^{page 1 note 1} ‘Import some earth’ as distinct from throwing some dirt! Nobody can properly withhold admiration from the far-sighted method and meticulous care of Parry and Lord in recording the material and its back ground, nor from its editing by Lord in Serbocroatian Heroic Songs, Harvard U.P. (in progress). The whole work will remain as a model for such studies and an invaluable source for all students of oral literature.

^{page 1 note 2} It is no doubt true to say that epithets ‘pad out’ the line in Latin epic. The change from the shorter Saturnian to dactylic hexameter initiated by Ennius led to the rapid growth of epithets. Many of these are translations of Homeric epithets and the further development of epithet in Latin epic is not relevant to oral epic. The distinction between ‘written’ and ‘oral’ epic lies in the pressure which instant composition produces. This pressure causes a resort to that economy which the tradition prescribes. In instant composition large-scale originality is neither possible nor, in its own milieu, acceptable. The role of the audience as conservers of fixity and controllers of innovation has not been sufficiently stressed. The term aural is as valid as oral for non-written literature.

^{page 2 note 1} See below, pp. 16 ff.

^{page 2 note 2} The tales used here are quoted from the collection Silva Gadelica, edited with translation by Standish O'Grady (Williams and Norgate, 1892). These tales are romantic rather than heroic in concept but as far as storytelling is concerned the distinction is of little importance (see Lord, , S.T., pp. 219–20Google Scholar). Of these tales one, The narrow-striped Kern (referred to as K) was composed, according to O'Grady's suggestion (p. xiii), soon after 1537 or earlier. An Gilla Decair (referred to as GD), The Difficult Lad, and Bodach in Chota Lachina (referred to as BCL), The Churl with the Grey Coat, may also be placed in the sixteenth century. Generally speaking, this might be regarded as the last century in which the composing of ‘high class’ or court tales (hereafter referred to as literary tales because they survive in written manuscripts) flourished. We include also two short tales from the same collection, Bruiden Cheise Chorainn (referred to as BCC), The Magic Dwelling of Chéis Chorainn, and Bruiden bheg na hAlmaine (referred to as BA), The Little Fight at Allen. We extend this sixteenth-century core backwards by referring to The Chase of Sid na mBan Finn (referred to as SBF), in Fianaigecht, Royal Irish Academy, Todd Lecture Series, vol. xvi, ed. Kuno Meyer, 1910 (repr. 1937). This version of the tale, as Meyer points out (p. xxxi), is taken from a manuscript written in 1419; we extend forwards by referring to Eachtra Lomnochtain (referred to as EL), The Adventure of Lomnochtan, a late Finn Tale, probably about 1700 (ed. Bergin and McNeill, Connradh na Gaedhilge, Dublin, no date; for English version see Gaelic Journal, nos. 94–105). Finally, to help in demonstrating the formulaic character of phrases quoted, we shall refer to Agallamh na Senórach (referred to as Ag.), The Colloquy of the Ancients, the longest of the pieces in Silva Gadelica (140 pp.). This is not itself a tale but a compendium of Irish place-lore in which Finn anecdotes are told to St. Patrick by two survivors of the Fiana. The piece dates from about 1200 and the many formulas in the language show that in dealing with heroic narrative matter a traditional language was already well developed. Our eight pieces, therefore, abbreviated in the text of this article as Ag., SBF, K, BCC, BCL, BA, GD, EL, cover the period from 1200 to about 1700.

^{page 3 note 1} Sean-Sgéalta Ó Thír Chonaill, Béaloideas, vol. vii, No. 2, 1937,Google Scholar collected and written down in Donegal by Séan Ó hEochaidh. (Béaloideas is the Journal of the Folklore of Ireland Society, publ. by The Educational Company of Ireland, Dublin.)

^{page 3 note 2} Parry's definition of formula specifies metrical conditions, and Lord (S.T., p. 31Google Scholar) says ‘any study of formula must … properly begin with a consideration of metrics and music’. These are no doubt proper definitions since the writers are dealing with two poetic traditions; but metre has no essential connection with formulas, which I take to be identical sequences of words.

^{page 3 note 3} Cf. Delargy, J. H., The Gaelic Storyteller;Google ScholarSirRhys, John Memorial Lecture, British Academy, 1945, pp. 34–5.Google Scholar

^{page 3 note 4} I deliberately omit consideration of more than one tale from the same speaker. For example, No. X also contains this opening formula, but it is told by the storyteller of No. I. It seems better to avoid such duplica tion so as to show clearly that these formulas are not personal to one speaker.

^{page 4 note 1} See Silva Gadelica, An Gilla Decair, pp. 257–9.Google Scholar These are not formulas in the rigid Homeric sense. There are many passages in Irish tales which are better classified as fixed descriptions. These may contain short formulas and many identical ‘concepts’, but the wording, the number, and the ordering of such concepts, varies, and is deliberately varied.

