Slovenski jezik Slovene Linguistic Studies 13 2021 Znanstvenoraziskovalni center Slovenske akademije znanosti in umetnosti Inštitut za slovenski jezik Frana Ramovša Ljubljana, Slovenija in/and Brigham Young University College of Humanities Provo, Utah, USA Slovenski jezik – Slovene Linguistic Studies 13 (2021): 3–20 Henning Andersen University of California, Los Angeles, USA DOI: https://doi.org/10.3986/sjsls.13.1.01 PIE *peh 2 ur ’fire’. Two Slavic etymologies The article adds two Proto-Slavic derivatives of the Proto-Indo-European word for ‘fire’ to a small number of other, already identified derivatives (§1). Detailed phonological and morphological analysis of the words for ‘bat; moth, butterfly’ establish PS *netopyŕĭ ‘bat’ and PS *netopyŕĭ, *netŭpyŕĭ ‘moth (> butterfly), which support a Pre-Proto- Slavic (PPS) *[nekt-i+pūr]-ja- ‘[night-time fire] one’. Semantic interpretation posits the social and ecological context for the divergent nominations ‘bat’ and ‘nocturnal moth’. The variant word shapes and meanings of these lexemes defined intersecting isoglosses at the time of the Slavic Expansion (§§2–3). Similar analysis of PS *ǫpyŕĭ ‘revenant, monster’ supports the reconstruction PPS *[un-pūr]-ja- ‘[without fire] one’. Its semantic interpretation is based on the Slavic folk belief that the untimely dead were in the power of evil forces and were tools of evil. In pre-Christian times they were denied the pyre, they were ‘without fire’. This belief long survived the introduction of Christianity and its abolition of cremation and obligatory burial (§§4–5). The conclusion (§6) comments on the remarkable archaisms in these ancient lexemes, which were coined thousands of years ago: PPS *nekt- ‘night’ (cf. Hittite), *un- ‘no, without’ (cf. Germanic), and *pūr- ‘fire’ (cf. West Baltic). Key words: Etymologies, Proto-Slavic, Balto-Slavic, Indo-European, Folklore Članek majhno število doslej prepoznanih izpeljank iz praindoevropske besede za ‘ogenj’ dopolnjuje z dvema praslovanskima izpeljankama (§1). Natančna fonetična in morfološka analiza besed za netopirja in veščo, metulja pokaže na psl. *netopyŕĭ ‘netopir’ in psl. *netopyŕĭ, *netŭpyŕĭ ‘vešča (> metulj)’, ki skupaj odražata predpraslovanski (ppsl.) *[nekti+pūr]ja ‘tak, ki je v zvezi z nočnim ognjem’. Semantična interpretacija ugotavlja družbeni in ekološki kontekst za razhajajoča pomena ‘netopir’ in ‘nočna vešča’. Variantne oblike in pomeni teh leksemov kažejo na prepredene izoglose v času slovanske ekspanzije (§§2–3). Podobna analiza psl. *ǫpyŕĭ ‘povratnik iz smrti, pošast’ omogoča rekonstrukcijo ppsl. *[unpūr]ja ‘tak, ki je [brez ognja]’. Semantična interpretacija temelji na slovanskem ljudskem prepričanju, da so bili predčasno umrli v službi zlih sil ter sredstva zla. V predkrščanskih časih jim je bil odrečen sežig na grmadi; ostali so ‘brez ognja’. Prepričanje se je ohranilo tudi po uvedbi krščanstva, ko je sežiganje umrlih nadomestil obvezni pokop (§§4–5). V zaključku (§6) je podan komentar k izjemnim arhaizmom v obeh leksemih, tvorjenih pred več tisoč leti: ppsl. *nekt- ‘noč’ (prim. hetitsko), *un- ‘brez’ (prim. germansko) in *pūr- ‘ogenj’ (prim. zahodnobaltsko). Ključne besede: etimologija, praslovanščina, baltoslovanščina, indoevropščina, folklora 4 Slovenski jezik – Slovene Linguistic Studies 13 (2021) 1. Introduction It is interesting that of the pair of Indo-European words for ’fire’, some languages generalized the active PIE *h 1 ng w -ni- and others, the inactive doublet PIE *peh 2 ur- (an r/n-stem) (Gamkrelidze & Ivanov 1984: 274); e.g., Skt. agní-, Lat. ignis vs. Hitt. pahhur, pahhunis.gen.sg, Arm. hur, hr-oy.gen.sg, hn-oc ’oven’, Gk. pũr, pŭrós, Gmc. *fōr- ~ *fun-, Go. fon, funins.gen.sg, OIcel. fūrr, funi, fȳrr, fȳri, OE fȳr, OHG fiur, fuir, Gm. Feuer) (Kloekhorst 2007, s.v.; Kroonen 2009, s.v.). 1 When we turn to Baltic and Slavic, it is interesting and instructive that we find the active generalized in East Baltic, Li. ugnìs, Lv. uguns, and the inactive in West Baltic, OPr. Panno ’Feuer’, panustaclan ’Feuerstahl, fire- steel’, a neuter u-stem noun built on the oblique stem PIE *ph 2un-. In Slavic, the active PS *ogńĭ alone is productive as a lexeme, in derivation, and in idiom; but the inactive doublet has been preserved as PS *pyr-, PPS *pūr- in a number of derivatives (1). In the following pages I suggest adding to this evidence PS *netopyŕĭ, *netŭpyŕĭ ’bat; moth, butterfly’ (§§2–3) and *ǫpyŕĭ ’revenant, monster’ (§§4–5). (1) Cz. o. dial. pýří ’embers’, pyřina ’ashes’, pýřeti ’smoulder’, Sk., US pyr ’ashes’, P perzyna ’charred ruins, ashes’ (Stanisławski 1995, s.v.), SC píriti ’blow’, upíriti ’blow into, fan (fire)’, pìrivatra ’arsonist; instigator’, pìrjān ’braised meat’ (Skok 1972: 661–662). 2 1 Notational conventions. Italics are used for attested forms and for the standard Proto - -Slavic comparative reconstructions labeled PS. Other reconstructed forms are in normal font. Hyphens (-) mark morpheme boundaries, plus (+) marks the internal word boundary in juxtapositions. Internal reconstruction of Slavic wordforms yields Pre-Proto-Slavic (PPS) forms that suggest a chronological stage before the Common Slavic velar palatalizations and vowel shifts. Intermediate reconstructed forms are labeled CS (Common Slavic). Abbreviations. Arm. (Armenian), Bg. (Bulgarian), Br. (Belarusian), C (consonant), ChS (Church Slavonic), CS (Common Slavic), Cz. (Czech), dial. (dialectal), Fr. (French), Gk. (Greek), Gm. (German), Gmc. (Germanic), Go. (Gothic), Hitt. (Hittite), Lat. (Latin), Li. (Lithuanian), OR (Old Russian, Old East Slavic), LS (Lower Sorbian), Lv. (Latvian), o. (old, obsolete), OCS (Old Church Slavonic), OE (Old English), OHG (Old High German), OIcel. (Old Icelandic), OP (Old Polish), OPr. (Old Prussian), OR (Old Russian, Old East Slavic), P (Polish), Pb. (Polabian), PIE (Pro- to-Indo-European), PPS (Pre-Proto-Slavic), PS (Proto-Slavic), R (Russian), RChS (Russian Church Slavonic), SC (cited as Serbo-Croatian), Sk. (Slovak), Skt. (Sanskrit), Sn. (Slovene), Sp. (Spanish), US (Upper Sorbian), Srb. (Serbian), U (Ukrainian), V (vowel). 