These infant Callicantzari are particularly prone, it is said, to attack and kill their own brothers and sisters. Hence comes the by-name by which they are sometimes known, ἀδερφοφᾶδες, ‘brother-eaters,’ as also, according to Polites’ interpretation, the name κάηδες, which is an equivalent for Callicantzari in several islands of the Aegean Sea. This word Polites holds to be the plural of the name Cain, and to denote ‘brother-slayers’; but inasmuch as a longer form καϊμπίλιδες appears side by side with κάηδες in Carpathos[563], I hesitate to accept this interpretation of the one while the other remains to me wholly unintelligible. At any rate to the people themselves the word has ceased to convey any idea of murderous propensities; for in the island of Syme, where the name is in use, the beings denoted by it are held to be harmless[564].

The issue before us is well summarised in two popular traditions which Polites adduces from Oenoë and from Tenos, and which are in clear mutual contradiction. The tradition of Oenoë begins thus: ‘“Leave-us-good-sirs” (Ἀς-ἐμᾶς-καλοί) is the name which we give them (the Callicantzari), though they are really evil demons (ξωτικά).’ The tradition of Tenos opens with the words: ‘The Callicantzari are not demons (ζωτ’κά)[565]; they are men; as New Year’s Day approaches, they are stricken with a fit of madness and leave their houses and wander to and fro.’ How are we to decide which of these two traditions is the older?

The evidence in favour of either is at the present day abundant; the two chief authorities on the subject, Schmidt and Polites, both acknowledge this; and, in my own experience, I should have difficulty in saying which view of the Callicantzari I have the more frequently heard expressed. On the mainland they are most commonly demons; in the islands of the Aegean, more usually human. But in a matter of this kind it would be of no value to count heads; even if the whole population of Greece could be polled on the question, the view of the majority would have no more value than that of the minority. The issue must be decided on other than numerical grounds.

And clearly the first consideration which suggests itself must be the nature of the earliest evidence on the subject. The earliest authority then is Leo Allatius[566], and his statement is in brief as follows. Children born in the octave of Christmas are seized with a kind of madness; they rage to and fro with incredible swiftness; and their nails grow sharp like talons. To any wayfarer whom they meet they put the question ‘Tow or lead?’ If he answer ‘tow,’ he escapes unhurt; if he answer ‘lead,’ they crush him with all their power and leave him half-dead, lacerated by their talons.

Thus far the testimony of Leo Allatius distinctly favours the belief that Callicantzari are human and not demoniacal in origin; but at the same time it must be admitted that his statement was probably founded upon the particular traditions of his native island only and carries therefore less weight. The barbarous custom however which he next proceeds to describe is of some importance. He states that children born during the dangerous period between Christmas and New Year had the soles of their feet scorched until the nails were singed and so they could not become Callicantzari. Now there is a small but obvious inconsistency in this statement. Persons who scratch one another use, presumably, not their toe-nails but their finger-nails; and animals likewise employ the fore feet and not the hind feet. To scorch the feet therefore, and particularly the soles of the feet, is not a logical method of preventing the growth of talons. But on the other hand the treatment adopted might well be supposed to prevent the development of hoofs, such as in many parts of Greece the Callicantzari are still believed to have. In other words, the custom which Leo Allatius describes was not properly understood in his time. But a custom which has ceased to be properly understood and has had an inaccurate interpretation set upon it is necessarily of considerable age. Already therefore in the first half of the seventeenth century the custom which Allatius describes was of some antiquity; and the belief that children turn into Callicantzari, which is implied alike by the original meaning and by the later interpretation of the custom, was equally ancient. In Chios then at any rate the human origin of Callicantzari is a very old article of faith.

But more important for our consideration is the answer to be made to the following question; is it more probable, that Callicantzari, if they were originally demons, should have come in the belief of many people to be men, or that, being originally men, they should have assumed in the belief of many people the rank of demons? Here, if I may trust the analogy of other instances in Greek folklore, my answer is decided. I know of no case in which a demon has lost status and been reduced to human rank; but I can name three several cases in which beings originally human have been elevated to the standing of demons. The human maiden Gello was the prototype of the class of female demons now known as Gelloudes. Striges (στρίγγλαις) are properly old women who by magical means can transform themselves into birds, but they too both in mediaeval and in modern times are frequently confused with demons. ‘Arabs’ (Ἀράπηδες), as the name itself implies, were originally nothing but men of colour, but they now form, as will be shown in the next section, a recognised class of genii. And if we turn from modern Greek folklore to ancient Greek religion, there also we find the tendency in the same direction. There men in plenty are elevated to the rank of hero, demon, or god, but the degradation of a demon to human rank is a thing unknown. In view of this strongly marked principle of Greek superstition or religion, it is impossible to come to any other conclusion than that the Callicantzari were originally not demons but men—men who either voluntarily or under the compulsion of a kind of madness chose or were forced to assume the shape and the character of beasts.

Having thus disposed of the problem presented by the various types of Callicantzari, we must next investigate the origin of the name itself. This investigation too is not a little complicated by the fact that the dialectic varieties of the name are fully as manifold and divergent as the various shapes which the monsters are locally believed to assume. There can be few words in the Greek language which better illustrate the difference in speech between one district and another. The most general form of the word, and one which is either used side by side with other dialectic forms or at least is understood in almost every district, is the form which I have used throughout this chapter καλλικάντζαρος or, to transliterate it, Callicantzaros; but in reviewing all the dialectic varieties of the word, I find that there are only two out of the fourteen letters composing this word, which do not, in one dialect or another, suffer either modification of sound or change of position. The consonant κ in the first syllable and the vowel α in the third are the only constant and uniform elements common to all dialects.

These dialectic forms demand consideration for the reason that some of the derivations proposed take as their starting-point not the common form καλλικάντζαρος but one of the rarer by-forms—a method which is evidently open to objection when it is seen, as the accompanying table of forms will show, that καλλικάντζαρος, besides being the common and normal form, is also the centre from which all the dialectic varieties radiate in different directions. In compiling my list of forms, however, I may abbreviate it by the omission of those which are a matter of calligraphic rather than of phonetic distinction. Thus the first two syllables of καλλικάντζαρος are often written καλι- or καλη-, but since ι and η represent exactly the same sound and λλ is very seldom distinguished from λ, I have uniformly written καλλι- even where my authority for the particular form uses some other spelling. On the other hand, as regards the use of τζ or τσ, between which there is a real if somewhat subtle difference in sound, I have retained the particular form which I have found recorded.

Starting then from the normal form καλ-λι-κάν-τζα-ρος, which I thus dismember for convenience of reference to its five syllables, I may classify the changes which the word undergoes in various dialects as follows: