It is not a matter on which we can properly legislate. Simon Young has shown convincingly how little agreement there is about the meanings of terms for fairies collected by folklorists in nineteenth-century Cornwall, concluding that “there is enormous blurring in lore and very often taxonomic categories misrepresented the beliefs of a given area”; if this is true of a single well-documented English county in a recent century, what hope can there be of our reconstructing a coherent fairy taxonomy for the whole of the European Middle Ages with the far scantier evidence that is available to us? As Young writes, “anyone who studies history has to constantly remind themselves that those people living hundreds of years ago did not structure their experience as we do.”6
Even in the Middle Ages fairy taxonomy seems to have been problematic. Thomas of Cantimpré, for instance, tries to categorize fairies in the final section of his mid-thirteenth-century book of moral instruction, De bonum universale de apibus [On the Universal Good of Bees], but the enterprise quickly falls apart.7 Turning from his admirable bees, he sets out, under the headings of ‘wasps,’ ‘cockroaches,’ ‘hornets,’ and ‘beetles’ [vespae, blattae, crabrones, and buprestes], to describe the depredations of various kinds of demon. Wasp demons, he says, cause tempests, and cockroach demons cause bad dreams, but when he turns to what we might call ‘fairies’ (under the heading of ‘hornets’), we discover that these too can cause tempests and bad dreams. Hornet demons, he says, can be divided into four classes: neptuni, who swim in water; incubi, who roam the earth; dusii, who live under the earth; and spiritualia nequitie in celestibus, who inhabit the air. This is already an eccentric classification since for Saint Augustine (as for most of the medieval commentators who followed him), incubi and dusii were clearly one and the same thing.8 Moreover, none of these terms is likely to have been used at the level of popular speech, nor is his classification likely to have represented any kind of popular taxonomy. Neptuni is evidently a commonization from the Roman god, and it is just possible that some such term was in popular usage, if only as a folk etymology for the French neton.9 On the other hand, there does not seem to have been anything particularly aquatic about netons,10 and whatever wassergeister Thomas of Cantimpré may have had in mind were probably called something quite different in common speech. The word dusius may well be the Latin form of a Gaulish word current in Augustine’s day, but it seems to have died out in European vernaculars by the thirteenth century.11 Incubus, probably the most widely used general scholastic term for ‘fairy’ in the Middle Ages, derives etymologically from the sense of being weighed down (incubitus) or smothered in sleep. The closest equivalent in English for this specialized sense would have been ‘nightmare’ (in French cauchemar and in German nachtmahr), but incubus underwent semantic generalization early—though not, as Thomas of Cantimpré would have it, to ‘earthbound spirit’ but rather (as we shall see in Chapter 3) to ‘fairy lover.’ Finally, when he comes to ‘the wicked spirits of the air’ (evidently Thomas has no specific name for them and must resort to Ephesians 6:12), he starts out by describing demonic tempests (not obviously different from those caused by the vespae) and then falls back on the general category of illusions (which turn out to include blattae-like dreams and incubi-like seductions!). After all this we should not be surprised to find that Thomas seems to have completely forgotten about the beetles (buprestes) with which he started out. In the end he simply gives up and launches into a recital of miscellaneous marvels, some of which he claims to have experienced personally. At least the taxonomy supplied by John Walsh, a Devonshire cunning man, in 1566 has the virtue of simplicity: “[he] saith that ther be .iii. kindes of Feries, white, greene, and blacke … Wherof (he sayth) the blacke Feries be the woorst.”12 Incidentally, the question of fairy coloring is its own mare’s nest. In addition to white, black, and green (green is sometimes mentioned—as with the green children of Woolpit—but it is by no means universal), we also have gray (in The Merry Wives of Windsor), red (in an account from Thomas Walsingham),13 and polychrome (as with Tristram’s fairy dog Petitcriu). Perhaps the key to all this is the innate volatility of fairies: they can be any size (or shape) they wish, and, as in Petitcriu’s case, their color is inherently unstable.
No attempt, whether medieval or modern, to impose a logical order on spontaneous local traditions can ever be totally satisfactory (though those who still feel themselves in need of such answers can always turn for help to Katharine Briggs or Claude Lecouteux).14 My own solution to this problem, however, is functional: for the purposes of this study I am concerned primarily with that class of numinous, social, humanoid creatures who were widely believed to live at the fringes of the human lifeworld and interact intermittently with human beings.15 In this they differed from those solitary creatures who inhabited the wilderness (giants and the like) or the social creatures who lived among humans (the various kinds of house hold spirit). Of the multitude of potential terms for these creatures in both Latin and the vernacular, few, if any, seem to have had a generally agreed and fixed meaning, but they were most commonly referred to as ‘elves’ in English (at least down to the middle of the fifteenth century)16 and in French as fées. The equivalence of these two terms in learned usage is nicely illustrated by a spell “Pur faies” from a thirteenth-century Anglo-Norman medical treatise that begins, “Conjuro vos, elves.”17 From the middle of the fifteenth century they were increasingly referred to as fairies in England,18 but other terms for them (wodwoses, pouks, goblins, or hobs, for example) were sometimes used. Whatever the name, I shall treat as fairies all creatures who behave in the way I have just described.
A second issue that will not detain me in this study is the Celtic origins of fairy lore. As far as I can see just about every region in medieval Europe had its fairy traditions. Fairies are to be found from Iceland to Sicily and from the Pyrenees to the Ruhr, but the notion that Wales, Scotland, and Ireland have a particular claim on them is deeply ingrained in the consciousness of the English-speaking world. Two main factors seem to be responsible for this view. In the first place, stories concerning King Arthur, many of which are filled with fairy lore, have proved to be among the most enduring of all medieval legends; since they clearly arose among Celtic speakers, particularly the Welsh, we tend instinctively to locate the source of all medieval fairy beliefs in Wales. On the other hand, had the legend cycles of Huon of Bordeaux or Godfrey of Bouillon, both of which contain prominent fairy elements, attained a postmedieval reputation comparable to King Arthur’s, we might now have a very different notion of the epicenter of European fairy beliefs. Second, fairy traditions have survived more tenaciously in Celtic-speaking countries, perhaps most notably Ireland, than anywhere else in the English-speaking world; since Irish fairies have a firm place in modern popular culture, we tend to assume that they somehow preempt similar beliefs elsewhere. The fairy beliefs of the Nordic world, particularly Iceland, have proved to be just as long-lived, however, and only their far greater cultural distance from English speakers has made them appear more tangential. No doubt other factors—the relative prominence of Celtic scholars among English folklorists, for instance,19 or the literary prestige of writers such as W. B. Yeats and J. R. R. Tolkien (whose elves are unmistakably Celtic)—have reinforced the impression that fairies are primarily a Celtic phenomenon.20
It is not even necessarily true that texts in Welsh and Irish contain the earliest medieval literary references to fairies, at least not if we go by manuscript date rather than the hypothetical dating of the material itself. At one point in the Latin epic Waltharius, for example, when Walter of Aquitaine is defending himself in a pass in the Vosges against the Burgundian King Gunther and his men, one of his enemies, a man named Ekivrid, taunts him with the implication that he is a fairy: “You seem just like a woodland sprite [saltibus assuetus faunus] to me” (line763).21 This is evidently a clever Latin pun on Walter’s name (woud heer or woudt-her =‘faunus agrestis’), and he “is saying in effect ‘So you’re called Walthere, are you? Well you really (quippe) do seem to be Waltheer, the wood-sprite, by the incorporeal way you’ve been avoiding arrows and lances.”22 There could hardly be anything less ostensibly Celtic than this poem about Germanic heroes, set in the Vosges mountains