2 SC pìrivatra exemplifies a third Indo-European word for ‘fire’: PS *vatra (P watra, Cz. vatra, dial. vátra, Sk. vatra, U vátra, SC vȁtra ‘bonfire, camp fire, hearth'), Rum. vatră, Alb. vatrë, votrë, Av. ātar- ’fire’, Skt. átharvan- ’priest, Gm. Feuerpriester’, Lat. āra ’altar’ (Vasmer 1953, s.v. vatrúška; Skok 1973, s.v. vȁtra; Vaan 2008, s.v. āra ; Rix 2001, s.v. *h2eh1-). H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 5 2. PS *netopyŕĭ ’bat’ The presumed Proto-Slavic word for ’bat’ has been the object of much etymological speculation. A handful of well articulated hypotheses have been advanced; for a summary, see Dingley (2006: 90–91). One of these, first proposed by Miklosich (1886: 214), is promising (Derksen 2008: 250). It resolves the apparent compound into CS *nekt-o- ’night’ and a form of *per- ’fly’ with a computed meaning of ’night flier’. One other proposal has been so widely accepted and cited that it cannot be ignored (§2.3). It interprets the word as a sentence PS *Ne to pyrĭ ’That is no bird’; it is a typical popular etymology based on unanalysed phonetic similarities (promoted by Trubačev 1997). There are a couple of obvious difficulties with the more promising hypothesis and – besides – several loose ends, phonological, morphological, and semantic. The obvious difficulties are the vocalism of the two roots. PPS *nekt- for ’night’ is not attested otherwise in Slavic or Balto-Slavic, or indeed outside Hittite (Kloekhorst 2007, s.v. Neku- zi ); see §4. Its attestation only there makes it an extraordinary Proto-Indo-European archaism in Slavic. The root *per- ’fly’ occurs in OCS pero ’feather’ and per-ǫtŭ.prs.3pl ’fly up’, and in R par-it’ ’hover’. PS *-pyŕĭ- ’flier’ presupposes a lengthened zero grade vowel of this root. A zero grade CS *pŭr- is unexpected; we expect CS *pĭr- (thus Vaillant 1974: 655). The lengthening reflected in R parit’ and PS *-pyŕĭ (PPS *pār-, *-pūr-) is unexplained. There will be more to say about these points below, as well as about the loose ends. 2.1 Phonological expression The standard language forms of PS *netopyŕĭ are in (2). They appear to have been codified with a view to the most conservative attestations. Some standard languages do not use a reflex of this etymon for ’bat’ (e.g., Br. kažán, Srb. šìš-miš, Bg. prílep), but reflexes of it are attested in all Slavic languages in one form or another (Dingley 2006: 97). A selection of dialect forms is offered in (3). (2) P nietoperz, LS ńetopyŕ, US njetopyŕ, Cz. netopýr, Sk. netopier, Sn. netopír, U netopýr, R netopýr’. 6 Slovenski jezik – Slovene Linguistic Studies 13 (2021) Dialects of several Slavic languages have a number of nonstandard variants. Far from sowing doubt about the validity of the accepted reconstruction of the word – as some scholars have feared – many of the variants actually presuppose the generally accepted reconstruction. The multiple variants have been thought to be taboo motivated (Bezlaj 1982, s.v.), but some are motivated semantically (3.a–b, d) or phonetically (3.d–e), and one might well judge the remainder to be results of folk etymology, that is, of efforts to render (parts of) an odd, opaque word with a more familiar segment sequence. In some dialect forms initial ne..., presumably interpreted as a negation, has been omitted; some results of this shortening have later become prefixed (3.a); or the initial n... has been reinterpreted as another sonorant (3.b). In some variants, the initial syllable has been replaced (3.c). Commonly, the intervocalic /t/ between the unaccented V 1 and V 2 has become lax/voiced (3.d). In some instances, syllable onsets have been interchanged, C 1 VC 2 > C 2 VC 1 or C 2 VC 3 > C 3 VC 2 (3.e). Some variants have a changed C 3 or V 3 (3.f). (3) (a) Sn. dial. topír, dopír, dupír, vdopír, podlopír; U dial. topýr (B). 3 (b) Cz. dial. letopéř, letopýř, Sk. dial. letopier, ňetopier, ňetopér, ňetopír; Sn. letopír; SC dial. metòpir. (c) P dial. latopierz (B), mętopyrz (B), Cz. dial. latopyř (B), Sk. litopéř; Sn. dial. latopír, natopír, matapír, matofír; SC metòpir; OR dial. natopyr’. (d) P dial. niedoperz; LS dial. ńedopeŕ, ńedopyŕ; Sk. dial. nedopier, nedopiar; Sn. dial. ladoper (B), nadopir, nadopér; R dial. niedopyr’. (e) SC dial. danopir (B); nepotjer (B); U dial. nepotýr. (f) P dial. -pierz (B); SC dial. -pijer; Sn. dial. -pér; Cz. dial. -péř; Sk. dial. -pér, -piar, -pier; The oldest attestations of the word are RChS netopyrĭ, nepŭtyrĭ, notopyrĭ; OCz. netopýř, OP nietopyrz, nietoperz. Of these, RChS nepŭtyrĭ (1500s) with its weak jer has no known reflexes meaning ’bat’ in the later textual record, nor in modern dialects; but it is significant for the reconstruction; see §3. 