original scientific article	DOI 10.19233/ASHS.2016.46
received: 2015-10-15
DIAL. SLOVENE *krves- AND THE ACCENTUAL HISTORY OF PROTO-SLAVIC *kry 'BLOOD'
Luka REPANŠEK
University of Ljubljana, Faculty of Arts, Department of Comparative and General Linguistics, Aškerčeva cesta 2,
1000 Ljubljana, Slovenia e-mail: luka.repansek@ff.uni-lj.si
ABSTRACT
Starting from an originally sigmatic inanimate noun as the likely source of the later Slavic feminine *kry 'blood' (= Old Irish cru), the author traces the accentual history of Late Proto-Indo-European *kru-s0, Csg *kruu-es (< *kruH2-es-s) on its way to the emerging complexity of the Proto-Slavic reflexes, viz. *kry, Csg *krhu-es-e (projected into Proto-Slavic as a peripheral variant of the ubiquitous non-sigmatic stem), and the athematic/i-stem feminine *kry, Csg *krhue/*krhui.
Keywords: Proto-Slavic *kry, Slovene dialectal material, Old Irish cru, sigmatic stems, accentology, accentual
mobility, analogy
LO SLOVENO DIALETTALE ^ktves- E LO SVILUPPO ACCENTUALE DA PROTOSLAVO *kry 'SANGUE'
SINTESI
Il materiale dialettale sloveno rivela come per il sostantivo protoslavo *kry 'sangue' (= antico irlandese cru) sia necessario postulare due paradigmi paralleli: una base lessicale sigmatica *kry, gen. sg. *krhu-es-e di genere neutro e una base lessicale semplice, senza affissi derivativi, ovvero un sostantivo con tema in -i *kry, gen. sg. *krhue/*krhui di genere femminile. Nel presente articolo si vuole avvalorare con lo studio dell'evoluzione accentuale l'ipotesi, giä avanzata nella letteratura scientifica, secondo cui proprio per tale duplicitä il sostantivo protoslavo *kry 'sangue' sarebbe da interpretare come riflesso della trasformazione analogica dell'originario sostantivo sigmatico di genere neutro *kru-s-0 (gen. sg. *kruu-es < *kruuess).
Parole chiave: protoslavo *kry, dialetti sloveni, antico irlandese cru, basi lessicali sigmatiche, accentologia,
accento mobile, analogia
Luka REPANŠEK: DIAL. SLOVENE »kfves- AND THE ACCENTUAL HISTORY OF PROTO-SLAVIC »kry 'BLOOD', 639-646
The1 Proto-Indo-European background of PSI. *kry and its congeners is relatively clear (cf. especially NIL s.v.; Nussbaum, 1999, 402; Stüber, 2002, 176-177). At the beginning of the derivational chain must lie an abstract feminine root-noun »kruH^- 'Blutiges' vel sim.^ (either acrostatic *kröuH^-/*kreuH^- (normally ^ *kruH2-), proterodynamic *kreuH2-/*kruH2-', or possibly *kruH^-/*kruH^-' of the OInd. g-fr-type), which to all appearances seems to indeed be continued by YAv. Asg xrüm 'raw meat' (< »kruuam < *kruH2-m or *krüm < *kruH2-m by Stang's Law). The corresponding verbal value of the radical is probably preserved at least in Ved. °krü-/°kruv- < »"krüH^-.^ There is a relatively rich system of derived Caland adjectives in *kruH2-rö- (Av. xrüra- 'bloody, blood-stained', Ved. krürä- 'bloody, raw, cruel'), *kruH^-mö- (Av. xrüma- 'cruel'), *kruH^-(u)ent-(Av. xruuant- 'terrible', Lat. cruentus 'bloody'), *kröuH2-o-(PGmc. *xrau(u)a- 'raw') and its corresponding deadjectival abstract noun *krö/euH2-i- (Av. xruui°/Ved. kravi°'blood(iness)', Lat. crü-dus 'bloody, cruel') with the pertaining set of adjectival derivatives *kro/euH^-i/ei-(n)ö-(Ved. kravya- 'bloody' = Lith. kraüjas 'blood', kruvinas 'bloody' etc.). This system is parallel to that established for, e.g., PIE *Hlreudh- 'redness; rubor' and quite expectedly also includes a proterodynamic inanimate sigmatic abstract noun *kreuH2-s- / *kruH2-es- (Schindler, 1975a, 263-264, pace Hamp, 1977),4 coupled with the regular internal amphidynamic collective as continued by Lat. cruor, Gsg cruöris '(clotted) blood, gore' <
Finally, the YAv.
