How Celtic are the Fairies?

Alaric Hall, University of Leeds

This is the original, English-language version of an article published as 'Hoe Keltisch zijn elfen eigenlijk?', trans. by Dennis Groenewegen, Kelten, 37 (February 2008), 2-5.


Of the many people with a stake in the idea of Celtic culturefolk-singers, scholars, neo-pagans, nationalists, or the Irish tourist boardall seem to agree that Celts come hand in hand with fairies. This is especially true if we understand fairy in the general sense of any human-like supernatural being: as supernatural beings go, the Leprechaun is probably rivalled only by Chinas dragons in its success in coming to symbolise a whole nation. But even if we define fairies more strictly as human-sized, human-like supernatural communities living alongside those of ordinary peopleprototypically the es sde (people of the fairy-mounds) of medieval Irish storiesthen Celts seem still to rule the roost.

Ostensibly, this simply reflects the evidence. In modern times, stories about fairies have been collected in far greater numbers in the Celtic-speaking regions of the British Isles than in England, William Butler Yeats commenting accordingly that the personages of English fairy literature are merely, in most cases, mortals beautifully masquerading. Nobody ever believed in such fairies. They are romantic bubbles from Provence. Nobody ever laid new milk on their doorsteps for them (quoted in Thuente 1981: 78). And even these romantic bubbles from ProvenceFrench fesare generally seen as originally Celtic. Sometimes the assumption is inheritance from (unattested) pre-Roman Celtic culture (e.g. Ginzburg 1992: 109, 212). Sometimes borrowing from Britanny and Walesthe stated place of origin of many of the lais of Marie de France, several of which feature fes as protagonistsis assumed (Ewert 1995). Either way, Celtic origins are to the fore; and certainly within the British Isles before the fourteenth century, most of our stories about fairies come from Ireland (on which see below) or Wales (on which see Wood 1992).

Constructing Celts

I suggest, however, that in the medieval period, belief in fairies was not distinctively Celtic, and that the Anglo-Saxon neighbours of Celtic-speakers had a long tradition of similar beliefs. But before addressing the medieval evidence, it is worth briefly considering our modern folkloric evidence, to suggest the context in which medieval fairy-lore has more recently been discussed. Folklorists have increasingly recognised that our corpus of folk-story, largely established in the nineteenth and early twentieth centuries, was defined as much by what collectors sought as by what was to be found. Thomas Crofton Crokers Fairy Legends and Traditions of the South of Ireland established already in 1825 an expectation that Ireland would produce stories of fairies (Croker 1859: vii). A lack of stories about fairies in England, then, may to some extent mean simply that such stories were not collected.

Meanwhile, the association of Celtic-speakers with fairy-lore suited various ideological interests during the period of National Romanticism in which most folklore collecting took place. Thus Yeats, whose views on English fairies I have quoted above, sought to discover and foster a sense of Irelands distinctive cultural identity, and one of his strategies was to collect and publish stories about fairies. For Yeats, claiming fairies for Ireland was a means of claiming authentic peasant culture for Ireland in contradistinction to England. The idea affected his own writing, informing his use of the medieval Irish story of C Chulainns encounter with the sde, Serglige Con Culainn, in his play The Only Jealousy of Emer, which premieredfittingly enough in the present journal, in Amsterdamin 1922 (1952).

However, the alignment of fairies with the Irish also played into English conceptions of Celtic nations as rural, backward, and in need of the rational governance provided by the English, with their Teutonic and Norman racial and cultural heritage (e.g. Hammond 2006). It may not be entirely co-incidental, then, that the interest in medieval Celtic fairies was not matched by searches for Germanic equivalents. Admittedly, because of the wealth of medieval Icelandic mythological literature, most research on medieval Germanic traditions took Scandinavia as its starting-point, but Scandinavian sources offer little by way of stories about elves (lfar). However, our texts from Anglo-Saxon England include a number of medical texts associating elves with causing ailments. These did draw scholars attention, but not as evidence for anything like the Old Irish es sde. Although most of Yeatss contemporaries did consider the idea of elves as causes of illness a barbaric superstition, it nonetheless seemed to reflect a rational Germanic focus on mundane health problems. Moreover, the elves in question were assumedon no real basis of evidenceto have been small, flying sprites. This made them similar to Christian demons, belief in which was still considered relatively respectable around the early twentieth century (Hall 2007: 69, 96130).