^{page 4 note 2} Od. 8. 499.Google Scholar

^{page 4 note 3} ‘… gléasadh cuirm a mhair naoi n-oidhche agus naoi lá, agus b'fheàrr an oidhche dheireannach nó an chéad oidhche.’ Béal, vol. vii, 2, p. 205.Google Scholar We exclude two other uses of this formula in the collection because the same storyteller or his father is involved; but it may be found also in other collections, e.g. Béal, vol. vi, 2, p. 292.Google Scholar

^{page 4 note 4} Op. cit., p. 224.Google Scholar Compare the tone of these endings with a typical ending quoted by Lord from a Yugoslav singer: ‘The next day Meho and Anica were married. May they have many children. The following day there was a horse race. Finally the wedding guests dispersed’ (Lord, , S.T., p. 234, App. I). Lord also gives an alternative ending from another singer of the same tale: ‘I heard this song from the Turk Huso Ćoravi. I have not heard from that day to this such a singer. There he is and there is his song. If it is worthy then I too am pleased.’ This sentiment is paralleled by a common Irish formula (not found in the collection we are dealing with) to the effect that ‘that is my story and if there be a lie in it, let it be. It was not I who composed or invented it [the tale].’ This well expresses the storyteller's peculiar position of both claiming and disclaiming authorship.Google Scholar

^{page 4 note 5} Op. cit., p. 232.Google Scholar

^{page 4 note 6} e.g., ‘that that is the adventure of GD’, or ‘of BCL’, or ‘that that is the “Bruiden” of Ch6is Chorainn’, or ‘of Allen up to that point’. See Silva Gadelica, pp. 275, 289, 310, 342.Google Scholar

^{page 4 note 7} The last line of the Iliad—‘Thus they held the funeral of Hector, tamer of horses’—does not seem like an end formula for the whole poem, effective though it is from a literary point of view. One scholiastic tradition provides a continuation line. Continuation (in an oral milieu) may account for the lack of a formal end. For example, the Finn tale Eachtra Lomnocht´in, as printed by McNeill and Bergin, continues for 147 paragraphs, but the majority of manuscripts end the tale after paragraph 50 (and we do not follow the text beyond this point as a source of reference).

^{page 5 note 1} Op. cit., p. 198.Google Scholar This formula also occurs in tale No. IX in the following form: ‘They raised their sails bulging and billowing to the top of the straight masts; and to make a long story short,’—a formula of curtail ment which strikes at the very heart of the heroic storyteller!—‘they made neither sea stop nor long pause until they reach the court and castle of the King of the Underlands.’

^{page 5 note 2} Op. cit.: ‘d'àirdigh sé a chuid seólta mar rinn sé aroimhe; agus ní theàrn sé stad mara nó mór-chomhnuidhe gur shroichidh sé cúirt agus caisleán Ri Innse Tuile ‘san Domhan Thiar.’

^{page 5 note 3} It may be convenient here to summarize other short formulas occurring in the five Donegal folktales selected. The expression ‘ní thearn sé (sí, etc.) stad mara nó mór chomhraidhe’-‘he (she, etc.) made neither sea stop nor great pause’-occurs 8 times in all; in one case the story says ‘she walked on and made neither sea stop’, etc. showing a type of contradiction alleged in respect of certain Homeric formulas. ‘Nior bhain siad méar dá sróin’, a formula of similar meaning occurs 6 times. The phrase ‘cúirt 7 caisleán’, court and castle, occurs 6 times. The traditional formula ‘bhi go maith 7 ni raibh go holc’—things were well and were not ill—occurs 5 times; the expression ‘neoin bheag 7 deireadh and lae’—little evening and the end of day, 6 times; the doublet ‘sughach sáthach’—merry replete—whether as adverb or epithet, 5 times. Other alliterating doublets are ‘ithte ólta’, in the expression his fill eaten (and) drunk, twice, and ‘fá dheireadh agus fá dheidhneach’, in the end and at last, twice.

^{page 5 note 4} Od. 2. It has been pointed out that Rieu's translation is not the best for the purpose of illustrating epic technique. It does not do complete justice to the formulaic character of the poems but on the other hand its vividness and charm represent an aspect of Homer very lacking in more literal transla tions.

^{page 5 note 5} Od. 2. 422–9, Cf. 4. 577–680.Google Scholar

^{page 6 note 1} See Singer of Tales, p. 92.Google Scholar

^{page 6 note 2} As in Il. 16. 130–9. Such descriptions are more summarily dealt with elsewhere, for example in Il. 10. 147–8; 177–8; with 254 ff. and 333–5. A shorter description is Il. 7. 206–7.

