Academic journal article
By Butler, Francis
Journal of Slavic Linguistics , Vol. 13, No. 2
Abstract: The paper adduces strong evidence that Russian vurdalak ('vampire') entered the language thanks to Puskin, who formed it from models in the work of Prosper Merimee and Lord Byron. It also surveys the distribution of related forms in Slavic and suggests that the Croatian surname Vrdoljak may not be related to any of them. These conclusions have significant consequences for a hypothesis of Johanna Nichols regarding the ultimate Iranian origin of vurdalak and related forms.
In an article published in 1987, Johanna Nichols argues against a widespread scholarly hypothesis that Russian vurdalak ('vampire' or 'werewolf'), volkolak ('werewolf' or 'vampire') and related forms in Slavic and neighboring languages may be traced back to Indo-European via proto-Slavic. The development of this hypothesis may be summarized as follows: Vasmer (1906: 403, 410; 1907: 225-26) suggests that (Old) Bulgarian (distinguished by Vasmer from Old Church Slavic) *vl[??]kolak[??] 'werewolf' (where vl[??]k- is clearly 'wolf'; o is a link vowel, and the meaning of -lak- is unclear) was transmitted to Greek, where numerous attested forms with r instead of l arose through regular sound change of r to l before a consonant (cf. Nichols 1987: 165-66). Bulgarian v[??]rkolak and Church Slavic vurkolak (the latter listed in Miklosich 1862-1865: 79) (1) are back-borrowings from Greek. Preobrazenskij (1910: 91-92) accepts Vasmer's derivation of the Bulgarian and Church Slavic forms, but also attempts to explain the second element by linking forms containing a second element -dlak instead of -lak (e.g., Serbian vukodlak) to Serbian dlaka 'hair', 'fur', and Slovenian dlaka 'fur' (Preobrazenskij reads "dlaka"). (See Miklosich 1862-1865: 162 for Church Slavic dlaka 'skin', 'color', with suggested connection to vl[??]kodlak[??].) Vinogradov (1954: 11-12) remarks vaguely that Russian vurdalak is a variant of "volkolak--volkodlak, vr[??]kolak" which "became fixed in the Russian literary language in the 1820s-1830s". He also asserts that "volkodlak (volkolak) entered the Russian literary language from South Slavic", and he follows Preobrazenskij regarding the etymology of the second element. Vasmer (1964: 338-39) summarizes previous literature without adducing a new etymology. In an editorial addition to Vasmer (1964: 365-66), Trubacev asserts that "[t]he form vurdalak, which appeared in Russian literature in the 1820s to 1830s, apparently owes its origin to Puskin and constitutes a distortion of a form resembling volkolak, vurkolak". Trubacev appears to be the first scholar explicitly to suggest that the Russian form originated with Puskin. Finally, Trubacev, ed. (1978: 63), provides an elaborate discussion of "*dolka?/*d(')laka?" with passing mention of "*v[??]lko-dolk[??]".
Nichols criticizes the above scholarly conclusions on grounds that may be summarized thus. (1) The explanation of the first element in Russian vurdalak is ad hoc. (2) No explanation is given for Serbian and Croatian *vurdoljak, which Nichols reconstructs from the American surname Vrdolyak. (3) The "explanation of the appearance of the r in Church Slavic and Bulgarian forms as due to back-borrowing from Greek is convoluted and does not account for [Greek] forms with r not adjacent to a consonant". (4) No explanation is given for English warlock. (1987: 167) More generally (5), these conclusions rest on "an implicit assumption that the preferred type of explanation in Slavic etymological studies is one which traces a word back to Proto-Indo-European, i.e., treats it as pristine native Slavic. This assumption enhances rigor only as long as it is not used to justify dismissing data" (1987:174-75).
Table 1 reproduces a summary of compound types from Nichols (1987: 170), adding a Type C (with first elements similar to Type A and second elements similar to Type B) which Nichols characterizes but does not include in her own diagram.
Nichols argues that the Type A forms, which are found in East and South Slavic, are primary rather than secondary to Type B forms; that none of the elements in Type A are originally Slavic; and that these elements derive from a Scytho-Sarmatian language or languages. …