Medieval origins for modern fairies

It is certainly possible to show that there are cultural connections between, for example, Irish and Scottish fairy-lore, and that modern fairy-lore is at least partly rooted in medieval Irish tradition. To give an example which has not been noted previously, the confessions to witchcraft made by one Elspeth Reoch on Orkney in 1616 (Stuart 184152: vol 2 part 1, 18791) show some striking similarities with the literary narrative of Serglige Con Culainn (the wasting-sickness of C Chulainn), surviving in the eleventh- and twelfth-century Irish manuscript Lebor na hUidre (Dillon 1953).

At the age of twelve, Elspeth went from her home in Caithness (Northern Scotland) to stay at her aunts house on an island in Lochaber. As she waited at the lochside for the boat home one day, two men came to her, one wearing black and the other with a green tartain plaid. The latter said to her she wes ane prettie And he wald lerne her to ken and sie ony thing she wald desyrewhich he did, Elspeth coming to be able to know things which she should not be able to. Two years later, Elspeth had just delivered a baby, and was visited by the man in black on three consecutive nights. He prevented her from sleeping, called himself a farie man, and on the third night convinced Elspeth to have sex with him. Hereafter, Elspeth had to be dum for learning to sie and ken ony thing she desyrit.

Meanwhile, in Serglige Con Culainn, C Chulainn is by a lake at the autumn festival of samuin, when two birds land there, linked by a gold chain. They sing, and almost everyone present falls asleep. C Chulainn, having recently captured enough birds to give two to each woman present apart from his wife, ill-advisedly shoots stones and a spear at the birds, but for the first time in his life, his projectiles miss. C Chulainn then falls asleep mysteriously, sitting against a pillar-stone, and sees two women approaching, one wearing a purple or red mantle, the other a green one. The women beat him with horse-whips, laughing, until he is almost dead, and then leave; they are doubtless identical with the two swans which appeared earlier. C Chulainn subsequently awakens, but is mute and too weak to move. A year later, after a visit by Oengus, the son of ed Abrat, the king of the es sde, C Chulainn regains some of his strength. His first action upon arising is to expound a poetic brathar-thecosc (preceptual instruction), after which he returns to the stone. There he meets the woman in green who explains that Fann, the daughter of ed Abrat, has fallen in love with him. The rest of the story concerns Fanns wooing of C Chulainn and the subsequent struggle for C Chulainn between Fann and C Chulainns wife Emer.

Various aspects of C Chulainns serglige are paralleled in early Irish and perhaps Welsh sources (Carey 1999), but here I want to emphasise its similarities to Elspeths account. Elspeths story is a problematic source: Elspeths somewhat confused narrative survives only insofar as it was recorded by the court notary, and she produced it under the duress of interrogation by people with different world-views from her own (most strikingly signalled by the fact that whereas Elspeth spoke of a farie man, her interrogators speak of Illusiounes of the Devell). Moreover, her narrative is presented as personal experience rather than a legendary taleand although I would argue that she was constructing her experiences through existing narrative ideas, this still complicates the comparison. However, the similarities are striking. In both narratives, there is an encounter between a person and two fairies of the opposite sex by a loch (in Serglige Con Culainn initially in the form of swans), with the distinctive detail that each of the fairies is dressed in a different colour, and in each case one wears green. Both Elspeth and C Chulainn are subsequently harassed for sexthough in Serglige Con Culainn this is done by proxy, whereas in Elspeths account it is done in person. Although C Chulainns year of mute disability precedes sex with Fann rather than following it, his serglige is nonetheless reminiscent of the dumbness imposed on Elspeth following sex with the farie man; likewise, although no explicit connection is drawn, C Chulainns sudden and unexpected propounding of wisdom when he arises from his sickness recalls the association of Elspeths illness with learning to sie and ken ony thing she desyrit (cf. Carey 1999: 19598). The similarities between C Chulainns first encounter with the sde in Serglige Con Culainn and Elspeths first encounter with the the fairy men strike me as substantial enough that we can consider traditional and ancient precedents for Elspeths narrative as established.

