Quiche discourse, areas comparable to the Popol Vuh (divining, praying, announcing, but also narrating). Tedlock calls the process ethnopaleography.4 In the course of comparing the reconstructive methodologies of Hymes and Tedlock, Foley is able to capture the general criteria for the exercise of ethnopaleography very well under two heads: there must be continuity between the living "verbal art and its frozen, deracinated textual avatar"; and "the assumption that the performance of contemporary poetic genres can inform the performance of an ancient text bespeaks a conviction that the paralinguistic features are traditional" (1995:22).

Admittedly the possibilities offered by the model of Tedlock's Quiche ethnopaleography appear severely limited as I turn to the putative music of "eddic song." The witnesses I will be interrogating actually deal with a wider range of Old Icelandic verse than implied by "eddic" in this phrase, but I follow the title of Jon Helgason's great essay on this topic in taking "Eddasang" as shorthand for the several relevant types of verse in Old Icelandic (1972). In general the poetry in question is older than the fourteenth century, and some of the eddic poems proper must belong to the pre-Christian period (before 1000 in Iceland). Later examples in the same genres and meters may date from the late middle ages or the early modern period; in fact, we do not always know how to conceptualize "age" for this basically oral and orally-derived poetry. What can be said with some conviction is that medieval poetry has been known in manuscript in Iceland ever since the thirteenth century; but an absolute terminal date can hardly be set since manuscripts were circulated and copied well into the modern period. In short, no dark age totally interrupted eddu frdi or eddic learning; in that sense both Foley's criteria-"direct genetic relationship" and "tradition"-would seem to apply here. This conclusion is, however, premature for the epiphenomena of musical performance.

But wait a minute! Iceland is an advanced western society; it has no analogue of, for example, Quiche's traditional ajk'ij "daykeepers," "men and women who know how to interpret illnesses, omens, dreams, internal bodily sensations, and the workings of the Maya calendar" (1987:144). (Asatru and similar modern forms of Nordic paganism are revivals and fail the criterion of connection.) We cannot expect the Icelandic native speaker of today to contribute more than, perhaps, an intonation pattern for natural prose, and even there we cannot be sure that the centuries have not changed everything. No contemporary Icelander could be expected to testify about "eddic song," except via the usual scholarly means. The witnesses I intend to call are, however, not contemporary, though still post-medieval; but as Icelanders of the Enlightenment period they would seem to stand in a temporal relationship to the cultural heritage in question similar to Tedlock's daykeepers and to be, like them, experts.

The question of "eddic song" has been discussed by musicologists since the late nineteenth century, and nineteenth-century Icelandic sources report several apparently traditional tunes for medieval poems.5 I will ignore these late and secondary sources and follow instead two unsurpassable modern philological essays on the topic by concentrating on the major eighteenth-century witnesses. The first of these modern articles, that by Dietrich Hofmann (1963),6 deserves to be known by every student of early Germanic verse and goes well beyond the historical evidence to a far-reaching examination of the philological and musicological implications of "eddic song," while the second, by Jon Helgason (1972), is in large part a skeptical rejoinder to Hofmann, but restricted to the historical (as opposed to musical) evidence. My remarks here are intended to make both the eighteenth-century witnesses and these two remarkable modern studies better known and, where I feel I can, to adjudicate between them.
The most direct eighteenth-century witness, then, is the compendious ethnomusicological study (in four volumes) published in Paris in 1780 under the editorship of Jean Benjamin de Laborde, Essai sur la musique ancienne et moderne. Among the national musical traditions sampled there, Icelandic appears with five tunes ("Chansons anciennes Scandinaves, comme on les chante encore en Islande" [II:397 (ch. 9)]), which accompany one sample stanza each from-in order-the skaldic (1) Death-Song of Ragnarr Lodbrok and from the famous eddic poems (2) Voluspa and (3) Havamal, a skaldic stanza by (4) King Haraldr Hardradi, and a stanza from the beloved skaldic hymn (5) Lilja. The order of items just given was, I suggest, probably meant to be chronological: Ragnarr lived in the nineteenth century (though this poem is now considered a twelfth-century composition); the two eddic poems were (and are) widely assigned to the tenth; Haraldr died in 1066; and the Christian poem is from the late middle ages (c. 1360). Laborde's introduction (presumably based on his Scandinavian sources) makes clear that the tunes are regarded as ancient, that the early medieval poets sang their compositions, and that contemporary Icelanders are preservers of that tradition.7 All students agree that the unnamed source of the tunes was a certain Jon Olafsson, an Icelander living in Copenhagen, and that a series of intermediaries between the informant and the printed product increases the chances of inaccuracy or tampering.8 In fact, the Laborde texts contain quite a few obvious errors, and the text-tune fit poses problems even when we allow for the modern Icelandic linguistic forms. Nevertheless, we have here on the face of it something like the kind of unbroken tradition Tedlock's paradigm depends on, but how reliable is the explicit information that "the ancient Scandinavian songs are still sung in Iceland" and how firm the implicit idea that the singing tradition of these poems reaches back to their origins?
Hofmann has a curious but convincing approach to the question of continuity. First he hypothesizes maximum rupture before a birth of anti-quarian interest in the seventeenth century, laying out the case against continuity (88-90). Then he contravenes this skeptical view by eliminating any other explanation than continuity for the evidence of 1780: The claim that old poems were still sung in contemporary Iceland is unlikely to have been an invention of Jon Olafsson or of the Danish intermediaries or even of an earlier generation of scholars. Knowledgeable Icelanders would have contested any such fraud, and then there is another contemporary witness, to be discussed, supporting the claim of the Laborde book (90-91). As a second rejected possibility Hofmann considers that the old poetry might have come to be sung under the influence of sung rimur verse, which takes hold in Iceland in the late middle ages. Hofmann's rejection of this explanation is motivated by the nature of rimur tunes, but in my opinion the putative connection between sung rimur verse and sung "eddic" verse has been inadequately studied and remains a gap in Hofmann's argumentation. Finally Hofmann rejects influence from other sung verse horn, outside the complex of folk material. He concludes: "It is therefore very difficult to explain the musical performance of alliterative poetry as a secondary phenomenon on the basis of Icelandic conditions of the eighteenth century or an earlier time. And so one must return to the questions of whether after all old and genuine traditions could not be behind the melodies transmitted by Laborde" (1963:92).

