Night-mare: on the origin of a trope in Celtic and Germanic (a response to Stephen Pax Leonard)

Размещено на http://www.allbest.ru/

Night-mare: on the origin of a trope in Celtic and Germanic (a response to Stephen Pax Leonard)

Tatyana A. Mikhailova

Moscow State University / Institute of Linguistics RAS

This paper has been conceived as a response to Stephen Pax Leonard's article “Hipponyms in Indo-European”. The idea of contrasting names for 'horse' in the `language of the gods' and the `language of men' certainly seems interesting. Nevertheless, empirical diachronic study of the use of the relevant terms in Continental and Insular Celtic is conducive to a different hypothesis. Analysis of the cognates of PIE *mвrkos in Germanic and Celtic makes me propose the idea of semantic convergence between *markos and the name of the female demon *mara as a result of paronymic attraction.

Keywords: Indo-European reconstruction; horse names; borrowing; wandering words; Celto- Germanic isoglosses; Altaic languages; etymology; metaphors; semantic shifts.

Т. А. Михайлова. Night-mare: о происхождении одного тропа в кельтских и германских языках (ответ С. П. Леонарду)

Статья представляет собой полемические рассуждения по поводу работы С. П. Леонарда «Гиппонимы в индо-европейском». Идея противопоставления обозначений лошади в «языке богов» и в «языке людей» представляется интересной. В то же время, более детальный анализ употребления лексем в Островном и Континентальном кельтском данной гипотезе противоречит. Анализ употребления когнатов и.-е. *mвrkos в кельтском и германском заставил меня высказать предположение о возможной семантической конвергенции между *markos и *mara -- обозначением демона женского пола, которая возникла в результате паронимической аттракции.

Ключевые слова: индоевропейская реконструкция; обозначения лошади; заимствования; бродячие слова; кельто-германские изоглоссы; алтайские языки; этимология; метафора; семантический сдвиг.

Stephen Pax Leonard's paper on “Hipponyms in Indo-European” is concerned with the subject of a very long-standing, more than well studied, and yet still relevant issue: reconstruction of several synonyms with the general meaning `horse' reflected in a variety of IE dialects. The author's proposal to distinguish between the reflexes of IE *hiйk'uo- and *mвrkos as respectively going back to the `language of the gods' versus `language of men' seems intriguing, yet one can hardly accept it at face value. Without any intention to engage in direct polemics, but rather in the spirit of further elaboration of the many observations made in this paper, I would like to contribute my own, somewhat different, view on the problem set out by the author.

In Celtic and Germanic, there is a word for `saddle-horse' and/or `female horse', attested only in these two language families and reconstructed by Julius Pokorny at the PIE level as *mark-o- (IEW 700; see also Watkins 2011: 52). A similar etymon is suggested in the Etymological Dictionary of Proto-Celtic by Ranko Matasovic who cites data from all extant and extinct Celtic languages (OIr. marc, MW march, OBret. marh, MoBret. marc'h, OCo. march, Gaul. markan), admitting, however, cross-linguistic differences in the use of this lexeme (Matasovic 2009: 257). Thus, already in the Middle Irish period marc “is a rare, poetic word, but the derivative mar- cach 'horseman' is attested in OIr.” [ibid.]. Nevertheless, in Goidelic the noun has survived into the present, and, moreover, produced the abstract noun marcachas `riding' as well as a pseudoverbal noun derived from the non-existent verb marcadheacht, which in some dialects also refers to riding a donkey and even to traveling by cart, by car, or sometimes even to sailing; see also Scottish Gaelic marcachd `act of riding'. Despite the relatively frequent use of this lexeme (or perhaps because of it), it has acquired the generic meaning `transport' rather than `horse' or even `saddle-horse'. In Brittonic languages, the derivatives of *mвrko- are well represented; however, Middle Welsh march is not the basic word 'horse' but rather a narrow term for `saddle-horse' (cf. also Welsh marchog `horseman, knight', while epic and poetic texts would rather use the generic term ceffyl, see Jones 1997). In Breton, marc'h designates `horse' as such (see LEIA III: 20). In other words, while this lexeme is present in nearly all the Insular Celtic languages, its connotations differ.

Furthermore, according to Patricia Kelly,

“The simple contrast of OIr. ech versus MW march implied here must, however, be modified to accommodate an Early Irish form of *marko-, namely marc, and the associated marchach `rider'.

Greene (1972), observing the contrast between the societies depicted in the two insular literatures, concluded that riding, and the terms denoting it, were borrowed into Ireland from Britain. According to this theory, OIr. marc is not inherited from Continental Celtic, but is a Welsh loan-word, and marcach a calque on W. marchawc `rider'” (Kelly 1997: 46).