Anglo-Saxon elves and the es sde

However, Serglige Con Culainn also bears some surprising resemblances to the Anglo-Saxon medical texts mentioned above, which may make us question how different its story of the es sde was from those which Anglo-Saxons might have told about elvesor, to give them their Anglo-Saxon name, lfe. Comparison of the lfe with the es sde is encouraged by the fact that the noun-class to which lf belonged in Old English was reserved for words denoting people and peoplessuch as Seaxe (Saxons), Engle (Angles), Egypte (Egyptians) and lde (people). This already suggests that far from being demonic sprites, elves were, like the es sde, human-like beings living as neighbours to normal people. This is consistent with Anglo-Saxons use of feminine forms of lf to translate Latin words for nymphs (Hall 2007: 6263, 7588). It also sits well with our later Scottish evidence. As I have suggested, there is much in the Scottish witchcraft trials which recalls medieval Irish texts concerning fairies. But it is also worth noting that all our records are in English and the vast majority are from lowland, English-speaking areas. Maxwell-Stuart, in his study Satans Conspiracy: Magic and Witchcraft in Sixteenth-Century Scotland, referred to all fairies with the Scottish Gaelic word sthean (2001, esp. 1017)but we might be often be dealing as much, if not more, with ideas originating on the English-speaking side of Scottish culture.

Shedding our preconceptions that Anglo-Saxon elves were demonic sprites, we can return to some of the medical texts and rethink what beliefs they might presuppose. To give one example, a salve from a late ninth- or tenth-century collection known as Leechbook III, specified to be wi feondes costunga (against the tribulations of the Enemy), ends with the note that eos sealf is god wi lcre feondes costunga & lfsidenne & lenctenadle (this salve is good against each tribulation of the Enemy and lf-siden and Lent-illness/tertian malaria; Hall 2007: 121). The word lf-siden, on the basis of its Old Norse cognate sa (to work magic) and its Welsh cognate hud (magic), implies that elves caused illness by magical means. Meanwhile, the remedy suggests that a salve against the Devil may also work against something elves magic and tertian malaria. This suggests that elves were similar to demons, but different enough to require separate mention, and that the symptoms caused by both might be similar to the fevers cased by malaria. Fever, associated with hallucination, would be similar to the vision inflicted on C Chulainn by the sde-women.

These readings are consistent with another remedy, this time wi lcre yfelre leod-runan & wi lf-sidenne (against each evil magical song and against lf-siden), which again associates lf-siden with magic (Hall 2007: 124). This text also includes a subsection gif mon mare ride (if a mre rides someone). This further associates lf-siden with the mre, a female supernatural being which came upon men in the night, suffocating or raping them. The sexual connotation here recalls the ultimate point of the visit of the sde-women in Serglige Con Culainn. Meanwhile, the mre rides its victim, while the sde-women beat C Chulainn specifically with echflesca (horse-whips), perhaps implying a similar conception.


I have argued here that we can identify impressive cultural continuties from our medieval Irish evidence for the es sde to more recent beliefs in Celtic-speaking regions: there is no doubt that Celtic-speakers have had strong fairy-traditions. However, in the medieval period they may have been no different in this from their English-speaking neighbours. There are clear ideological reasons why it suited scholars in the nineteenth and twentieth centuries to accentuate the differences between fairy-lore in the medieval Celtic-speaking regions and medieval Germanic-speaking regions. But the primary evidence more plausibly suggests similarities. Anglo-Saxon elves seem likely to have been magically powerful people living alongside normal human communities, just like the Irish es sde. Meanwhile, although the causing of illness has not been prominent in how scholars think of the es sde, this emerges as a trait which they share with wider comparative material, and which might bear further scrutiny.

It has not been possible here to explore how far similarities in Anglo-Saxon and Irish fairy-lore may reflect the influence of Celtic-speakers on their Anglo-Saxon neighbours, how far the similarities may reflect some older shared cultural heritage, and how far they are consistent enough with beliefs around the world simply to suggest that we are dealing with general tendencies in human behaviour. I have shown elsewhere that early Anglo-Saxon ideas about elves can be paralleled in early Scandinavian evidence, suggesting that they are older than the Anglo-Saxon migrations to Britain (Hall 2007: 2166). On the other hand, some of the similarities between the texts which I have studied here seem too specific simply to be coincidence, so we should maybe look to shared features of Germanic and Celtic cultures in Antiquity.

None of this is to say that medieval Irish fairy-traditions, or other traditions from Celtic-speaking regions, had no distinctive features. Rather it is to say that those distinctive features will only emerge from new and revisionist readings of the full range of our medieval evidence.

Works cited

microsoft windows 7