Such continuity is plausible, Hofmann continues, if instead of firm associations of particular melodies with particular poems, there were looser meter-melody connections so that "poems in the same meter could be sung to the same melody," except that the melodies themselves need not be imagined as fixed but may have varied according to melodic formulas within a narrow range (1963:92). Even more loosely, it could have been only the custom of sung performance and the type of singing that was passed on in traditional association with poems of a given meter or with the meter itself as long as it remained in use-that is, was sometimes employed for new compositions (1963:92). Hofmann then surveys early modern Icelandic for metrical continuity and finds it strong for fornyrdislag (exemplified in Laborde by Voluspa), for the meter of Lilja, and for the main skaldic meter, drottkvaett. For the second eddic meter, Ijodahattr (exemplified by Havamal), Hofmann found only weak and indirect clues suggesting continuity. The remaining meter, hattleysa, a variant of drottkvastt, is even less securely attested in modern usage and presents special problems as it is exemplified in Laborde (1963:93-95). Hofmann's inference that continuity of meter involves continuity of melody or singing performance is supported by the second witness, to be discussed below.

Jon Helgason's brilliant study does not directly disagree on the question of continuity; in fact, his pages on the afterlife of the five meters (1972:39-45) add evidence for the actual singing of poems in the meters that did remain in use in post-medieval Iceland though for ljodahattr such evidence remains weak (1972:43-44). (But see the end of this paragraph for the purpose of this survey.) Jon Helgason's skepticism focuses instead on the basic motivation for the five ancient songs of 1780. For he makes it extremely likely that in four of the five cases the choice of text was dictated by a popular and influential early book on "the Vikings," Paul-Henri Mallet's Monumens de la mythologie et de la poesie des Celtes, et particulierement des anciens Scandinaves of 1756. The Danish scholar who collected the tunes "did not say to Jon Olafsson: 'Choose for yourself some Icelandic folk melodies to sing for us,' but rather, 'Sing for us these ancient songs which Mallet has made accessible to the French public!'" (1972:24). Further on Jon Helgason provides a convincing metaphorical elaboration of this scene (1972:35), but his conclusion that "Meter and melody belonged together, not poem and melody" (1972:35) is essentially the same as Hofmann's position. he goes further, however, with the assurance that Jon Olafsson "had certainly never heard Voluspa as a song"; "Instead he had heard the singing of other, newer poems in the same verse form" and would not have considered the exchange of texts a falsification (1972:36). So Jon Helgason's survey of surviving meters (and singing) is framed as an indication of the tunes the singer might have had at his disposal for a nonce application to the poems specifically requested by the Danish "fieldworkers."
The authentic case of Lilja seems to me to caste some doubt on the extremes to which this skeptical conclusion might be pushed. Jon Helgason (1972:24-25) imagines a plausible scenario connecting Laborde's interest in the phrase "Danish language" for what we call Old Norse (II:397) with the inclusion from the poem Lilja, which is not mentioned by Mallet, precisely of a stanza alluding to the dansk tunga; but the influence could as well have run from the stanza to the introductory comments. In any case, the circumstances of collection here do not put a sung Lilja in doubt. Jon Helgason offers a good explanation for the absence in Laborde of a fifth poem from Mallet, in other words for the incomplete agreement with the Laborde repertoire: since this poem, Hakonarmal, was in the same meter as Havamal, the singer lacked another tune for it (1972:24). At the same time, however, this argumentation would seem to support the authenticity of the ljodahattr melody itself and for the singing of ljodahattr. So even though Jon Helgason's main claim-the influence of the Mallet canon on the choice of songs collected from Jon Olafsson-can be considered proven, it seems to leave untouched the central questions of whether old poems were "still" sung in the eighteenth century and how far back that custom might go.

All this relates to another disagreement between Hofmann and Jon Helgason. For during the 1770's there were two Icelanders by the name of Jon Olafsson living in Copenhagen, both learned philologists and noted experts on older Icelandic lore, Jon Olafsson the Elder ("fra Grunnavik," 1705-79) and Jon Olafsson the Younger ("fra Svefneyjum," 1731-1811). Hofmann follows the older scholarship in assuming that the informant for the Laborde book was the Elder on the basis of the only contemporary (or nearly contemporary) testimony, a 1798 reference by Rasmus Nyerup to the supplier of the Laborde tunes as "den lserde Islsenderjohn Olafsen" (43; Hofmann 1963:87-88; Jon Helgason 1972:20-22). The reference could of course still be ambiguous, but the nineteenth-century Icelandic musicologist Olafur Davidsson concluded that "the learned" was probably the Elder because elsewhere Nyerup refers in unambiguous circumstances to the Younger simply as "Olafsen."9 In his earlier book-length biographical account of Jon Olafsson the Elder, Jon Helgason (1925/1926:323) too explicitly follows this tradition in the scholarship of the subject, attributing the tunes to the singing of the Elder. In his biographical dictionary accounts of both Jons, Jon Helgason did not mention the tunes of 1780, but he described the Elder as a man of prodigious memory and legendary learning in Icelandic lore (Jon Helgason 1939:367-69). In the article of 1972, however, Jon Helgason switches Jons and presents, in the strongest possible terms, the case for the Younger, whom I will discuss below, as the second eighteenth-century witness to "eddic song." The evidence on both sides is slight and inferential; but the question of whether the direct testimonies were vested in one or two persons is of some importance. I will return to it, but first to the less obscure Jon Olafsson the Younger and his important book.10
He was born in Svefneyjar, a group of islands in Breidafjordur near Flatey, in 1731, the year after the death in Copenhagen of Arni Magnusson and three years after the great Copenhagen fire of 1728. Jon's older brother Eggert became well known as one of the mainstays of the Enlightenment in Iceland and as a poet. After his Icelandic school and student years, Jon journeyed in to Copenhagen in 1753 where he matriculated as a student at the University, taking his first degree in 1756 and a First on his theology exam in 1765. Beginning in 1768 he received a yearly stipend from the Arnemagnaean Fund for his work on the Institute's editions, and over the years he had important roles in the publication of Heimskringla (1777-83), Hungrvaka (1778), and the great Saemundar Edda (beginning in 1787). Along the way he edited two Latin ecclesiastical treatises (1770 and 1771); his later decades were devoted mostly to an etymological dictionary which burned in the English bombardment of 1807. Jon appears to have weathered this bit of bad luck, so typical of the age; near the end of his life he became a member of the Arnamagnaean Commission, respected and admired by contemporaries. He was buried in Copenhagen in 1811.