3 The examples in (2) are mainly from Trubačev (1997); a few from Bezlaj (1982) are marked (B). H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 7 RChS notopyrĭ is notable. Its noto- can represent an inherited CS *nokto- or a *nekto- modified by vowel assimilation (e > o /...o) either before or after the loss of /k/. Similar possibilities might be considered for the reconstructed PS *neto-: either it is indeed an archaism *nekto-, or it is the result of a vowel dissimilation (CS *nokto- > *nekto-) as in Fr. secourt ’aids’ (< succurrit), quenouille ’distaff’ (< conucula), Sp. hermoso ’beautiful’ (< formosus), redondo ’round’ (< rotundus), reloj ’watch’ (< (ho)rologium), etc. Such dissimilations are ’natural’, but typically irregular or sporadic, being limited by morphological and lexical constraints (Grammont 1931: 272). A single example of dissimilation in the language, in a unique, morphologically opaque stem such as netopyŕĭ, would not be surprising; it is implicit in Vaillant’s (1974: 655) reconstruction ”nokto-”. However, the existence of the byform PS *netŭpyŕĭ (§3), in which the connecting vowel is not -o-, rules out this possibility. 2.2 Morphology The interfix -o- is a minor matter. Already in prehistory, -o- (PPS *-a-) had become generalized for all stem types in Slavic compounds, including i-stem (OCS sŭmrŭt-o-nosĭnŭ ’death-bringing’), u-stem (syn-o- božĭstvie ’son-godliness’), and C-stem nouns (srŭd-o-bolja ’heart-ache’); cf. Vaillant (1974: 747–750). Hence, whichever stem class PPS *nekt- may have belonged to we are not surprised by the interfix -o-. There is one widely attested PPS *nokt- compound that appears to have a PS -ĭ- interfix: R nočleg, P nocleg, Cz. nocleh ’night lodging’ point to PS *noť-ĭ-legŭ. In Vaillant’s judgement, this was not a compound, but a juxtaposition (R sopoloženie, Gm. Zusammenrückung) of temporal adverb PPS *nakt-i.loc.sg and noun, PPS *nakt-i+lega- ’bed+at-night’, cf. OR nočĭ adv. ’at night’ (Olander 2015: 170). More serious is the derivational suffix that is responsible for the stem- final /ŕ/ of PS *netopyŕĭ. This morphological constituent – actually the word’s morphosyntactic head – is PIE *-jo- (PPS *-ja-); after the Common Slavic deiotation it is reflected as a consonant mutation, /r/ → /ŕ/. It is overlooked in most etymological studies of this word, which present it as if it were an i-stem noun, viz. *netopyrĭ (thus Trubačev 1997). There is no basis for this; the descendants of PS *netopyŕĭ reflect a jo-stem (Vaillant 1974: 505). Vaillant (1958: 157; 1974: 655) thought 8 Slovenski jezik – Slovene Linguistic Studies 13 (2021) the /ŕ/ might be analogical; but his point of departure for this analogy (a reconstructed ”*-pir-” from the root *per- ’fly’) is semantically unacceptable. The effect of the PIE *-jo- suffix is to form a relative (possessive) adjective, in this instance evidently substantivized, yielding the literal reference potential: ’one relating to a night-time X’. Once we recognize the suffix and its meaning, it puts the supposed ’night-flier’ in doubt and prompts us to rethink the meaning of the X, PS *-pyŕ-, PPS *-pūr-(ja-). 2.3 Meaning A widely accepted etymology (Brückner 1927, s.v. nietoperz) and the preferred interpretation of Trubačev (1997), reads the Proto-Slavic reconstruction as a sentence *Ne to pyŕĭ ’That is not a bird’. Trubačev praises this construal for being in ”complete accordance with Slavic word-formation and semasiology”. But for the construed sense the element order is wrong; it should be *to ne pyŕĭ. And for this sense the demonstrative pronoun is quite redundant and hard to justify; no clause is needed to express the negative sense; a simple compound would suffice, cf. R neveža ’ignoramus’, nedug ’ailment’, neprijazn’ ’enmity’, nerjaxa ’sloven’, etc. and the ancient Slavic tradition of negated names, e.g., OR Nedanŭ, Nekrasŭ, Neljubŭ, Nemilŭ, Nesulŭ. Trubačev does not mention how the inscrutable PS *-pyr- ’bird’ would conform to ”Slavic word formation and semasiological tradition”. ’That’s not a bird’ is a (scholarly) folk etymology, an attempt to make sense of the surface phonetics without a morphological analysis (see §2.2), without regard to the actual attestation (see §2.1), and at the cost of making up a brand-new word for ’bird’, unknown to any Slavic language, past or present. Although it is popular among academics, the great number of variants of *netopyŕĭ in Slavic dialects that were sampled in (3) shows that this interpretation has not seemed obvious to the folk. If we reconstruct PS *netopyŕĭ as PPS *[nekt-i-pūr]-ja-, we can recognize its second constituent as the Slavic correspondent of PIE *peh 2 ur- ’fire’ (see further §3.3, §5) and posit that the word for ’bats’ was coined with the reference potential ’the ones by the night-time fire’. This seems a H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 9 realistic hypothesis: Bats typically fly out as night falls, are often seen against the darkening sky, but can also be seen, fleetingly, in the light of an out-door fire even though they would not directly have anything to do with the fire; but see further §3.3. 3. PS *netopyŕĭ, *netŭpyŕĭ ’moth, butterfly’ Most discussions of *netopyŕĭ ’bat’ mention that in some Slavic languages or dialects it means ’butterfly’. Mostly not much fuss is made over this fact, but it does involve a few phonological and morphological details that should be mentioned, in addition to the more interesting semantic one. 3.1 Phonology PS *netopyŕĭ is attested in Polabian in various spellings that reduce to Pb. netüpar ’butterfly’ (Olesch 1983: 646); Sk. netopier is dial. ’moth’ (Bezlaj 1982: 221) and Srb. dial. netopir, ’hawk moth Sphingidae’ (cited in Dingley 2006: 86). The byform ChS nepŭtyrĭ, reconstructed as PS *netŭpyŕĭ ’butterfly’ is reflected in Croatian, Bosnian and Serbian (Northeast, Vojvodina, standard) with initial /n/ > /l/ (3.b), interchange of C 2 and C 3 (3.e), and loss of the weak jer: lȅptȋr. In some other Western South Slavic dialects the consonant interchange did not occur, and after the loss of the jer the C 2 C 3 cluster was simplified: lȅpȋr, lȅper (Montenegro). These dissyllabic forms have been exposed to diverse kinds of popular etymology (Skok 1972: 289). 