But this supposed borrowing into Irish has not superseded the old Indo-European term reflected in Proto-Celtic as *ek"os > OIr. ech `horse', MW ebol `colt' (cf. Latin equus `horse', Greek hippos with the same meaning, Sanskrit вsva, Old English eoh, Lithuanian asvа `female horse, mare'); the PIE form for all these cognates is usually reconstructed, with some degree of approximation, as *hiek'wo- (see Mallory and Adams 2006: 139, 154, and, for a survey of literature on the subject, Pereltsvaig and Lewis 2015: 170-171). In Irish, the generic term ech was later superseded by capall, whose primary meaning was `carthorse, draught horse'. The origin of the latter is not quite clear. Although a folk etymology interprets it as a Latin loanword, Joseph Vendryes believed that Latin caballus was itself borrowed from Celtic (see Gaulish Nomina Loci Cabillonum, Caballio etc.), while the rare Greek word KapaAAqз was also borrowed from the Celtic-speaking Galatians (LEIA VI: 33-34). In addition, Welsh ceffyl (also dialectal cafal and carall) is attested with the generic meaning `horse'. Naturally, there is a possible link with Russian kobyla `mare' and Lithuaninan kumлle of the same meaning, but it seems hard to establish either regular correspondences or a scenario of borrowing in this case. J. Loth once wrote a paper on the functional history of this noun and the evolution of its phonetic variants (Loth 1933). He suggested to analyze it in the context of the entire group of Slavonic lexemes like konj-i, komon', kobyla and made a number of hypotheses on possible sources of borrowing of the root under discussion, without settling on any of them definitely. As a cautious guess, he offered parallels with Finn. hebo `mare' and Norv. hoppe `mare'. In my opinion, we are dealing here with a Wanderwort in the European area, and tracing its ultimate source is likely impossible: a designation of such an important domesticated animal is expected to migrate from dialect to dialect, becoming part of the so-called “cultural lexicon”, whose status is language- independent (see also *kobyla in Trubachev 1983: 93-98; Derksen 2008: 231-32)).

In this particular case, I do not focus on the actual origin of the word or ways of its expansion, but rather stress that it seems to have become a successful competitor of the generic term for `horse', with the prevalent meaning of `draught horse'. Thus, in Goidelic the opposition between different words for `horse' conveys both functional and sociolinguistic differences. This opposition could indeed be treated as an equivalent of “Gods' vs. Men's language”, just as Leonard proposes, but it is formally expressed through different lexemes, of which at least one has no reliable PIE etymology.

Incidentally, almost the same observations apply to the PIE word for `horse'. What seems to be an unequivocally Indo-European and archaic term for `horse', *(hi)ek'wo-, at a deeper level is often tracked down to the stem *dk'u `quick' (Hamp 1990: 212); at the same time, Sergei Starostin (1988) has offered strong arguments in favor of the word having been loaned from Proto-North Caucasian *ыi[n]cwо- with some phonetic adaptation; this borrowing must have occurred at the earliest stage of Indo-European, i.e. even before the split with Anatolian. The original meaning seems to have denoted a `beast of burden', cf. Sumerian ansu `donkey', also a loanword from an unknown language (Kullanda 2008: 674). At the same time, in the Brit- tonic languages the cognates of this PIE stem were gradually marginalized and survived only in the words for `foal' (MW ebawl, Bret. ebol, OCo. ebol) and in the old tribal name Epidii.

Summing up, it is likely that in Insular Celtic we are dealing on every occasion not with the dichotomy *ekwo- ~ *marko-, but with a more sophisticated distribution of three stems: *ekwo-, *marko- and *kobil- (?). In the course of language evolution, one of the three words had to shift to the semantic periphery, as it happened to marc in Goidelic and equos in Brittonic, while the remaining two became competitors, both surviving into the present with the status of either synonyms or dialectal variants. In Irish it is each vs. capall, while in Welsh it is march vs. ceffyl.

Some aspects of the portrayal of the horse in Celtic Insular culture

Designations of `horse' are often borrowed, possibly in order to specify the exact referent: `horse' as an animal, a type of transport, as a unit of metaphorical phraseology etc. This is quite natural if we consider the importance of the horse in migrations and the overall life of humanity for almost six millennia, alongside the fact that the technique of saddling (for both riding and packing) must also have migrated from culture to culture.

Returning to the words for `horse' in Insular Celtic languages, we must note that the Medieval Irish tradition typically portrays the hero riding a chariot, while horse-riding was rather perceived as an otherworldly activity. This was pointed out, for instance, by Ann Ross in her book “Everyday Life of the Pagan Celts” (1970) with reference to the tale Togail Bruidne Da Derga (“The Destruction of Da Derga's Hostel”):

“Conaire the King, at a significant moment in the drama, sees `the three Reds'. These are three men having red hair, dressed entirely in red, and riding red horses” (Ross 1970: 72).