The biographical facts suggest that Jon Olafsson the Younger was a steady scholar, a thorough academic, and an organization man. His importance in the present context is based on a book published in 1786 entitled "On Scandinavia's ancient art of poetry: its ground rules, types of verse, language, and mode of delivery" [Om Nordens gamle digtekonst: dens grundregler, versarter, sprog og foredragsmaade]. The book was extensively shaped by the rules of an academic contest (which it won); Jon alludes to this context in his introduction as follows:
This last especially, and I mean historical knowledge, has brought Scandinavian poetry and poetic art into deserved regard by many scholars, both foreign and native, and long ago awakened the wish to see an extensive explanatory description of its nature and structure. The Royal Academy of Sciences (det Kongelige Videmkabernes Selskab) also directed its attention to this when in the year 1782 it advertised the following prize problem: "To demonstrate the essential nature of ancient Nordic poetry and to inquire into its difference from Greek and Roman; then in what it agrees with Anglo-Saxon and primeval German [poetry]-or not" [Monstrare indolem poeseos septentrionalis antiquae, & in ejus a Graeca Romanaque discrepantiam inqvirere, tum in quo cum Anglo-Saxonica priscaque Germanica conveniat vel secus] (p. xi).

In accord with this plan, the core of Jon's study comprises a descriptive poetics. It is justly called "the first modern treatise on Old Icelandic . . . poetry" in Stefan Einarsson's single-sentence evaluation (1957:208). Framing the central poetics are chapters of wider cultural relevance. The still very readable introductory essay is the most general and philosophical part11 and most remarkable in its time for its attitude of moderation on most of the matters it discusses, especially its effort to negotiate a middle way between classicism and gothicism.
The "foredragsmaade" of Jon's subtitle seems very promising; but apart from some discussion of improvisation in relation to poetic difficulty and of the prestige of court skalds, Chapter 5 ("Poetic presentation," "Det poetiske Foredrag") is more concerned with what we would call structure and genre. Jon does, however, conceive of Nordic poetics as a unity based on something he calls, with greater suggestiveness than precision, "song-speech":
In the following I have given the name of Syngesproget to the oldest Nordic poetry; for it is probable (troeligt) that our forefathers, when they kvadu-that, is delivered their poetic language-have used a singing voice quality, which naturally belonged to it. It is for this reason that these poems were called by the ancients galdr, kvida, ljod-words which according to their essential meaning entirely agree with the name I have given. (1786:6)
With this explicitly invented term Jon Olafsson manages to suggest that a relation to music, which he hardly mentions directly, inheres in a poetry, which nevertheless is grounded in unmarked language. More technically, "Syngesproget" applies to the older and simpler verse-fornydalag (our fornydislag plus ljodahattr) and Old English-but skaldic verse too is referred to as "sung" when performed. This does not mean, however, that Jon Olafsson actually argues for ancient musical performance; instead he assumes it, apparently on the basis of the practice of his own age. Another root of his idea of "Syngesproget" may be the elevated sense provided by Latin cano, carmen, and the like, but he probably also has in mind that the verb kveda had come to designate a breakthrough into performance in a rhythmic, half-singing style. Though Jon keeps his focus strictly on ancient verse, he does allude to later Icelandic tradition, and it appears that this constitutes the major influence on his central concept of "Syngesproget."
Both the modern philologists make full use of Jon Olafsson's book in their own arguments. Hofmann maximizes Jon Olafsson's reliability as a witness to "eddic song": "Even if he (against probability) should also stand behind Laborde's melodies and therefore be our sole witness to sung performance of alliterative poetry in the Iceland of his day, we would have to believe him. He was without doubt a conscientious author in pursuit of truth; moreover, it is precisely on this point that he would not have been able to prevaricate without having to fear the contradiction of his fellow countrymen" (1963:89). Hofmann cites the major places where Jon Olafsson speaks explicitly of singing (e.g., "Fornyrdalag of eight units [i.e., fornyrIislag] is still sung and used in Iceland under the name of Liuflings-lag, that is elf-song. . ." [1786:56; Hofmann 1963:93; further example, 1963:95]) and of singing in old-time entertainment (1963:96).

Jon Helgason sees the same glass as half empty: "In Digtekonst Jon Olafsson nowhere discusses the problem of whether ancient poems were sung; yet he touches it indirectly, especially pp. 184-85. The verbs 'read out' and 'sing' are used here interchangeably. [. . .] The expression [Syngesproget, etc.] is based on an hypothesis, not on knowledge of inherited song" (1972:36). But Jon Helgason, like Hofmann, also discusses the passages where Jon Olafsson directly mentions how verses were sung in his own time (1972:37; 1786:24-25, 56-57, 194-95).