3.2 Morphology In the variant PS *netŭpyŕĭ the interfix -ŭ- stands out. The loss of the /ŭ/ documented by the modern lȅptȋr and lȅpȋr forms shows that -ŭ- in the attested RChS nepŭtyrĭ was not a mere spelling variant for -o-; the -ŭ- represented a real jer in weak position. Several late (Old) Church Slavonic texts document an assimilation of /ĭ/ and /ŭ/ to a following back or front vowel (called ’Umlaut’; Diels 1932: 108–112), undoubtedly a real sound change in dialects spoken by some scribes. However, in PS *netŭpyŕĭ, CS *nekt-u+pūr-ja-, a presumable V 2 PPS /i/ must have been assimilated to V 3 (i > u /...pū) much earlier, prior to the change of 10 Slovenski jezik – Slovene Linguistic Studies 13 (2021) /kt/ before /i/; the vowel assimilation may have been facilitated by the intervening labial plosive /p/; contrast PPS nakt-i+lega-, PS *noťĭlegŭ (§2.2). PS *netŭpyŕĭ, perhaps no longer recognizable as a compound, escaped the generalization of the -o- interfix. In a comparison of the byforms *netopyŕĭ and *netŭpyŕĭ we would perhaps select the latter, which is unaffected by morphological analogy, as our final Proto-Slavic reconstruction. But it seems more interesting to compare both with PS *noťĭlegŭ and reconstruct a PPS *[nekt-i+pūr]-ja- ’night-time+fire-one’. 3.3 Meaning Most etymological studies of these words take the change ’bat’ > ’butterfly’ in stride as an uncomplicated semantic shift, perhaps with ’moth’ as an intermediate stage. But whether the supposed original meaning was ”That’s-not-a-bird” or ”night flier” (cf. §2.1), why would ’bat’ necessarily be the basic meaning? It would seem possible there was no change from ’bat’ to ’moth’ at all. The posited PPS *[nekt-i+pūr]-ja- ’night-time-fire-one’ prompts us to consider more explicit contexts. (i) Moths navigate by available light, such as the northwestern and northeastern horizon during summer nights; as a consequence of their orientation in relation to a light source, they tend to circle artificial lights repeatedly. (ii) Bats have no interest in a bonfire as such, but with their echosignals they track moths. If moths flutter around a fire, bats will pursue them there. Here is the basis for two divergent nominations or two divergent semantic specializations, ’bat’ and ’moth’. The subsequent extension from ’moth’ to ’butterfly’ would be made possible by the existence of day-time moths. In South Slavic dialects the ’moth’ word has been associated with Srb. lepètati ’flutter’; its extension to ’butterfly’ may have been aided by an association with Srb. lȇp ’pretty’ (Skok 1972: 289). 3.4 Time and space The geographical distribution of the ’moth; butterfly’ words (§3.1) shows that both the semantic difference ’bat’ vs. ’moth’ and the phonological difference PS *netopyŕĭ vs. *netŭpyŕĭ may have existed before the historical Slavic Expansion that began in the 500s CE. They H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 11 may have formed intersecting isoglosses prior to the migrations that brought PPS *nekt-i-pūr-ja- into the western Balkan Peninsula and PPS *nekt-a-pūr-ja- northwestward to West Lechitic (Polabian), presumably from Central Slavic. The Western South Slavic and West Lechitic distribution of the ’moth; butterfly’ meaning is somewhat similar to that of the PS dial. *-ny- (vs. *-nǫ-) verb suffix (Andersen 1999, 2020). These are some of the small prehistorical dialect differences that some time in the future may be integrated to give us a fuller understanding of prehistoric Slavic and the historically attested Slavic Expansion. The existence of these differences before the Expansion does not tell us when the ’bat’ and ’moth’ words were coined. Their relation to the much earlier differentiation – when pre-Slavic lects went with East Baltic in generalizing the (formerly active) PPS *agni- while giving a nod to West Baltic by conserving formations with the (formerly inactive) PPS *pūr- gives pause to thought; see §6. 4. PS *ǫpyŕĭ ’revenant, monster’ The similarity of PS *netopyŕĭ and *ǫpyŕĭ has been noted by etymologists in the past (cf. Dingley 2006), but none of the proposed interpretations of *netopyŕĭ has helped clarify *ǫpyŕĭ; it has remained a puzzle (Vasmer 1957, s.v.). It is not surprising that we do not find it in the Etymological Dictionary of the Slavic Inherited Vocabulary (Derksen 2008). Quite apart from a possible connection with *netopyŕĭ, however, a real problem with *ǫpyŕĭ is that the word vampire is derived from some of its modern descendants (4), which since the 1700s has allowed their meaning to be influenced by the Western fascination with vampirism, to such an extent that in most of the modern Slavic languages they simply mean ’vampire’. The obsession with vampirism has made some scholars simply take the notion of vampire as point of departure in the investigation of PS *ǫpyŕĭ (thus Stachowski & Stachowski 2017). The meaning of P upiór is ’ghost, spectre, phantom’, senses that recur in its derivatives (Stanisławski 1995, s.vv.); they appear covered by that of ’revenant’. By contrast, Cz. upír is ’vampire bat; nightly creature that sucks blood’ (Travníček 1952, s.v.). U upýr ’dead person that sucks blood form the living’ (Hrynčenko 1909, s.v.); R upýr’ and vampir are synonyms (Smirnickij 1992, s.vv.). 