*kreuH2os/*kruH2-s
'to thirst after blood' < *kruuiš-ia- wi
denominative
xruuis-iia- 'to thirst after blood' < *kruuis-iä- will rather than going back to a monosyllabic *kruiš- or to a full-grade *kruuiš- < *kr3uiš- (on the latter cf. De Vaan, 2003, 228) continue a secondary, internally motivated zero grade to **krsuuiš- < *kreuH'-s- = Ved. kravis-, which is not at all necessarily based on xruui- = *xruui- < *kruH2-i-5 (see, however, Stüber, 2002, 177), but is in any case unoriginal for the expected **xrüs-iia- < *kruH^-s-ie-(or possibly/ **xr3uuiš-iia- < *kreuH2-s-, cf. in this respect Av. mqs°/mšng° 'wisdom' < *mens° beside maz-da- 'wise').
The columnal Ved. kravi-s- 'raw meat'6 (= Av. *krduuiš-) and Greek Kpea-o- 'id.' (each time evidently concretised from an older meaning *'that which is bloody') can best be viewed as generalisations of the Proto-Indo-European full-grade variant *kreuH2-s-0 ^ *kreuH'-s-, i.e. as replacements of the originally mobile pattern *kreuH2-s-0, Gsg *kruH2-es-os, Dsg *kruH2-es-ei: Ved. krav-is-0, krav-is- (kravisas, kravise etc.), Gr. Kpeaq, Kpef-ao- (Kperoq, Kpeg etc.). This scenario is much more likely7 than it would be to assume for both respective paradigms a regular phonetic reflex of an amphidynamic paradigm (viz. *kreuH2-s-0, *kr(e)uH2-s-es ^) seeing that one would in such a case almost certainly expect a generalised zero-grade in the root, and that a secondarily mobile, "amphidynamic" pattern is normally (and quite sporadically)8 only encountered in originally acrostatic, not proterodynamic neuter nouns. Cases such as the seemingly secondarily amphidynamic Hitt. Gsg lamnaš
The manuscript has been prepared with the input system ZRCola (http://ZRCola.zrc-sazu.si) developed by Dr Peter Weiss, a fellow of the Scientific Research Centre at the Slovenian Academy of Sciences and Arts (http://www.zrc-sazu.si). It is based on the talk presented at the 10th International Workshop on Balto-Slavic Accentology (IWoBA X), 16 October 2014, held at the Faculty of Arts, University of Ljubljana.
But cf. Nikolaev (2010, 139ff.), who analyses *kreuH2- as a secondary derivative to *(s)ker-. A similar idea can be found in Scarlata 1999, 88, ft. 124.
Judging from the position of the accent in the Vedic hapax mitra-kruvas 'bloodying/hurting vel sim. the mitram', but cf. Scarlata 1999, 88-89, Stüber 2002: 176, who regardless of the accent take it as a bahuvrihi. Looked at in the context, nothing decisive can admittedly be said about the exact meaning of mitra-kru-. Semantically, however, it does seem to be on a more or less equal footing with drögha-mitra-, itself an indubitable case of a bahuvrihi, in the twelfth stanza of the same hymn: RVS X.89.14
karhi svit sä ta indra cetyasad aghasya yad bhinado raksa esat mitra-kruvo yac chasane na gävah prthivyä ä-pfg amuyä sayante
»When, oh when, Indra, will there be this retribution/punishment of yours
when you will split asunder the harmfulness of evil as it strives to reach (us) [if *°H,/-H,i-s-nt-],
when those who stain with blood (their) alliances like cattle in slaughter
will lie there in that way (as one) mixed (= joined) with the earth?«
X.89.12c-d
aSmeva vidhya diva ä sTjänas tapisthena hesasä drögha-miträn
»Like a stone which has been released from heaven pierce with the hottest weapon those whose alliance is a deception!«
Proto-Indo-European ablauting s-stems belong to the older layer of such suffixal formations, only rather later receiving a non-ablauting counterpart in the ubiquitous *CeC-os type.