Chiefly, however, Jon Helgason's citations from Jon Olafsson are intended to reinforce his identification of the younger man as Laborde's source singer. Several of these citations ought to be examined one by one. Page 27: Laborde's references (1780:412-14) to drottkvaett could be from the author of Digtekonst; on the other hand, this terminology is also standard. Page 28: Digtekonst's usage of capital letters to mark alliteration could be (very imperfectly) reflected in Laborde. But this reasoning implies that Jon Olafsson the Younger had an extensive written hand in the development of the Danish contribution to Laborde, whereas all the other information shows that two Danes collected from live singing (presumably helped along by a manuscript from the Icelander, their singer). The later book may have been following a custom, or the apparent vestiges of systematic capitalization in Laborde may be deceptive. P. 29: Modern Icelandic spellings and a few orthographic oddities interpretable as anachronisms could stem from either of the two Icelanders in question. Page 32: One spelling seems to suggest an orthographic reform recommended by Jon Olafsson's brother Eggert. This is an ingenious argument, but very tenuous, given the state of the text. Pages 32-33: The general coincidence of terminology (demi-rime, rime entier; halvt Rim, heelt Rim) would, to have significance, again have to imply extensive influence from Jon Olafsson the Younger through the Danish intermediaries. Pages 34-35: Reflection of the facts of early Icelandic verse and condescending attitudes toward ballads could have issued from either of the Icelanders in question. None of these passages seem to me to produce a "smoking gun" connecting the author of Digtekonst with the tunes of 1780, and another (1972:37-38) seems rather to show the incompatibility of the strange metrics of 1786 with the musical practice of 1780.

Contrasting textual readings where the same stanzas appear in the publications of 1780 and 1786 were cited by Hofmann as a positive indication that the two witnesses were not from the same person (1963:88). Jon Helgason considered this inconclusive (1972:22), and presumably the other differences between the two sources could be interpreted as Jon Olafsson the Younger's development in understanding.12 But when Digtekonst uses "fornyrdalag" for fornyrIislag plus ljodahattr while Laborde lumps these two meters together as "togmaelt," one could just as well interpret the situation as Digtekonst's correction of the older scholar's mistake. "Togmaelt" and related terms are used quite differently in Digtekonst. Certainly it is remarkable that, given the small world of Icelandic philology in Copenhagen of the 1780's and given Digtekonst's extensive references to the scholarly literature of its time (especially in "Indledning," but also passim), Jon Olafsson the Younger nowhere refers to Laborde's book of 1780 or its tunes despite his basic theme of poetry as "musical language." In addition his metrics seem incompatible with the music of 1780, as even Jon Helgason remarked. Could not differences between the two witnesses be better understood as a silent response of the younger scholar to the older?13

That Jon Olafsson the Elder was now deceased might have muted the younger man's response. But if Jon Olafsson the Younger had been so involved in Laborde's production as to have supplied the system for indicating alliteration by capitals or something as specific as a grammatical form based on his brother's orthographic ideas, would he not have mentioned the publication that preceded his own by only six years? (What scholar could refrain from citing himself?) Nor does Digtekonst make any use of the grammatical reform supposedly introduced (in a single instance) into Laborde's text.
The differing treatment of what was, in the eighteenth century, probably the single most famous poem of the Viking Age, "The Death Song of Ragnarr LoIbroak," may be further evidence that Laborde's singer was not Jon Olafsson the Younger. That poem had been given pride of place by Laborde, his intermediaries, or his informant; but its ten-line stanza had been trimmed to eight lines, presumably to fit the available melody. In Digtekonst, on the other hand, this famous poem is never quoted and mentioned only once in the context of vaguely mythical precursors (p. 15). Jon Olafsson the Younger passed up many opportunities to cite it, for example, in connection with the meter hattleysa (pp. 57-58), with variations in stanza length (pp. 216-17), or with refrains and similar repetitions (e.g., pp. 140-44, 252-53). Above all, it is surprising that Jon Olafsson the Younger passes up other opportunities to cite Ragnarr Lodbrok along with Starkadr as an early Danish poet (p. 11; pp. 239-40). After all, one of his sources, Skaldatal, had placed Ragnarr with Starkadr at the very beginning of Nordic poetry; and given the contemporary fame of Ragnarr's ridens moriar14 and the probable expectations of Jon's Danish judges, these absences are remarkable. The evidence is negative, of course, but for me the best explanation would be that this severe scholar-though not quite a Jon Helgason15 was skeptical of the popular fraud, and his silence is best construed as a correction, not only of his predecessor of 1780, but of all the pre-romantic enthusiasm around this poem. He can cite Mallet's famous book of 1756 approvingly for his own purpose (pp. vii-viii); but he explicitly corrects Mallet's egregious fusion of Celtic and Germanic (pp. 204-05, p. 212). This could be taken as a correction of Laborde as well since Laborde's effort at literary history after the skalds (II:406) includes a Malletesque allusion to "la religion celtique."16 This notion could not have come from the author of Digtekonst, but of course it may be chargeable to Laborde or to the intermediaries. In any case, this idea of the relationship of the two witnesses-that the book of 1786 can be interpreted as revealing reservations about that of 1780-should be given some consideration.17

In his 1972 article Jon Helgason built a case against Jon Olafsson the Elder as informant on very general grounds: Though in youth he had been famous for his memory (20) and knowledge of things Icelandic,18 in his later days he was a "psychically crippled man," cared for by an aging widow, isolated and without means, filling volume after volume with wild etymological speculations (1972:20). It seemed to Jon Helgason unlikely that M. Jacobi, "Secretaire de la Societe Royale des Science de Copenhague" since 1776 (1972:17), "un homme de la Cour," and M. Hartmann, "savant Musicien du Roi de Dannemarck," would have sought out "this frail old man in his solitude in order to have him sing" for them (1972:20). But one could object that this same Jacobi seems to have had no scruples about appealing directly to das Volk for tunes; Laborde mentions, in words that imply some persistence, that Jacobi tried without success to find a Danish folk singer ("M. Jacobi n'a pu parvenir a engager aucun paysan a chanter l'air de cette chanson, pour qu'on la put copier. 'Ils croient, dit-il, qu'on veut se moquer d'eux, quand on les en prie, sur-tout si c'est un homme de la Cour qui leur fait une pareille proposition'" II:398). Jacobi had a collection of 200 folk ballads "which only peasants know how to sing today" (Laborde 1780: II:406-407; Jon Helgason 1972:34), and the Norwegian songs are explicitly attributed to peasants (Laborde 11:407-12, 415). The notion of "the folk" and its connection with antiquity had not attained romantic maturity in 1776-79, the years when the Laborde's Scandinavian collection must have been made, but in his references to the survival of medieval poetry in Iceland and to peasant singers in Denmark and Norway one can sense it just over the horizon. And so it seems to me it would be precisely the "superstitious" (as we learn elsewhere)19 older man, a veritable repository of Icelandic lore, to whom Jacobi would have turned for ancient peasant-preserved songs, just as folklorists have classically associated old age with wisdom and lore.