The horsemen who cannot be overtaken lure the protagonist king into the Otherworld where he is to die. Naturally, they are referred to as marcach. The horse-riding character who serves as a mediator between the Otherworld and the human world survived into later Celtic folklore where he would often be linked to the motif of water (a river or a lake where the hero is carried by a supernatural riding-horse, see, for instance O'Reilly 1991: 83-90).

The Welsh tradition is somewhat different; its medieval legends often portray horseriding as a conventional and unmarked activity (which, in my opinion, is a later influence of the French chivalric tradition, see Davies 1997). Nevertheless, it also describes a supernatural horsewoman on a white mare who cannot be overtaken. Thus, the tale Pwyll, Prince of Dyuet belonging to the Mabinogi epic contains the following episode (Thomson 1957: 8):

Ac wal y bydynt yn eisted, wynt a welynt gwreic ar uarch canwelw miwr aruchel, a gwisc eureit llathreit o bali amdanei, yn dyuot ar hyt y prifford a geredei hab law y'r orssed “As they were sitting, they saw a woman dressed in shining gold brocade and riding a great pale horse approaching on the highway which ran past the hill”.

This was the goddess Rhiannon, who arrived from the Otherworld. Later she married the king, was wrongly accused of their son's death and had to offer as punishment for herself to carry guests and strangers to the king's court on her back for seven years.

Most importantly, modern Welsh folklore has a character called march gwynn `white horse', a demonic apparition in the form of a white horse waylaying travelers on the roads at night. Thus, the rider as well as the horse itself act as liminal characters, mediators between worlds, and personifications of a female deity and of fertility.

The link between the female horse and the female deity was described as a “commonplace” in the Celtic tradition. Literature on the subject is quite vast; I would limit the references to Tatar 2007, which sums up numerous conclusions and observations by the earlier authors and draws multiple, sometimes unconventional, parallels. It may be substantiated by the story of a local goddess named Macha competing with the king's horses in The Debility of the Ulidians” (Hull 1968), or by the Irish ritual involving a white mare, described by Giraldus Cambrensis in the treatise Topo- graphia Hibernica in the late 12th century and compared with the Indian asvamedha ritual by

F. R. Schrцder (1927). According to him, the king of Ulidians had to perform a ritual sex act with a white mare at his inauguration ceremony, then to slaughter her and boil her meat in a cauldron. Comparison with Old Norse examples leads the authors of a monograph on Horse sacrifice in Indo-European cultures to the following idea:

“In the same way as the Irish sacrifice described by Gerald of Wales (Giraldus Cambrensis) in the 12th century, the horse sacrifice and the ritual eating of the horse's flesh appear to have been a religious act that forged special bonds between the king and his people” (Kaliff & Oestigaard 2020: 225).

The author of “Hipponyms in Indo-European” also refers to this well-known source, drawing further parallels from the Vedic asvamedha and the Roman October Equus. It is perhaps worth noticing that Giraldus in his description of the ritual sacrifice of the horse goddess refers to her not as equa `mare' (a term from the `language of the gods'), but rather as jumentum, in Classical Latin `an animal used for pulling or carrying, beast of burden' (Glare 1968: 981). However, taking into consideration that the text was compiled in the late 12th century and, moreover, by someone certainly familiar with Old French, where by that time the Latin stem had acquired the meaning of `female horse' (jument), we can assume that this was the meaning intended by the author, especially given the overall erotic context of the scene. This case, in my opinion, supports the dynamic nature of various lexemes with the overall broad semantics of `horse'. Note that the discussion above features parallel examples of semantic change involving the reconstructed PIE basic term.

Horse in Continental Celtic

In Gaulish dialects, both competing hipponyms are well represented, and it is quite difficult to determine which of the two can be called the generic term. The linguistic evidence is widely dispersed both in time (from the 3d century BC to the 3d century AD) and in space. For instance, the following piece of evidence is offered by Pausanius in his “Description of Greece” (2nd c. AD):

каі. ltctcov to фvoqa Iotw tlз qapKav цvxa ьno xffiv KeAxffiv (Paus. 10. 19.11)

“for I would have you know that marca is the Celtic name for a horse”.

This remark follows the description of a Galatian military unit consisting of three horsemen and three horses and called trimarkisia. According to Loth 1933, this implies that marcos meant `saddle-horse', although the generic meaning `horse' cannot be ruled out either. We should also consider Gaulish place-names Marco-durum, Marco-magus, Marco-lica (Delamarre 2003: 217) and personal names Marco-marus, Marco-sena, Marco-mani, Marcus, Marcula (Schmidt 1957: 123; Delamarre 2007: 226), whose precise meaning cannot be reconstructed. Thus, according to A. Falileyev, the toponym Marcomagus means `horse market' (Falileyev 2010: 157), yet it is impossible to determine whether it was a market for saddle-horses or for horses in general, which is probably more likely.