12 Slovenski jezik – Slovene Linguistic Studies 13 (2021) Since ’revenant’ is a cultural term – unlike the zoological ’bat’ and ’butterfly’ words – we expect to have to peel back layers of semantic change reflecting a history of cultural development in order to reach the original meaning of PS *ǫpyŕĭ. Fortunately, as we will see in §5.3, Slavic folklore – collected, coincidentally, since the 1700s – points to a more concrete meaning of the word and provides enough information about its use in former times to lead us back more than a thousand years, beyond the introduction of Christianity, to a plausible cultural context for its creation. In any case, however, our first step must be a morphological analysis aiming to recover the original, literal reference potential of the word. 4.1 Phonology The standard languages have the following reflexes of PS *ǫpyŕĭ: P upiór, Cz., Sk. upír, Bg. văpír, U upýr, Br. upír, R upýr’. A fair amount of variation is attested in Slavic dialects (4). (4) P wąpierz, wypiór, lupiór, o. upierz; SC o. upir, upirina; Bg. văper, vapir, vapirin, vipir, vepir, voper, jepir, ljapir, lipir(in), lepir(in), vlepir(in), lampir, lempir, upir, o. vǫpyr, vampyr, vepyr; U opyr, vopyr, oper, upir, pir; Br. vupar; R upír (Valencova 2013: 92). It is difficult to see any regular phonological or semantic motivation behind the many dialect variants; they may be results of taboo avoidance or folk etymology, a common result of the fraught transmission of a morphologically opaque lexeme. The Proto-Slavic form is reconstructed (on the basis of the principle of lectio difficilior) from P dial. wąpierz, Bg. văpír, o. vǫpyr, R upýr’; U upýr is compatible with these. A perfect set of correspondences would be P **wąperz, Cz. **upýř, Sk. **upýr, SC o. upir, Bg. văpír, U upýr, Br. **upýr, R upýr’. P upiór, attested since the 1700s may be an acquisition from Russian (Brückner 1927, s.v.). The spellings of Cz., Sk. upír are unetymological, but the word may have been introduced in its pronunciation form or may be acquired from South Slavic. The man’s name Upir attested in Old Russian may be a phonological adaptation of an Old Swedish name Ófeigr (literally ’undaunted’) (Sjöberg 1982). H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 13 4.2 Morphology The stem of PS *ǫpyŕĭ divides into prefix, root, and suffix. Two unproductive, homonymous noun prefixes are attested: (i) PS *ǫ 1 - represents PS *vŭ(n)- (PPS *un-) ’in(to)’, e.g., OCS ǫdolĭ ’valley’, P wądół ’ravine’, R dial. udól ’hollow’; ChS utŭkŭ ’weft, woof’, P wątek, R utók ’idem’. (ii) PS *ǫ 2 - represents PIE *ṇ- (PPS *un-), cf. Gk. a(n)-, Lat. in-, Eng. un-, PS *ne- ’not’ (PPS *ne-) ’negated, lacking’, privative or caritive (Vaillant 1974: 763). Both prefixes reflect a vowel-lowering CS /un/ > /on/ in nominal formations, where they were univerbated; the change did not affect the clitic preposition or preverb PPS un- ’into’, PS *vŭ(n)-. 4 Here we have PS *ǫ 2 -, the privative prefix. It is attested in historical Slavic in a handful of lexemes, in part with metaphorical extensions (5.a); these have parallels in other languages, e.g., Gm. Wetter ’weather’ vs. Unwetter ’stormy weather’, Tiefe ’depth, deep’ vs. Untiefe ’shoal; bottomless pit, abyss’, Zahl ’number’ vs. Unzahl ’huge number’. One South Slavic text known in an East Slavic copy (Svjatoslav Izbornik, 1073) documents extensions with color adjectives (5.b). (5) a. (i) OCS ǫrodŭ adj. ’mad’, R uród ’freak, monster; ugly person’ (lit.: not born > not born right’); (ii) OCS ǫtĭlŭ adj. ’holey, leaky’ (lit.: ’without bottom’ > unsteady, weak), P wątły ’weak, frail, sickly; flimsy’, Cz. útlý ’frail, tender’, Sn. vótel ’hollow’, SC dial. ȕtal ’hollow’, R útlyj, ’fragile, frail’; (iii) ChS ǫsobĭ adv. ’(mutually) opposed’ (lit.: ’without reciprocity’), OCS ǫsobica ’strife, sedition’ (Vaillant 1974:763). b. u-črŭmĭnŭ ’reddish’, u-zelenĭ ’greenish’, jǫ-sinjĭ ’dark blue’ (Vaillant 1974: 764; Boryś 1975: 149–169). The root PS *-pyr- (PPS *-pūr-) ’fire’ is weakly attested in the Slavic languages, as mentioned (§1). Like the prefix, it has lost all productivity, being superseded – as an independent lexeme and in derivation and 4 The vowel-lowering was presumably just word-initial and preconsonantal. If this account is right, either the prefix of PS *onuťa ’foot-cloth’, R onúča (PPS *un 1 -au-t-jā-) reflects a generalization of the o-vocalism, or it can be understood as diphthongized before initial vowel (/ǫ-/ → /on-/); for a parallel, cf. PS *telę ’calf’, ESl. dial. telen-ŭkŭ ’diminutive’, R telënok; see also Majer Forthcoming. The proposed vowel-lowering is similar to the one posited by Kortlandt (1979) to account for PS *ognĭ ’fire’, cf. Li. ugnìs (viz. ugni- > ungni- > ongni- > ogni). 14 Slovenski jezik – Slovene Linguistic Studies 13 (2021) idiom – by descendants of its ancient synonym PS *ogńĭ (PPS *agni-). Some examples are in (1). PS *-pyr- has been suspected to be the root of *ǫpyŕĭ before; the idea is rejected out of hand by Trubačev (1997), but it has been advocated by several scholars in recent years (e.g., Lukinova 1986: 123; Gluhak 1993, s.v. vampir; Bezlaj 2005, s.v. vampir; Snoj 2015, s.v. vampir) but without a persuasive explanation that would connect the word’s literal sense to its meaning in use. The suffix is the relative or possessive PIE *-jo- (PPS *-ja-) (Vaillant 1974: 505); after the Common Slavic deiotation, it appears as a mutation, here /r/ → /ŕ/. We reconstruct a PPS *[un-pūr]-ja- with the literal reference potential ’one without fire’. 