Digtekonst, on the other hand, is a distinctly modern, analytical work, learned in a broadly European way, and the younger Jon, who was nicknamed "Hypnonesius" in sophisticated interlingual play on the name of his home, is described as a "quiet, frugal researcher" (Jon Helgason 1939). It is not hard to imagine a contrast between the older man, more endowed with folk learning, and the younger, a product of the Enlightenment generation. Hofmann touches on a related point when he quotes Jon Helgason's early biographical entry on JOn Olafsson the Elder: "J.O. disposed over a rich supply of verses, songs, and other elements of folklore; and much of that is introduced on occasion into his works; obviously he would have been able to narrate much more if he had not been ashamed before his enlightened contemporaries.' When Jacobi and Hartmann, commissioned by the French scholar, had asked him for old Icelandic melodies, he for once did not need to be ashamed" (1963:88; Jon Helgason 1939).

Finally, I would like to suggest that Nyerup's phrase "den laerde Islasnder" reflects the Icelandic "frodr madr"-meaning "learned man," of course, but a title more appropriate to Jon Olafsson the Elder's folk-near, home-grown lore. As it happens Jon Olafsson the Younger gives a contemporary definition of this kind of learning:
There is, rightly considered, a survival of this custom, about which the native historical writing speaks here and there, which is still usual today in many places in Iceland. It is that people pass the long winter evenings with old stories and heroic poems, which are not only read out of books but often enough narrated out of the memory alone. Then, if the speaker wants to earn the approval of the listeners, he must strive for a handsome and coherent presentation. On this occasion both practice and natural intelligence are in evidence. Everywhere among the people he [such a speaker] is honored with the name of the Wise and the Knowledgeable, frodr, frdi madr [sic], one who has enriched his mind with old histories and stories. When such a one, old people in particular, arrive on a journey and seek accommodation with a farmer, the stranger is eagerly asked at skemta, that is, to sing or narrate something ancient for entertainment-which the man of the house himself and his wife, as well as their household, will hear with great pleasure. (1786:8; Hofmann 1963:96).
It is perfectly possible that no reference to this tradition underlies Nyerup's phrase; but if it does, the title surely belongs rightly to the older man.20
There is no way to know with certainty who the Labordian informant was or whether there were two, or only one, eighteenth-century Icelander who believed that the ancient poetry was sung in its time as it "still" was in his. It appears, however, from Jon Helgason's investigations (1972: 45-47) that there was a third early (1826) attestation of this belief from a great Danish student of Icelandic, the grammarian Rasmus Rask. Rask, commenting on the poor tune-text fit of the poem by Ragnarr Lo[partial differential]brok, observes that the poem must have been sung to a totally different melody or else that the first two lines of the (ten-line) stanza must have had their own, lost melody. In other words, neither Rask nor Rafn (whose edition of the poem was under review and whose reply Jon Helgason also quotes) seem to doubt either the contemporary or the ancient musical presentation itself. In this negative way, Rafn must be considered a fourth early witness.
The essential question-how far back does the tradition of singing really go-must remain no closer to an answer. But even if Jon Olafsson the Younger was also the Labordian singer, I find that his belief in "Syngesproget" (and the analogous beliefs of Rask and Rafn) makes any kind of fraud unlikely. Jon Helgason may well be right when-in an almost Homeric simile that attempts to capture the "desire for origins" operative here-he purports to hear the cliffs echo back the answer to that essential question with "late Icelandic" instead of "ancient Germanic" (1972:47). But this result itself is a qualified positive if "late Icelandic" refers to a late medieval-early modern continuous tradition rather than to an eighteenth-century scholarly prank or imposture. The Labordian singer and the author of Digtekonst may have projected what they (he?) knew or had heard about the performance of traditional verse forms back upon the Scandinavian "forefathers," and we remain in the dark about how, say, Voluspawns performed about the year 1000. Six or seven hundred years later, however, there is reason to believe that heirs of Voluspa in its meter were "sung" or "kve[partial differential]in."
Ethnopaleography without live informants seems, at least in this experiment, more a study of the informants than an application of their learning to the primary object. And after ethnopaleography reaches its limit there appears to be no recourse but to return to classic philological-musicological speculation, as Hofmann does in the latter parts of his studies. This would be the method, I believe, with which to return with more determination to the rimur connection for some at least possible theories about "eddic song" in the late medieval period.

Harvard University

1. Foley (1995:17-27) has an excellent comparative discussion of the approaches of Tedlock and Hymes to text representation, reconstruction, and "ethnopaleography."
2. see Gottert 1998, especially 15-17. Some American scholars seem to think desire drives only approaches to be deemed erroneous.
3. Tedlock 1982/1983, 1983a, 1987.
4. "Ethnophilology" might seem more appropriate. Cf. Hymes 1965/81 and Tedlock's discussion in Tedlock 1983a:129.
5. For a brief treatment of this topic within the context of eddic performance in general and for further references see Harris 2000; for the nineteenthcentury musicologists see Hofmann 1963 and Jon Helgason 1972. Translations in this article are mine unless otherwise credited.
6. Hofmann 1963 is part of a series with Jammers 1964 and Hofmann and Jammers 1965. The latter parts of Hofmann 1963 and all of the two later members of the series go beyond the historical evidence reconsidered in this article. see also the recent, wide-ranging study by Gade (1994).