There is also a Gaulish inscription MARCOSIOR - METERNIA (Lambert 2002: 117), where marcosior is interpreted as 1sg. fut. dep., derived from a hypothetical verb meaning *'to ride' > `to copulate' with the desiderative suffix. This implies that the meaning `saddle-horse' had already developed a metaphorical aspect, and that, therefore, the word under discussion was already firmly established in the language. At the same time, it is worth noting that in the Gallo-Latin tradition Apollo, the sun and healer deity, is accompanied by horses; in Mauviиres (Indre), Apollo is called by the Celtic surname of `Atepomarus' or `great horseman' (Green 1986: 172). This epithet is again compatible with the designation of `saddle-horse'.

Yet this Continental Celtic stem is less frequent than reflexes of the PIE root *h1йk'uo-. Witness the abundance of proper names containing the element epo-: Epasnactus, Epaticus, Epato, Epetina, Epillius, Eporedorix, Eposognatus, Epotsorouidus, Eppamaigus etc. (see Evans 1967: 197198). Ellis Evans also observes “the sporadic alternation with -ku-” in a number of personal names (ibid. 197), e.g. Equaesus, Equesus, Equonus, as well as the name of the ninth month in the Coligny Calendar - EQUOS (Olmsted 1992: 199). This phenomenon can be explained by either dialectal variation or archaization, or, rather, phonetic conservation of the sacralized form in some lexemes. For instance, the proper name Equaesus has two identifiable components equo- `horse' and aesus `Aesus, theonym', which implies the literal interpretation “horse of (the god) Aesus”. Contrast the `regular' names like Eporedorix “king of riders”, Eposognatus “one who knows horses” etc.

In its derived feminine form, the PIE stem produced the name of the Celtic horse-deity par excellence and one of the most popular Gaulish deities in general, Epona, whose identity is dependent upon the presence of the horse emblem. She appears on nearly three hundred stone monuments in Gaul, being favored particularly in the east, side-saddle, astride or between two horses or foals. One of her major functions was that of a mother-goddess (Green 1986: 173).

Thus, we can reconstruct a Proto-Celtic female horse deity, certainly having fertility functions, whose cult traces survived in Insular Celtic legends as well (see above). However, this mythical figure seems not to have been tightly linked with any specific stem for `horse', since in Welsh, for instance, her name is derived from *marco- (see Green 1986: 72-102)..

Germanic data

The Indo-European stem *marko- has clear reflexes in Germanic (Proto-Germ. *marxaz according to Orel 2004: 261), which occur in West as well as North Germanic: Old Norse marr, Old English mearh (m.), mere (f.), Old High German marah `horse', mariha `mare', and Middle High German Marah > Mдhre f. (cf. Old French marahscalc > French marйchal `marshal', borrowed from Germ. *marha-skalkaz `horse-servant', Watkins 2011: 52). In Germanic, there is partial conflation of the meanings `saddle-horse' and `mare' (*marhjon-), and in Old English one finds the same obscene meaning that we have suggested for Gaulish (see above): `to ride' > `to copulate'.

F. Kluge hypothesized that all the Germanic derivatives originated from Proto-Germanic *marhп-, which supposedly was the generic term for `horse' at an earlier stage (Kluge 1957: 454). As a basic word, in English and German it would be superseded by later innovations (German Pferd, English horse). Presumably, the reflexes of Old Germanic *marhп had already been specialized and could not be used to denote `horse' in general. This opinion is shared by Mallory and Adams, who believe that at least in Proto-Germanic, this word could have been basic, but that it was later extended with additional suffixes denoting animals of different sex; in particular, the word for `mare' got the suffix *-eha (Mallory & Adams 1997: 274).

Furthermore, in Germanic languages this stem has become conflated with the word for `female ghost', Mare (ON. mara, OE. mare, Mod. Germ. mahr `nightmare') through some kind of paronymic attraction. The Germanic forms can be reconstructed as *mara, going back to the Celtic-Germanic stem *mora f. `witch, malicious supernatural female being' (IEW 736; De Vries 1962: 379). In Middle English the two stems were conflated, yielding nightmare `night-fiend', literally “night (female) horse”. On the face of it, this conflation seems accidental, but if we take seriously the idea of the Freudian Ernest Jones that the white mare represents the most archaic symbol of deep-seated human fears (see Jones 1971: 241-341), it becomes logical. The designation of the ghost horse may now be interpreted as an overlap of two semantic fields, `female horse / saddle horse' and `female ghost'. Jones proposed a tentative etymology for both Germanic words, deriving them from a more archaic element *M-R with a broader semantic field also involving such concepts as `death' and `moisture' (see Jones 1971: 327-8). Admittedly, this part of Jones 1971 belongs to the domain of “folk linguistics”, which does not, however, imply that his observations are to be dismissed easily and unequivocally (cf. the final section).