4.3. Semantics The literal sense of the etymon must be interpreted in the social or cultural context for which it was coined. The reconstructed meaning ’revenant, monster’ presupposes cultural contexts that surround death. The literal reference potential ’one without fire’ seems to imply a culture in which cremation was the norm, and in which perhaps a meaningful distinction was made between the deceased that were cremated and those who were not. This could be a variant of the wide-spread existence of distinct funeral rites for the normal passing (the good death) and deviations from it (the bad death), e.g., fatal accident, manslaughter, capital punishment, suicide; an elaborate example from Indonesia is in Andersen (2001: 25). We find this opposition expressed in Russian folklore, which distinguished sharply between the generations of honored ancestors, who could be invoked as protectors and defenders of the living, and the ’untimely dead’ (R ’založnye pokojniki’), the living dead, the R upyri, who would harm the living (Rybakov 1994: 142). Their bad death demonstrated that they had come into the power of evil forces; hence they could spread evil, do damage, bring loss, death, separation, discord, the death of the cattle, drought, or fire to their community. Their spirits were presumed to dwell where they met their untimely death or in the moor or the quag. Traditionally the community would deny such a person a normal funeral ceremony even despite their priest’s insistence that it was H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 15 only suicides that were not to be buried in hallowed ground, and his protestation that there was nothing sinful in dying any other untimely death, whether you were killed, went astray in the woods, or fell from a tree. The folk believed that if such a person’s body was buried in the cemetery, it could cause endless general misfortune, and if misfortune did occur, the presumed upyr’ would have to be unburied, pierced with an aspen stake, and cast away from the cemetery. They did not allow the bodies of the drowned or killed to be buried, but dragged them into the field, pierced with stakes”. And if the ”spring winds were cold” and threatened to damage the crops, ”if they could find some recently buried victim of drowning or manslaughter [...] they would dig up the poor body and cast it away somewhere (Maksim Grek, 16 th century). Oh wicked unreason, oh faithlessness! [...] Is this a way to pray to God, to unbury the drowned or accidentally dead? Is this a way to atone for God’s punishment (Bishop Serapion, 1237). (Rybakov 1994: 143; my translation, HA) The Slavic understanding of the difference between good and bad death evidently preexisted the introduction of burial as the sole approved method of disposing of the dead, which was part of their Christianization. Their pre-Christian custom had been cremation. We find a description of this soon-to-be-abolished custom in the Old Russian Tale of Bygone Years, written by a cleric, a representative of the victorious inhumation culture and believer in the resurrection of the flesh. The Radimiči and Vjatiči and Sěver had identical customs. [...] And if someone died, they made a memorial feast for him, and afterwards they made a large pyre, and they laid him on the pyre; they burnt the corpse, and afterwards they gathered the bones and put them in a small vessel and put it on a post by the road, as the Vjatiči do even to this day. These, too, were the customs of the Kriviči and the other heathens, who did not know the law of God but made a law for themselves. 5 (Ostrowski 2003: 72–74) This passage describes the funeral rite for those who had died a good death. The body of one who had died in a bad way would be disposed of ’without fire’. In the Christianization of the Slavs the replacement of cremation with burial was obviously easier than eradicating the Slavs’ age-old 5 I radimiči i vjatiči, i sěverŭ odinŭ obyčai imjaxu. [...] I ašče kto umrjaše, tvorjaxu triznu nadŭ nimŭ, i po semĭ tvorjaxu kradu veliku, i vŭzložaxutĭ i na kradu, mertveca sožĭžaxu, i posemĭ sobravše kosti vložaxu v sudinu malu i postavljaxu na stolpě na putexŭ, eže tvorjatĭ vjatiči i nyně. Si že tvorjaxu obyčaja kriviči i pročii poganii, ne vědušče zakona božija, no tvorjašče sami sobě zakonŭ. 16 Slovenski jezik – Slovene Linguistic Studies 13 (2021) understanding of good and bad death. Rybakov (1994: 144) dates the logic of this distinction to the Late Stone Age. Who knows? In any case, it is relevant here that the Slavs had cremated their dead, presumably, at least since the early Iron Age. In Sedov’s (1994) chronological account of Slavic prehistory, the formation of the Slavic ethnos is signaled in the archaeological record by the introduction of bell graves, a variant of the urn-field culture, around 500 BCE. During the following some 1500 years, cremation continued as a distinctive feature of archeological finds that can be identified (hypothetically) as Slavic, though in periods other forms of funeral are attested as well (Sedov 1990; 1994, passim). 