7. I quote parts of Laborde 1780, 11:397, interspersed with some remarks of later commentators: "11 n'y en a point de modernes dans ce petit recueil [referring to the whole chapter on Scandinavia]; car celles que l'on fait maintenant, ressemblent aux barcarolles de Venise, & se chantent de meme. [¶] Celles que M.Jacobi a choisies, meritent l'attention des curieux, sur-tout les cinq premieres [the Icelandic tunes]. Elles sont faites par les anciens Poetes Scandinaves, appelle Scaldes [Bjarni -Dorsteinsson (1906-1909:462) comments that Laborde probably is to be understood here as having switched from the tunes to the poems themselves]; il reste un grand nombre de cette sorte de poemes dans les vieilles chroniques du pays. [¶] Elles sont ecrites dans l'ancienne langue Danoise, qui etait celle de tout le Nord, y compris l'Angleterre, & qu'on parle encore en Islande. [Old Norse or even Old West Norse was indeed called "Danish" in some medieval sources, including in the stanza from Lilja quoted by Laborde.] Mais malgre cet avantage, un Islandais a pourtant de la peine a les comprendre, car les Poetes s'etaient forme un langage a part . . . [¶] Ces Scaldes, qui etaient des personnages illustres, chantaient leurs chansons dans les Cours des Princes ... & les accompagnaient du son de divers instrumens; aussi etaient-ils appeles Harpax, c'est-a-dire, Joueurs de harpe." [Bjarni Dorsteinsson, rendering the passage in interpretative paraphrase, seems to capitulate here; no commentator has adequately explained "harpax," and the whole connection with instruments runs counter to the main conclusions of modern scholars. Cf. Joan Helgason 1972:19.]
8. The intermediaries Schutze, Jacobi, and Hartmann, discussed by Hofmann 1963 and more fully byjon Helgason 1972, will be mentioned again below.
9. Olafur Davidson 1888-92:249 and n. 3 there; his conclusion is discussed by Hofmann 1963:88 and at more length by Jon Helgason 1972:21. Olafur Daviosson pointed out that Nyerup regularly and "often" cites Jon Olafsson the Younger's book simply as "Olafsen," and uses the form "den terde Islaender John Olafsen" only once, namely when identifying the source of Laborde's tunes. Jon Helgason attempts to invalidate this evidence by showing that the only meaningful contrast can be between the first citation of the author of the 1786 book, Digtekonst, and that of the Labordian singer so that "often" does not apply. But Olafur Davidsson's observation still stands since the Younger is cited first by Nyerup on p. 4 ("Om disse garnie Skalder anmasrker Olafsen i sit Skrift . . . ") while the attribution of the Labordian melodies to "the learned Icelander Jon Olafsson "follows it on Nyerup's p. 43. Jon Helgason, who knew the sources as no one will again, tried to strengthen his point by showing that in a later book (1808) Nyerup referred unambiguously to the Elder as "den forste Stipendiarius Magnseanus Johannes Olavius fra Grunnevig." But this passage only proves the felt necessity of making the distinction between the like-named scholars. In the crucial passage, 1798:43, I feel sure Nyerup did intend a contrast; and see further my remarks on the title "den kerde" below.
10. My account is based chiefly on Pall Eggert Olason 1950,111:239-40; but also on Jon Helgason 1939.
11. This is the only part cited in Svanhildur Oskarsdottir's recent essay on the Icelandic Enlightenment-to-Romanticism transition as viewed through the treatment of "eddufraedi." This is the only fairly extensive recent discussion of Digtekonst I have found.
12. The gap between the two books, and therefore the likelihood of a "development of understanding," is reduced when one considers that the contest was announced in 1782 and the printed book of 1786 is a reworking of the manuscript submitted for the prize (Jon Olafsson 1786:xi). The reworking appears to be augmentation for a more general audience (pp. xi, xii-xvi). Cf. n. 17 below.
13. At least one explicit instance of the younger man's criticism of the work of the older is recorded in a letter mentioned by Jon Helgason 1925/1926:317.
14. Ragnarr and his death song or Krakumal may be traced in histories of romanticism such as Clunies Ross 1998 and Fell 1996. The two most important seventeenth-century books on Norse culture both celebrated the poem: Bartholin 1689, whose Latin translation of the refrain line I have quoted, and Worm 1636.
15. Stefan Einarsson characterized Jon Helgason as a "sharp-tongued satirist": "A perfectionist as scholar and editor, Jon Helgason's task often was to weed out uncritical theories and assumptions, rather than to advance theories of his own" (1957:313).
16. Laborde was far behind the times on this famous question; even Percy had already refuted Mallet in the introduction to the English translation (1770).
17. A few further observations along the same lines: (1) Digtekonst's characterization of the language of poetry (passim) does not emphasize word order as Laborde does. (2) One passage (pp. 197-98) might be construed as Digtekonst's comment on or elaboration of Laborde's note on "l'ancienne langue Danoise" (II:397). (3) Although the reservation about ballads Jon Helgason heard in "au lieu de ces Poesies . . . on se contentait des rimes" (1972:34) could have been common to Icelanders in general and is echoed in Digtekonst (p. 196, where the specific topic is lack of alliteration in ballads), other references in Digtekonst seem to put ballads in more varied contexts (p. 200: Faroese ballads, which have much alliteration, are evidence of the earlier layer of alliterative poetry in Scandinavia even where that style has died out; p. 247: ballads and related are a poetry of the common people; pp. 252-53: ballads in Iceland resemble older French ballads and Syv's Danish collection). It is impossible from these references to show disapproval of Laborde's remarks on the ballad, but in references to the latest collections of Faroese and Danish ballads (253: "some [are to be found] in the handsome collection which one of our learned and deserving anti-quarians, Herr Sandvig, has brought to light") we might be justified in hearing a note of one-upmanship. This reference, by the way, is to the first volume (1780) of a book edited by C. B. Sandvig and would probably have been written before 1784 when the second volume came out since that volume was edited by Rasmus Nyerup; Sandvig died in 1786. Cf. n. 12 above.