In Celtic, a reflex of the PIE stem *mora is attested in the name of the war and death goddess, Morrigain (LEIA III: 64), but one should also compare the aforementioned Welsh folklore character of the white mare attacking travellers at night. In my opinion, it is hardly possible to tell which of the two meanings is present in the line below, which belongs to the well-known Old English “Journey charm”, believed to have been modeled after Insular Celtic charms:

Ne me mer ne gemyrre (Storms 1948: 216) - `May no nightmare disturb me'.

The word mer, translated by the editor as nightmare, looks like the Middle English word for `mare, female horse' (OED VI: 158), but the author of the text seems to have meant `demon' or, even more likely, the same spooky figure of the `night horse' haunting travelers on dark roads. Jones indeed wrote about the term mare, naturally in terms of psychophysiology rather than myth; cf. also Modern French cauchemar < Picard. chauche-mar, derived from Lat. calcare > Old French chaucher `to press' and Germanic *mara).

Celtic vs Germanic?

As Edgar Polomй once pointed out,

“There are quite a few reservations that can be made about the assumption that all these (Celto- Germanic - T.M.) terms were borrowed from Celtic into Germanic. First of all, in the case of correspondences restricted to Celtic and Germanic there are always four possibilities that need to be investigated:

the terms represented either a common regional innovation in marginal areas of the Indo- European territory or the localized survival of an archaic term lost elsewhere throughout the Indo-European linguistic area;

the terms have been both taken over from the same third source - be it a pre-Indo-European (`substrate') language or a less well-documented Indo-European language in their vicinity;

the Celtic term was borrowed by Germanic;

the Germanic term was borrowed by Celtic”

(Polomй 1983: 284).

There is a variety of hypotheses that were expressed on this subject in previous literature. Thus, Vladimir Orel characterized Germanic *marxaz `horse' as a Celtic loanword (Orel 2003: 261). De Vries also tended to accept the idea of borrowing from Celtic into Germanic, but left open the possibility that it was an old Wanderwort (de Vries 1962: 380). The idea of a “Wanderwort of Eastern origin” was proposed in Matasovic 2009: 257. It is worth noticing that the borrowing of this lexeme into both Germanic and Celtic from an unknown source was first offered in Meillet 1926: 229. The idea of borrowing from the language of the Thracians, famous for their riding skills, was entertained on archeological grounds in Birkhan 1970: 393-402. As discussed at some length in Leonard's paper, a borrowing from an Altaic language has likewise been proposed in Gamkrelidze, Ivanov 1995: 832.

The fourth possibility (borrowing from Germanic into Celtic) seems to have never been considered seriously, although at least in theory it is not totally unlikely. I believe that it was ignored due to the lack of linguistic evidence for the earliest Germanic dialects, since the

Celtic languages are attested through much earlier inscriptions, dating back to the first millennium BC.

Can the Celtic-Germanic isogloss be reconstructed at the PIE level, thus vindicating Polomй's first hypothesis? Theoretically, this is possible, especially under the glottochronologi- cal scenario that places the divergence of Celtic from Proto-Indo-European around 3350 BC (Starostin apud Blazek 2007: 85), while keeping in mind that mastering the skill of horseback riding took place somewhat earlier, about 3700-3500 BC in the steppe regions (see Anthony 2008: 23). In practice, however, the emergence of a word with the specialized meaning of `saddle-horse / mare' seems unlikely to have occurred in the period before the divergence of Celtic, given its unclear provenance and its absence in Italic, Baltic and Slavic, although, of course, in some of these languages it might have been lost and superseded by other semantic innovations, given that the lexemes for `horse' often yield an astonishing number of synonyms, even within the same language. Even the original meaning of the lexeme under discussion remains unclear: `saddle-horse' or `female horse'? In principle, the idea of a common innovation whose derivatives in the descendant languages developed similar yet not identical meanings (`saddle horse' in Celtic vs. `horse' in Germanic) might still be acceptable, but only if we assume that initially the meaning of the word was different, for instance, `any domesticated horse', perhaps only used as source of meat and milk. A similar suggestion was made in Mallory & Adams 1997: 276, where the possibility of borrowing an “Eastern” word for `horse' into West PIE dialects is doubted and the proposal is made that Celtic and Germanic had their own lexeme for `wild horse', but their conclusions seem somewhat far-fetched.

Altaic or Nostratic?