6 Coincidentally, Sedov considered the change in funeral custom around 500 BCE an important element in the cultural divergence of the Slavs from the Western Balts. We can assume that this development occurred at a time when PPS *pūr- was still productive in the language and the word PPS *un-pūr-ja- ’one without fire’ was a transparent, perhaps matter-of-fact term, perhaps a euphemism for a community member who had died an untimely death and become an instrument for evil. 5. Conclusion In §§2–3 we saw how the variant shapes and meanings of PS *netopyŕĭ, *netŭpyŕĭ ’bat; moth, butterfly’ (PPS [*nekt-i+pūr]-ja- ’night-time+fire- one’) implied the existence of dialect differences in Slavic-speaking areas on the eve of the Historical Expansion in the 500s ce: an east – west isogloss *netopyŕĭ || *netŭpyŕĭ, and a north – south isogloss ’bat’ || ’moth, butterfly’. We can imagine a possible socio-cultural context for the creation of these words, the gathering around the out-door fire after dark, beside the sort of earth-lodge well known from Slavic archaeology. Also for PS *ǫpyŕĭ ’revenant, monster’ (PPS [*un-pūr]-ja- ’without-fire-one’) there is an implied socio-cultural context, though a very different, grim one. 6 Two comments on the inherited cremation custom seem called for. First, the Chronicle text refers to the deceased in the masculine gender. But there is some evidence that Slavic women accompanied their deceased husbands on the pyre, textual in medieval sources (Niederle 1956: 206–209; Litavrin et al. 1995: 417, 420) and archaeological (Herrmann 1985: 320). Secondly, the distinction between good and bad deaths and the etymology proposed here imply that the archaelogical evidence of cremation reflects only good deaths. H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 17 The purely linguistic context in which these words were coined is well beyond our reach. It was a pre-Slavic stage – or period – when PPS *pūr- was still a productive lexeme and a covariant of the roughly synonymous PPS *agni-, however their variation may have been conditioned; cf.§1, §5.2. PPS *nakt- ’night’ had an inherited allomorph *nekt-, perhaps no longer occurring in inflection but at least in the form PPS *nekt-i.loc.sg ’at night’, a lexicalized oblique case wordform turned adverb; cf. §1, §2. And the negation PPS *ne had an alternant PPS *un-, productive in nominal formations; cf. §5.2. A regular reflex of PIE *ṇ ’un-’, it has no known correspondents in either East or West Baltic. We will need to examine more data in this perspective before we can posit a place and time for the ancient linguistic system(s) of which these details were part. References Andersen, Henning. 1999. The Western South Slavic contrast Sn. sah-ni-ti || SC sah-nu-ti. Slovenski jezik/Slovene Linguistic Studies 2: 47–62. Andersen, Henning. 2001. Markedness and the theory of linguistic change. In: H. Andersen (ed.), Actualization. Amsterdam: Benjamins. 19–57. Andersen, Henning. 2020. On the formation of the Common Slavic koiné. In: V. Boček, N. Jansens, T. Klír (eds.). New perspectives on the early Slavs and the rise of Slavic: Contact and migration. Heidelberg: Winter. 11–42. Bezlaj, France. 1982. Etimološki slovar slovenskega jezika, vol. 2. K–O. Ljubljana: Mladinska knjiga. Boryś, Wiesław. 1975. Prefiksacja imienna w językach słowiańskich. (Monografie slawistyczne, 31.) Wrocław: Ossolineum. Brückner, Aleksander. 1927. Słownik etymologiczny języka polskiego. Cracow: Krakowska spółka wydawnicza. Reprint, Warsaw: Wiedza Powszechna, 1974. Derksen, Rick. 2008. Etymological Dictionary of the Slavic Inherited Lexicon. Leiden: Brill. Derksen, Rick (ed.) 2010. Etymological Dictionary of the Baltic Inherited Lexicon. Available at: https://dictionaries.brillonline.com/search#dictionary=baltic&id=bl ti694 (24. 3. 2021). Diels, Paul. 1932. Altkirchenslavische Grammatik, Teil 1. Grammatik. Heidelberg: Winter. Dingley, John. 2006. Slavic *netopyrь in a broader context. Journal of Slavic Linguistics and Poetics 44/45: 85–103 (2002–2003). Gamkrelidze, T. V., Ivanov, V. V. 1984. Indoevropejskij jazyk i indoevropejcy. Rekonstrukcija i istoriko-tipologičeskij analiz prajazyka i protokul’tury. Tbilisi: Izdatel’stvo tbilisskogo universiteta. 18 Slovenski jezik – Slovene Linguistic Studies 13 (2021) Gluhak, A. 1993. Hrvatski etimološki rječnik. Zagreb: August Cesarec. Grammont, Maurice. 1933. Traité de phonétique (8 th edition, 1965). Paris: Delagrave. Herrmann, Joachim (ed.). 1985. Die Slawen in Deutschland. Geschichte und Kultur der slawischen Stämme westlich der Oder und Neiße vom 6. bis 12. Jahrhundert. Ein Handbuch. Berlin: Akademie-Verlag. Hrynčenko, Borys [D.] 1907–09. Slovar’ ukrains’koji movy/Slovar’ ukrainskogo jazyka. Reprint, Kiev: Akademija Nauk Ukrajins’koji RSR, 1958–1959. Kloekhorst, Alwin (ed.). 2007. Etymological Dictionary of the Hittite Inherited Lexicon. Leiden: Brill. Available at: https://dictionaries.brillonline.com/search#dictionary- hittite&id=hi0725 (24. 3. 2021), First published online October 2010. Kortlandt, F. 1979. Three problems of Slavic phonology. Zbornik za filologiju i lingvistiku 22/2: 57–63. Kroonen, Guus (ed.). 2009. Etymological Dictionary of Proto-Germanic. Available at: https://dictionaries.brillonline.com/search#dictionary=proto_germanic&id=21064 (24. 3. 2021). First published online October 2010. Litavrin, G. G., S. A. Ivanov, V. K. Ronin (eds.). 1995. Svod drevnejšix pis’mennyx izvestij o slavjanax, vol. 2. (7.–9. st.). Moscow: Institut slavjanovedenija i balkanistiki & Vostočnaja literatura Lukinova, T. B. 1986. Leksika slavjanskogo jazyčestva, Ètimologija 1984. 