^{page 6 note 3} Od. 9. 471–2; 77–8.Google Scholar

^{page 6 note 4} Od. 1. 144–54. The preceding lines also describing the preparations are conventional.Google Scholar

^{page 6 note 5} Silva Gad., p. 268.Google Scholar

^{page 7 note 1} Ag., p. 201Google Scholar (ina chnedhaib ocus ina chréchtaib); Ag., pp. 136, 202Google Scholar (slemain sláinchréchtach); BA, pp. 336, 341Google Scholar (do réir a uaisle ocus a atharda); EL, p. 9Google Scholar (biada saora sochaithme); Ag., pp. 97, 104Google Scholar (urghair-diugad menman ocus aicenta); BCL, p. 293Google Scholar (re suan ocus re sírchotlad); Ag., pp. 189, 206Google Scholar (circhaill teintide); GD, p. 266 (na talman tromfódaige).Google Scholar

^{page 7 note 2} This line, a formula of transition, is properly used in the Odyssey passage to pass on to other matters. In the Iliad passage it is rather illogically incorporated in the feast description. There is no immediate transition, as the narration goes on to say ‘when the sun sank’, etc., and later, ‘when rosy-fingered dawn appeared’, etc.

^{page 8 note 1} In a similar description of a sacrifice in Il. 2. 410 ff. Homer omits the detail of pouring on the wine, possibly because of the first part of the line——as there was no old man there concerned. He simply says they burned (the sacrifice) on logs which had been stripped of their leaves.

^{page 8 note 2} This passage contains noun-epithet formulas, e.g. long luchtmar lánáibsech, and other formula types, e.g. cuan ocus calad-phort, but taken as a whole it is not a formula in the Homeric sense. It could be described as ‘a conceptual formula’. It will be noticed that the concept of sounds heard —seals, monsters, etc.—is the same as in the formula quoted from the Donegal tales (p. 5). Silva Gad., pp. 273–4.Google Scholar

^{page 9 note 1} There are very few cases of repeated similes in Homer. The most notable in the Iliad is the stallion simile of bk. 6 (Paris) and bk. 15 (Hector). The Odyssey out of its approximately forty-six long similes repeats two.

^{page 9 note 2} Even where the subject of comparison is the same, the simile may be quite different. The lion subject is the most obvious case in point. To compare a hero to a lion can suggest only a limited field of likeness in, say, the qualities of courage, strength, fierceness. But illogically the simile takes its own unpredictable path by making the lion's characteristics depend not so much on its nature as on the circumstances related. For example, in Il. 17. 109–12, the lion feels fear, and in 133–6 the helpless presence of its cubs inspires it to courage; and in Il. 18. 317–22, the lion's fierce longing for revenge is due to the fact that a huntsman has stolen its cubs in its absence. In such similes the lion's native qualities are largely irrelevant.

^{page 9 note 3} Silva Gad., p. 289.Google Scholar

^{page 9 note 4} See Ross, J., Modern Philology, lvii (1959-1960). pp. 4f.Google Scholar

^{page 10 note 1} Seanscéalta, Nos. I, VI, VII.

^{page 10 note 2} EL, p. 8,Google Scholar par. 11. The full phrase occurring there—‘They both fell’—i dtaisibh 7 i dtáim-néalaibh báis 7 buan-oidhidh-‘into swoons and deep comas of death and lasting doom’, doubly illustrates the point, for taise 7 tdim-néal is itself a formula [also in Ag., p. 136].Google Scholar

^{page 12 note 1} The development of cohering pairs of epithets enables them to be used as adverbial formulas (the adverb in Irish is formed by prefixing ‘go’ to the adjective). Accordingly we get go láidir lánchalma, with strength and valour (BA), or go calma curata, bravely and in warrior fashion (GD, whereas in SBF, BCC, calma curata appear as epithets).

^{page 13 note 1} Though EL is one of our sources, the above example is taken from an extension to the tale found in a few manuscripts. Else where in this article we use only the shorter tale found in most manuscripts which ends at paragraph 50 of the Bergin-McNeill text. See Cat. of Ir. MSS. in Brit. Mus., vol. ii, p. 345–6.Google Scholar

^{page 13 note 2} If we consider epithets which alliterate with slabrad, we find, Ag., p. 128,Google Scholar slabrada sniomacha senaircit, twisted chains of old silver; SBF, p. 86,Google Scholar co slabraduib snithi sesmacha senairgit, with twisted steadfast chains of old silver; and, BA, p. 337, slabrad sinte senaircit, a stretched chain of old silver.Google Scholar

^{page 13 note 3} Passages of verse often occur in the course of Irish prose tales but consideration of them lies outside the scope of this article. One may remark, however, that Bowra's suggestion (Heroic Poetry, p. 15), that these verse interludes are survivals of an earlier tradition of which the prose has filled the gaps, is one which few Irish scholars would entertain.Google Scholar

^{page 13 note 4} It is not possible on the basis of our examination of Irish formulas to give an accurate idea of the degree of thrift in Irish tales as compared with Homer. As Lord points out, the work of a single singer—and the material would need to be very extensive—is the proper basis for a comparison with Homer. In our case, considering we are dealing with what amounts to a random selection of manuscript tales ranging over many centuries, the abundance of persisting formulas must postulate a high degree of ‘thrift’ in the tradition.