At the same time, a fact worth considering is that a word of similar meaning is well represented in Altaic languages: Mongolian mцrin (wherefrom Russian merin `gelding'), Kalmyk mцrn, Evenki morin etc. (Ramstedt 1935: 266-67). The stem is reconstructed as Proto-Mongolic *mori, Proto-Tungusic *murin, attested in Korean (Middle Korean mar 'horse') and, in the most recent etymological corpus of Altaic, is traced back to Proto-Altaic *morV (Starostin, Dybo, Mudrak 2003: 945), already with the meaning of `horse'. In Sergei Starostin's Nostratic database on the Tower of Babel website (http://starling.rinet.ru), an even deeper, Nostratic-level, connection has been suggested with a potential cognate in Proto-Dravidian (*mur- 'buffalo, cow'), suggesting Proto-Nostratic *morV with the generic meaning 'livestock'. The word may have even deeper roots with additional evidence from Denй-Caucasian and Afroasiatic languages (same website, Long-range Etymologies database, filed under MVRV 'ungulate').

Theoretically, such a deep reconstruction is possible, and the semantic change `ungulate' ^ `horse' seems logical, but tracing the Celtic-Germanic etymon all the way down to a Nostratic origin and interpreting it as an inherited item of deep ancestry rather than a loanword, entails, as it seems to me, too many unprovable assumptions. In particular, such a deep reconstruction is only likely if one assumes parallel semantic shifts in different language families, which had already diverged by the time of horse domestication.

Yet another detail, however, is worth noticing. In Kalmyk and, more generally, in Mongo- lic the same lexeme (a homonym?) has the meaning `broad river' (Ramstedt 1935: 265), which curiously parallels Jones' link between the Celtic-Germanic stem for `(female) horse' and the PIE word for `water' or `moisture' (Jones 1971: 329). While linguistic speculations by Jones, who was a psychoanalyst writing over 60 years ago, must be taken with a grain of salt, he seems to have made an insightful observation on the mechanism of “paronymic attraction”.

He suggested that homophonie lexemes, despite being etymologically unrelated, can be drawn into the same semantic field whose borders are mostly blurred. This is presumably what happened to the Old English word for `female horse', linked to a female demon by folk etymology, which ultimately yielded the word nightmare. Folk etymologies are in general frequently based on paronymie attraction: thus, Russian merin `gelding' can be understood as kto versty merjaet “mile-measuring” > `fast-running'. Actually, the stem for `water moisture' (along with `sea') suggested by Jones may itself be of Nostratic origin (*mдrд, Illich-Svitych 1976: 60; for PIE *mor- and its reflexes, see IEW: 748). This stem is in turn parophonic with *mer- `death, disease, pain' as was also observed by Jones, even though his ideas are taken by many as bordering on junk science and lacking sound academic justification.

While this specific juxtaposition can hardly be taken seriously, it is worth considering a parallel in the languages of the Balkan area. According to H. Birkhan (1970: 393-402), the “Eastern” word for `horse' was introduced into the “Western language area”. Are there any traces of this phenomenon in the Balkan languages? Gindin & Kaluzskaja 1997 presents an intriguing and somewhat surprising attempt at unravelling the tangle of Balkan words with a generic (shifting) meaning of `livestock, horse, jade, carrion, garbage'. Starting off with Hungarian marha `cattle', which they assume to have been borrowed from Middle German market `property, commodity', they find an astonishing abundance of its reflexes all over the Balkan area, including Romanian dialects, Serbo-Croatian, Slovene etc. They see this lexeme as a case of `migration term radiating from the Hungarian language zone' (Gindin, Kaluzskaja 1997: 66). Without challenging these observations and conclusions, I would like to mention that the words cited in the work under discusson show amazing semantic consistency: `horse, old horse, jade, bad woman (fig.)'. Compare, for instance, Polish marcha `jade, old horse; carrion', Slovene mrha `jade' etc., a similar meaning is attested in West Ukrainian dialects for merha. Gindin and Kaluzskaja also suggest a conflation of meanings, but then, would not it be reasonable to reconstruct a similar kind of contamination at a much earlier stage of language evolution and linguistic/ethnic migrations? In other words, can these words be relics of a Nostratic stem meaning something like `livestock, horse'?

A conclusion?