18. Jon Helgason 1939:367-69; biographical entries on both men were signed by both Jon Helgason and Kristian Kalund (JH as first author on the Elder, Kalund on the Younger); the entries are almost unchanged in the third edition of Dansk biografisk leksikon though JH is the exclusive author.
19. Stefan Einarsson 1957:208: "Jon Olafsson Grunnvikingur . . . had been Arni Magnusson's amanuensis, but was as superstitious and uncritical as his master was free from those blemishes." Stefan Einarsson does not elaborate, but the charge of superstition and the uncritical memory attributed to the Elder agree with the picture I am attempting to suggest. On the other hand, cf. the recent publications cited in note 20 below.
20. Bjarni Dorsteinsson (1906-1909:462) was obviously following in the foot-steps of Olafur Davi[partial differential]sson (and of Nyerup, I believe) when he notes that the tunes were transcribed by Hartmann "as 'the learned Icelander' Jon Olafsson fra Grunnavik sang them before [the gentlemen]." Though this passage has no independent evidential value, it does show the reputation of the Grunnavikingur in Iceland, a value supported further on in this text: Laborde clearly states that the five tunes were currently still sung in Iceland, "and there is no cause to doubt that since they stem from Jon Olafsson [fra Grunnavik], who was the best of all men in all Icelandic learning" (463). Another example, with more evidential value, of the Elder's reputation may be drawn from an autobiographical sketch written by Skuli Magnusson; in 1732 he arrived penniless in Copenhagen and went to the student Jon Olafsson, "the honest, learned man," but got only philosophical advice such as would not lift him out of poverty; "yet never in his life was he allotted better or more profound teaching" (Dorkell Johannesson 1947, 11:59; cf. x). (This autobiography was written in Danish, but is available to me only in Icelandic translation.) Skuli Magnusson's testimony is contemporary, but of course falls short of proving that the fra[partial differential]r ma[partial differential] r title was peculiar to the Elder. As it happens another, but modern (apparently 1924), vita in the same collection (1957:VI: 52) could be interpreted as explicitly associating Nyerup's phrase with the Younger: ". . . Jons Svefneyings (Hypnonesius), er kalla[partial differential]ur var 'hinn laer[partial differential]i Islendingur' (+1811)."-Further modern signs of the Elder's reputation in Iceland are to be found in a contemporary group of "friends of Jon of Grunnavik" and their publications; I came across these recent publications only after completing this article, but a rapid reading suggests that they do not bear directly on the problems discussed here though they greatly enrich our understanding of the Grunnaviking himself: Dorunn Sigurðoardottir (1996; especially ix-xxv); Fririk Magnusson and Gu[partial differential]run Kvaran (1994); Gu[partial differential]run Ingolfsdottir and Svavar Sigmundsson (1999); Gu[partial differential]run Asa Grimsdottir (2001).
Bartholin, Thomas. 1689. Antiquitaium Danicarum de causis contempt a Danis adhuc genfilibus mortis libn ires ex velulis codicibus & monumentis bactenus ineditis congesti. Copenhagen: Bockenhoffer.
Bjarni-Dorsteinsson. 1906-1909. Islenzk pjodlog. Copenhagen: Muller.
Clunies Ross, Margaret. 1998. The Norse Muse in Britain 1750-1820. With an appendix on the periodical literature by Amanda J. Collins. Trieste: Parnaso.
Doane, A. N. 1991. Oral Texts, Intertexts, and Intratexts: Editing Old English. In Influence and Intertextuality in Literary History, ed. E. Rothstein and J. B. Clayton, Jr. Madison: University of Wisconsin Press. Pp. 75-113.
_____. 1994a. Editing Old English Oral/Written Texts: Problems of Method (With an Illustrative Edition of Charm 4, Wid Faerstice). In The Editing of Old English: Papers from the 1990 Manchester Conference, ed. D. G. Scragg and Paul E. Szarmach. Cambridge: Brewer. Pp. 125-46.
_____. 1994b. The Ethnography of Scribal Writing and Anglo-Saxon Poetry: Scribe as Performer. Oral Tradition 9:420-39.
_____. 1998. Spacing, Placing and Effacing: Scribal Textuality and Exeter Puddle 30a/b. In New Approaches to Editing Old English Verse. Ed. Sarah Laratt Reefer and Katherine O'Brien O'Keeffe. Cambridge: Brewer. Pp. 45-65.
Fell, Christine E. 1996. The First Publication of Old Norse Literature in England and its Relation to its Sources. In The Waking of Angantyr: The Scandinavian Past in European Culture. Ed. Else Roesdahl and Preben Meulengracht S0rensen. Actajutlandica 31:1. Aarhus: University Press. Pp. 27-57.
Foley, John Miles. 1995. The Singer of Tales in Performance. Bloomington and Indianapolis: Indiana University Press.
Fridrek Magnusson and Gudrun Kvaran, eds. 1994. Hraeringur ur ritum Grunnavikur-Jons. Erindi flutt a malpingi un Jon Olafsson ur Grunnavik laugardagin 16. april 1994. Reykjavik: Ordmennt og Godvinir Grunnavikur-Jons.
Gade, Kari Ellen. 1994. On the Recitation of Old Norse Skaldic Poetry. In Studien zum Altgermanischen. Festschrift fur Heinrich Beck, ed.. Heiko Uecker. Erganzungsbande zum Reallexikon der Germanischen Altertumskunde 11. Berlin & NY: de Gruyter. Pp. 126-51.
Gottert, Karl-Heinz. 1998. Geschichte der Stimme. Munich: Fink.
Gudrun Asa Grimsdottir. 2001. Laerdur Islendingur a Turni. Af Joni Olafssyni Grunnvikingi. Gripla 12: 125-47.
Guorun Ingolfsdottir and Svavar Sigmundsson, eds. 1999. "Vitjun sina vakta her". Safn greina eftir Jon Olasson ur Grunnavik. Reykjavik: Godvinir Grunnavikur-Jons og Haskolautgafan.