^{page 14 Note 1} Metrically, they cover the same ground but the expression is ruled out in 1. 414 as the initial would lengthen the preceding syllable.

^{page 14 note 2} Possibly the Homeridae, whose part in the transmission of the material is not now knowable, may have—if editing was done—eliminated rival variants.

^{page 15 note 1} Page comments at some length on the epithet (of Achaeans) and takes it to mean ‘black-eyed’ rather than deriving from any association with This epithet is not very common and occurs for the first time at Il. 1. 389. The fact that this obscure and indeed obsolete epithet should occur at all is puzzling, especially since already at Il. 1. 123 (or 135), the Achaeans have been characterized as , a metrical equivalent which thus makes an offence against the ‘law of economy’. The poet has already referred to Chryseis at Il. 1. 98 as , a unique feminine counterpart of The context of this line is the return of the girl to her father, the priest of Apollo, and the context of line 389 where the epithet is transferred to the Achaeans, is the same. The same scene in the bard's mind, when he has introduced an obsolete epithet on the first occasion, may explain his using the same epithet for the Achaeans in the second context. However, we are now encroaching on the fascinating dimension of theme composition, of motif links, motivation, and association techniques, which Lord has developed over and above the formulaic groundwork.

^{page 16 note 1} Epithet-noun, in that order. Otherwise than in this paragraph I have used -he more usual term noun-epithet, but the normal order is epithet-noun.

^{page 16 note 2} Les Origines indo-européennes des mètres grecs, Paris, 1923.Google Scholar

^{page 16 note 3} Od. 4. 13;Google ScholarIl. 23. 493; Il. 9. 5, etc. Meillet writes (p. 70): ‘Il est done probable que à la date où se sont fixees les formules, la liberté était plus grande qu'à la date où l'lliade et l'Odyssée ont été composées.

‘Derrière les usages stricts de l'époque historique, on entrevit une époque préhistorique où il était tenu un compte exact de la quantité, mais oú le versificateur avait des libertés qui ont été restreintes par la suite.’

^{page 16 note 4} ‘Indo-European Metrics and Archaic Irish Verse’, Celtica, vi. 194–249, Dublin Institute for Advanced Studies, 1963.Google Scholar

^{page 17 note 1} Two of the examples of paroemiacs (op. cit., p. 200Google Scholar) quoted by Watkins, an alternation of long and short line, form together a perfect dactylic hexameter:

Watkins, following Bergk, takes this couplet as paroemiac but others regard it as a normal dactylic hexameter. It is mentioned here without prejudice to emphasize how much—even to the extent of identity—the paroemiacs have in common with the epic hexameter.

^{page 17 note 2} See p. 18, and Kirk, G. S., loc. cit. One may perhaps assume the priority in time of prose tales to metrical tales.Google Scholar

^{page 17 note 3} Homer and other Epic Poetry: Wace, and Stubbings, , A Companion to Homer, p. 200.Google Scholar

^{page 18 note 1} In, say, the line of Archilochus beginning the accidental features of are given significance by the deliberate choice of the context which underlines diese accidents.

^{page 18 note 2} Apart from special effects like the unruly stone of Sisyphus, the only clear example of deliberate word pattern that comes to mind is Il. 3. 298–301:

This, with its various repetitions, is a prayer formula, no doubt traditional and reinforced by magical jingles.

^{page 18 note 3} Ross, James in an interesting article in Modern Philology, lvii (1959-1960), pp. 1–12,CrossRefGoogle Scholar examines song-poetry of the Western Isles of Scotland in respect of formulaic diction. He shows that this material exhibits ‘conceptual thrift’ (that is, the idea or mental picture remains rather fixed) ‘with considerable variety of diction rather than “thrift of diction”’. This leaves such formulas free to develop beyond one fixed and established form. In a short excursus on tales he stresses the part which memory and preparation play (as against the Parry-Lord image of the creative composer). This emphasis on recollection finds support elsewhere. See Delargy, J. H., The Gaelic Storyteller, pp. 11, 12, 14, 25, 27, 34.Google Scholar

^{page 18 note 4} Kirk, G. S., The Songs of Homer, 1962, pp. 108 f.Google Scholar