So, what sort of conclusions are we to draw from all this? Perhaps we cannot tell for sure whether the Nostratic isogloss is real or what were the actual ways of transmission in the case of this apparent Wanderwort. It is obvious that the word for `horse' can change its meaning and acquire further specifications in accordance with how the horse was seen: as a wild ungulate hunted and eaten, cart-horse, saddle-horse, etc. We suspect, however, that the relevant Indo-European stems not only underwent semantic changes over the course of history, but were also influenced by paronymic stems encoded by the consonants M-R. This process was likely influenced, at least in part, by extra-linguistic factors, such as the archaic fear of night demons emerging from darkness, on the one hand, and the metaphor of copulation as riding, on the other hand. Yet, in our view, the sources of the trope are neither limited to the similarity of consonants nor derived from it. Thus, the Scythian progenitor goddess, akin to Ishtar, “was seen both as a water deity and a patroness of horses”, despite being genetically unrelated to the words under discussion (Schaub 2007: 94). A broad and archaic semantic link between the notions of `feminine' and `moist' must have played a role in this case. Thus, several factors conspired in the emergence of the trope of a female demon, “mare”, conceptualized in folklore as both a woman and a female horse.

References

Anthony, David W. 2008. A New Approach to Language and Archeology: the Usatovo Culture and the Separation of Pre-Germanic. Journal of Indo-European Studies 36(1-2): 1-51.

Birkhan, Helmut. 1970. Germanen und Kelten bis zum Ausgang der Rцmerzeit. Wien: Hermann Bцhlau.

Blazek, Vaclav. 2007. From August Schleicher to Sergei Starostin: On the development of the tree-diagram models of the Indo-European languages. Journal of Indo-European Studies 35(1-2): 82-110.

Davies, Sioned, 1997. Horses in the Mabinogion. In: Sioned Davies, Nerys Ann Jones (eds.). The Horse in Celtic Culture. Medieval Welsh Perspective: 121-140. Cardiff: University of Wales Press.

Delamarre, Xavier. 2003. Dictionnaire de la langue gauloise. Une approche linguistique du vieux-celtique continental. Paris: Йdition errance.

Delamarre, Xavier. 2007. Noms de personnes celtique dans l'йpigraphie classique. Paris: Йdition errance.

Derksen, Rick. 2008, Etymological Dictionary of the Slavic Inherited Lexicon. Leiden / Boston: Brill.

De Vries, Jan. 1962. Altnordisches Etymologisches Wцrterbuch. Leiden: Brill.

Evans, D. Ellis. 1967. Gaulish Personal Names. A Study of some Continental Celtic Formations. Oxford: Clarendon Press.

Trubachev, O. N. (ed.). 1983. Etimologicheskij slovar slavianskikh jazykov. Vol. 10. [An Etymological Dictionary of Slavic Languages. Vol. 10]. Moscow: Nauka.

Falileyev, Alexander. 2010. Dictionary of Continental Celtic Place-Names. Aberystwyth: CMCS.

Gamkrelidze, Thomas V., Vjaceslav V. Ivanov. 1995. Indo-European and the Indo-Europeans. A Reconstruction and Historical Analysis of a Proto-Language and Proto-Culture. Translated by Johanna Nichols. Berlin / New York: Mouton de Gruyter.

Gindin, Leonid A., Irina A. Kaluzskaja. 1997. Ob odnom ungarizme v Karpatskom areale: *marha [On a term of Hungarian origin in the Carpathian area: *marha]. In: T. M. Sudnik, Ye. A. Khelimskij (eds.). Balto-Slavyanskije issledovanija 1988-1996 [Balto-Slavic Studies 1988-1996]: 66-71. Moscow: Indrik.

Glare, Peter G.W. 1968. Oxford Latin Dictionary. Oxford: Clarendon Press.

Green, Miranda. 1986. The Gods of the Celts. Surrey: Brameley Books.

Greene, David A. 1972. The War Chariot as Described in Irish Literature. In: Charles Thomas (ed.). The Iron Age in the Irish Sea Province: 59-73. London: Council for British Archeology.

Hamp, Eric. 1990. The Indo-European Horse. In: Thomas L. Markey, John A. C. Greppin (eds.). When Worlds Collide. Indo-European and Pre-Indo-European: 211-226. Michigan: Karoma Publishers.

Hull, Vernam (ed.). 1968. Noinden Ulad: The Debility of the Ulidians. Celtica VIII: 1-42.

IEW = Pokorny, Julius. 1959. Indogermanisches etymologisches Wцrterbuch. Bern: Franke Verlag.

Illich-Svitych, Vladislav. 1976. Opyt sravnenija nostraticheskikh jazykov. Sravnitel'nyj slovar'. [A comparative dictionary of Nostratic languages]. Moscow: Nauka.

Jones, Ernest. 1971. On the Nightmare. New York: Liveright.

Jones, Nerys Ann. 1997. Horses in Medieval Welsh Court Poetry. In: Sioned Davies, Nerys Ann Jones (eds.). The Horse in Celtic Culture. Medieval Welsh Perspective: 82-101. Cardiff: University of Wales Press.