119–124. Majer, Marek. Forthcoming. The origin of East South Slavic gerunds in -nik. Miklosich, Franz. 1886. Etymologisches Wörterbuch der slavischen Sprachen. Vienna: Braunmüller. Reprint, Amsterdam: Philo Press. Niderle [Niederle], Lubor. Slavjanskie drevnosti. Moscow: Izdatel’stvo inostrannoj literatury. Olander, Thomas. 2015. Proto-Slavic Inflectional Morphology. A comparative handbook. Leiden: Brill. Olesch, Reinhold. 1983–1984. Thesaurus linguae dravaenopolabicae. 1–3. Cologne: Böhlau. Ostrowski, Donald (ed.). 2003. The Povest’ vremennyx let, an interlinear collation and paradosis, 1–3. With David J. Birnbaum, Horace G. Lunt. Cambridge, Mass.: Harvard University Press. Rix, Helmut (ed.). 2001. Lexikon der indogermanischen Verben. Die Wurzeln und ihre Primärstammbildungen. (Second edition.) Wiesbaden: Reichert. Rybakov, B. A. 1994. Jazyčestvo drevnix slavjan. Moscow: Nauka. Sedov, V. V. 1990. Pogrebal’nyj obrjad slavjan v načale srednevekov’ja. In: V. V. Ivanov, L. G. Nevskaja (eds.). Issledovanija v oblasti balto-slavjanskoj duxovnoj kul’tury. Pogrebal’nyj obrjad. Moscow: Nauka. 170–182. Sedov, V. V. 1994. Slavjane v drevnosti. Moscow: Naučno-proizvodstvennoe blagotvoritel’-noe obščestvo. Fond arxeologii. Sjöberg, A. 1982. Pop Upir’ Lichoj and the Swedish rune-carver Ofeigr Upir. Scando- Slavica 28.109–124. Skok, Petar. 1971–1974. Etimologijski rječnik hrvatskoga ili srpskoga jezika. 1–4. Zagreb: Jugoslavenska akademija znanosti i umjetnosti. Smirnickij, A. I. 1992. Russko-anglijskij slovar’/Russian-English Dictionary. (16 th edition.) Moscow: Russkij jazyk. Snoj, Marko. 1997. Slovenski etimološki slovar. Ljubljana: Mladinska knjiga. H. Andersen, PIE *peh 2 ur ’fire’. Two Slavic etymologies 19 Stachowski, Kamil, Olaf Stachowski. 2017. Possibly Oriental elements in Slavonic folklore. Upiór ~ wampir. In Michał Németh, Barbara Podolak, Mateusz Urban (eds.). Essays in the History of Languages and Linguistics. Dedicated to Marek Stachowski on the occasion of his 60 th birthday. Cracow: Księgarnia Akademicka. 643–693. Stanisławski, Jan. 1995. Wielki słownik polsko-angielski. 1–2. Warsaw: Philip Wilson. Trávníček, František. 1952. Slovník jazyka českého. Prague: Slovanské nakladatelství. Trubačev, O. N. 1997. *ne(k)topyrĭ. Ètimologičeskij slovar’ slavjanskix jazykov 24: 143– 145. Moscow: Nauka. Vaan, Michiel de. (ed.). 2008. Etymological Dictionary of Latin and the other Italic Languages. Leiden: Brill. Vaillant, André. 1958. Grammaire comparée des langues slaves. Vol. 2.1. Flexion nominale. Lyon – Paris: IAC. Vaillant, André. 1974. Grammaire comparée des langues slaves. Vol. 4. La formation des noms. Paris: Klincksieck. Valencova, Marina M. 2013. Ètnolingvističeskij kommentarij k ètimologii slov mara i upyr’. In: Bohumil Vykypěl, Vít Boček (eds.), Etymology: An Old Discipline in New Contexts. Prague: Nakladatelství Lidové noviny. Vasmer, Max. 1953–1957. Russisches etymologisches Wörterbuch. 1–3. Heidelberg: Winter. Translated: O. N. Trubačev. 1964–1973. Ètimologičeskij slovar’ russkogo jazyka. Moscow: Progress. Received April 2020, accepted May 2021. Prispelo aprila 2020, sprejeto maja 2021. Prevod naslova, izvlečka, povzetka in ključnih besed v slovenščino: Miha Sušnik – pregledala Metka Furlan. Acknowledgments I am grateful to two anonymous readers for helpful advice and encouragement and to Dr. Marek Majer for several substantive improvements. The usual disclaimers apply. Summary PIE *peh 2 ur ’fire’. Two Slavic etymologies This study investigates two traditional Proto-Slavic etyma. (1.a) PS *netopyŕĭ, *netŭpyŕĭ ‘bat’. The former has widespread descendants in East and West Slavic and Western South Slavic; the latter, attested in Middle Russian, has no known modern reflexes. (1.b) PS *netopyŕĭ, *netŭpyŕĭ ‘moth > butterfly’. The former is attested as Pb. netüpar ‘butterfly’; the latter as SC lȅptīr, lȅpīr, Mn lȅper ‘butterfly’. Comparison with PS *noťĭlegŭ ‘night lodging’ points to a PPS *[nekt-i+pūr]-ja- ‘[night-time fire] one’, a 20 Slovenski jezik – Slovene Linguistic Studies 13 (2021) juxtaposition of PPS *nekt-i (cf. OR nočĭ adv. ‘at night’) and *pūr ‘fire’. Moths and bats naurally occur around an out-door fire at night. (2) PS *ǫpyŕĭ ‘revenant, monster’ is resolved as PPS *[un-pūr]-ja- ‘one without fire’. The socio-cultural context for the word’s creation is the ancient Slavic cremation culture, in which the untimely dead were thought to be tools of evil forces. The remarkable archaisms in these words are commented on in the Conclusion. Pide. *peh 2 ur ‘ogenj’. Dve slovanski etimologiji Razprava obravnava dva tradicionalna praslovanska etimona. (1.a) Psl. *netopyŕĭ, *netŭpyŕĭ ‘netopir’. Prva oblika je široko izpričana v vzhodno- in zahodnoslovanskih ter zahodnih južnoslovanskih jezikih; slednja, ki se kaže le v srednji ruščini, nima sodobnih odrazov. (1.b) Psl. *netopyŕĭ, *netŭpyŕĭ ‘vešča > metulj’. Prva oblika je izpričana v polabskem netüpar ‘metulj’, slednja pa v shr. lȅptīr, lȅpīr, črnogor. lȅper ‘metulj’. Primerjava s psl. *noťĭlegŭ ‘prenočišče’ kaže na ppsl. *[nekt-i+pūr]-ja- ‘tak, ki je v zvezi z nočnim ognjem’, tj. sklop ppsl. *nekt-i (prim. stru. nočĭ prisl. ‘ponoči’) in *pūr ‘ogenj’. Tako vešče kot netopirji se ponoči naravno pojavljajo v bližini ognja. (2) Psl. *ǫpyŕĭ ‘povratnik iz smrti, pošast’ je pojasnjen kot ppsl. *[un-pūr]-ja- ‘tak, ki je brez ognja’. Družbeno-kulturni kontekst nastanka besede je stara slovanska sežigalna kultura, v kateri se predčasno umrli smatrajo kot sredstva zlih sil. V zaključku je podan komentar k izjemnim arhaizmom v teh besedah.