Harris, Joseph. 2000. The Performance of Old Norse Eddie Poetry: A Retrospective. In The Oral Epic: Performance and Music, ed. Karl Reichl. Intercultural Music Studies, 12. Berlin: VWB. Pp. 225-32.
Hofmann, Dietrich. 1963. Die Frage des musikalischen Vertrags der altgermanischen Stabreimdichtung in philologischer Sicht. Zeitschrift fur deutsches Altertum und deutsche Literatur 92:83-121.
Hofmann, Dietrich and Ewald Jammers. 1965. Zur Frage des Vertrags der altgermanischen Stabreimdichtung. Zeitschrift fur deutsches Altertum und deutsche Literatur 94:185-95.
Hymes, Dell. 1965/1981. Some North Pacific Coast Poems: A Problem in Anthropological Philology. In Hymes 1981:35-64.
_____. 1977/1981. Discovering Oral Performance and Measured Verse in American Indian Narrative. In Hymes 1981:309-41.
_____. 1981. "In vain I tried to tell you": Essays in Native American Ethnopoetics. Philadelphia: University of Pennsylvania Press.
_____. 1994. Ethnopoetics, Oral-Formulaic Theory, and Editing Texts. Oral Tradition 9:330-70.
Jammers, Ewald. 1964. Der Vortrag des altgermanischen Stabreimverses in musikwissenschaftlicher Sicht. Zeitschrift fur deutsches Altertum und deutsche Literatur 93:1-13.
Jon Helgason. 1925/1926. Jan Olafssonfra Grunnavik. Dissertation, University of Iceland. Copenhagen: Muller (1925). Rpt. in Safn fraedafjelagsins um Island og Islendinga, 5. Copenhagen: Muller (1926).
_____. 1939. Olafsson, Jon [sic]; Olafsson, Jon. In Dansk biografisk leksikon, ed. Povl Engelstoft and Svend Dahl. Vol. 17. Copenhagen: Schultz. Pp. 367-70. Rpt. with slight additions in 3rd ed., ed. Sv. Cedergreen Bech (1979-84), vol. 10.
_____. 1972. Eddassang. Gardar: Arsbok for Samfundet Sverige-Island i Lund-Malmo 3:15-49.
[Jon Olafsson] John Olafsen. 1786. Om Nordens garnie digtekonst: dens grundregier, versarter, sprog og foredragsmaade. Et Priisskrift. Copenhagen: Det kongelige Videnskabers Selskab/Stein.
de Laborde, Jean Benjamin. 1780. Essai sur la musique ancienne et moderne. 4 vols. Paris: P.-D. Pierres.
Mallet, Paul-Henri. 1756. Monumens de la mythologie et de la poesie des Celtes, et particulierement des anciens Scandinaves: pour servir de supplement et de preuves a l'introduction a l'histoire de Dannemarc. Copenhagen.
Nyerup, Rasmus. 1798. Udsigt over Nordens aeldste Poesi og dens Litteratur. Copenhagen: Sejdelin.
O'Keeffe, Katherine O'Brien. 1990. Visible Song: Transitional Literacy in Old English Verse. Cambridge: Cambridge University Press.
Olafur Davidsson. 1888-92. Islenzkar skemtanir. Copenhagen: Hid islenzka bokmentafelagi/Mollers.
Pall Eggert Olason. 1950. Islenzkar aeviskrarfra landnamstimum til arsloka 1940, vol. 3. Reykjavik: Hid islenzka bokmenntafelags.
Percy, Thomas. 1770. Northern Antiquities; or, A Description of the Manners, Customs, Religion and Laws of the Ancient Danes and other Northern Nations Including those of Our own Saxon Ancestors. With a translation of the Edda, or System of Runic mythology, and other pieces from the Ancient Islandic Tongue. 2 vols. London: Carnan.
Stefan Einarsson. 1957. A History of Icelandic Literature. Baltimore: Johns Hopkins University Press and American-Scandinavian Foundation.
Svanhildur Oskarsdottir. 1996. Vardhaldsenglar Eddu. Eddufraedid skaldskap og bokmenntaumraedu a upply singarold. In Gudamjodur og arnarleir: Safn ritgerda urn eddulist, ed. Sverrir Tomasson. Reykjavik: Haskolautgafan. Pp. 221-59.
Tedlock, Dennis. 1971/1983. On the Translation of Style in Oral Narrative. In Tedlock 1983b:31-61.
4_____. 1982/1983. Reading the Popol Vuh over the shoulder of a diviner and finding out what's so funny. In Tedlock 1983b:312-20.
_____. 1983a. Translating Ancient Words: From Paleography to the Tape-Recorder. In Tedlock 1983b:124-55.
_____. 1983b. The Spoken Word and the Work of Interpretation. Philadelphia: University of Pennsylvania Press.
_____. 1987. Hearing a Voice in an Ancient Text: Quiche Maya Poetics in Performance. In Native American Discourse: Poetics and Rhetoric, ed. Joel Sherzer and Anthony C. Woodbury. Cambridge Studies in Oral and Literate Culture. Cambridge University Press. Pp. 140-75.
Dorkell Johannesson, ed. 1947. Merkir Islendingar: AEvisogur og minningargreinar, vol. 2. Reykjavik: Bokfellsutgafan.
_____. 1957. Merkir Islendingar: AEvisfigur og minningargreinar, vol. 6. Reykjavik: Bokfellsutgafan.
Dorunn Sigurdardottir, ed. 1996. Jon Olafsson Ur Grunnavik, Hagpenkir JS 83 fol. Reykjavik: Godvinir Grunnavikur-Jons og Hagpenkir, felag hofunda fraedirita og kennslugagna.
Worm, Ole. 1636. Runer, seu Danica literatura antiqvissima, vulgo Gothica dicta, luci reddita, opera O. Wormii, acc. de prisca Danorum poesi dissertatio. Amsterdam: Janson.