Kaliff, Anders, Terje Oestigaard. 2020. The Great Indo-European Horse Sacrifice. 4000 Years of Cosmological Continuity from Sintashta and the Steppe to Scandinavian Skeid. OPIA 72. Uppsala: Uppsala University.

Kelly, Patricia. 1997. The Earliest Words for `Horse' in the Celtic Languages. In: Sioned Davies, Nerys Ann Jones (eds.). The Horse in Celtic Culture. Medieval Welsh Perspective: 43-63. Cardiff: University of Wales Press.

Kluge, Friedrich. 1957. Etymologisches Wцrterbuch der deutschen Sprache. Berlin: Walter De Gruyter.

Kullanda, Sergej. 2008. Loshad' v praindoevropejskom [The name of the Horse in Proto-Indo-European]. In: Ilya Smirnov (ed.). Aspects of Comparative Linguistics. Orientalia et Classica. Papers of the Institute of Oriental and Classical Studies. Issue XIX: 669-678. Moscow: RSUH Publishers.

Lambert, Pierre-Yves. 2002. Recueil des Inscriptions Gauloises. Textes Gallo-Latins sur instrumentum. Vol. II. Paris: CNRS Editions.

LEIA = Vendryes, Joseph. 1959. Lexique йtymologique de l'irlandais ancien de Joseph Vendryes. Vols. I-VII. Dublin Institute For Advanced Studies.

Loth, Joseph. 1933. Les noms du cheval chez les Celtes en relation avec quelques problиmes archйologiques. Mйmoires de l'Institut national de France 43(1): 113-148.

Mallory, James P., Douglas Q. Adams. 1997. The Encyclopedia of Indo-European Culture. London: Fitzroy Dearborn.

Mallory, James P., Douglas Q. Adams. 2006. The Oxford Introduction to Proto-Indo-European and the Proto-Indo- European World. Oxford University Press.

Matasovic, Ranko. 2009. Etymological Dictionary of Proto-Celtic. Leiden: Brill.

Meillet, Antoine. 1926. Caractиres gйnйraux des langues germaniques. Paris: Hachette.

OED = Murray, J. H. et al. (eds.). 1933. The Oxford English Dictionary. Oxford: Clarendon Press.

Olmsted, Garrett. 1992. The Gaulish Calendar. Bonn: Rudolf Habelt.

O'Reilly, Barry. 1991. River Claiming Its Due & the Sod of Death Predestined: Two Legends of Fate in Irish Tradition. Bйaloideas 59: 83-90.

Orel, Vladimir. 2003. A Handbook of Germanic Etymology. Leiden / Boston: Brill.

Pereltsvaig, Asya, Martin W. Lewis. 2015. The Indo-European Controversy. Facts and Fallacies in Historical Linguistics. Cambridge University Press.

Polomй, Edgar. 1983. Celto-Germanic Isoglosses (Revisited). Journal of Indo-European Studies 11: 281-298.

Ramstedt, Gustaf John. 1935. Kalmьckisches Wцrterbuch. Helsinki: Lexica Societatis Fenno-Ugricae.

Ross, Ann. 1970. Everyday Life of the Pagan Celts. London / New York: B. T. Batsford LTD.

Schaub, Igor J. 2007. Mif, kult, ritual v Severnom Prichernomorje [Myth, Cult, Ritual in the North Pontic Area VII-VI BC]. Saint-Petersburg: Sankt-Petersburg University Press.

Schmidt, Karl Horst. 1957. Die Komposition in gallischen Personennamen. Tьbingen: Max Niemeyer Verlag.

Schrцder, Franz Rolf. 1927. Ein altirischer Krцnungsritus und das indogermanische Rossopfer. Zeitschrift fьr celtische Philologie 16: 310-312.

Starostin, Sergej. 1988. Indoevropejskie i Severno-Kavkazskie isoglossy [Indo-European and North-Caucasian isoglosses]. In: G. M. Bongard-Levin, V. G. Ardzinba (eds.). Drevnij Vostok: Etnokul'turnyje sv'azi [Ancient East: ethnic and cultural relationships]: 112-163. Moscow: Nauka.

Starostin, Sergej, Anna Dybo, Oleg Mudrak. 2003. Etymological Dictionary of the Altaic languages. Leiden: Brill.

Storms, Godfrid. 1948. Anglo-Saxon Magic. Nijmegen: Martunus Nijhoff.

Thomson, Robert L. (ed.). 1957. Pwyll Pendeuic Dyuet. Dublin: DIAS.

Tatar, Maria Magdolna. 2007. The Myth of Macha in Eastern Europe. Journal of Indo-European Studies 35(3-4): 323-344.

Watkins, Calvert. 2011. The American Heritage Dictionary of Indo-Europeans Roots. Boston / New-York: Houghton Mifflin Harcourt.

Размещено на Allbest.ru