Cf. Ved. tuvis° for expected **tavis° = Av. t3uuiš- 'raw power, force' beside the well-represented simplex tavisi- 'strength'.
With regular accent shift (pace Schindler, 1975a, 265): *krauH'-s- > *kraui-š- > *kraui-š- as in, e.g., rayi- 'wealth, possessions' < *reHl-i-
'what is given'.
Cf. Stüber (2002, 21-22, 1 77), who supposes inter- or, in the case of Vedic, intra-paradigmatic contamination: *kreuH2-os-0, *kreuH-es-^ Ved. (or generally Indo-Iranian?) *krav-i-š- (cf. arcts-), Gr. *kref-as-.
The normal result of a secondary transfer to the mobile paradigm is of course the proterodynamic pattern.
2
3
4
7
Luka REPANŠEK: DIAL. SLOVENE »kfves- AND THE ACCENTUAL HISTORY OF PROTO-SLAVIC »kry 'BLOOD', 639-646
'name' (ntr.) < »H^nömn-os or, for that matter, the likes of YAv. Apl psrs9uu5 'ford' (m.) < *pr-tu-ms,'9 do not in fact represent counterexamples to such an observation, seeing that the former certainly owes its ablaut pattern to the generalisation of the o-grade variant of the originally static10 rather than a dynamic paradigm (*Hlnöm-n/*Hlnem-n-(o)s ^ *Hlnm-en-s as in *u(e)d-en-os for »ued-n-s etc.) and as such reflects a virtual »H^nem-n-s ^ *Hnem-n-os (= type seen in Hitt. genuuaš < »gen-u-os, mehunaš < *meH2-un-os etc.) not *Hln(e)m-n-es, while the latter can still be unproblematically reconcilled with the normal proterodynamic paradigm *per-tu-/*pr-teu-, cf. Ved. Apl dyun 'heaven' < *di-u-ns (OAv. Apl paratUš < *pr-tü-ns) beside div-as 'id.' = Hitt. ši-mu-uš 'god' < *di-u-ms, both latter examples within an originally hys-terodynamic paradigm. The Hittite Gpl iš-ša-aš 'mouth', if it goes back to a proterodynamic sigmatic neuter *HieH3-s (remodelled to coalesce with the nepiš-type), Gsg *HlH3-es-os,11 must surely continue the latter par-oxytonon with the morphonologically restored vowel of the suffix (*H1 oH3-es < *H1 eH^-es : *HlH3-es-os) rather than a secondarily amphidynamic *HlH3-s-ös. The typical representatives of such a pattern, viz. the likes of Hitt. Gsg iš-ha-na-a-aš 'blood', however, are generally difficult to account for with reasonable certainty as such heteroclite neuters typically have an originally amphidynamic collective by their side, so that Hittite iš-ha-na-a-aš may be equated with the Ved. Gsg asnas 'id.' < *HlesH2-n-6s, a marginally mobile replacement of the acrostatic *H1 esH^-n-s (^ *H1 esH^-en-os), but may just as likely go back to *H1 (e)sH^-n-ös of the accompanying collective (that is if the sequence does not in the end stand for *išhan-, based on the locatival *H1 (e)sH-en-12).1^
In a 1009 lecture (handout dated 2002/2003), Prof. Furlan brought the Slavic and tentatively also the Old Irish comparanda (for an attempt at justification see Repanšek, 2010) into the question of the existence of a
proterodynamic sigmatic stem and proposed to see in PSl. *kry 'blood' and Old Irish cru 'id.' the generalisations of the theoretically predictable but until then unidentified zero-grade, which in any case would have been completely ousted in Indo-Iranian and Greek. The cumulative evidence would then by reciprocal reconstruction confirm the theoretically surmised proterodynamic ablaut distribution in the PIE sigmatic neuter *kreuH2-es- (see Repanšek op. cit., Furlan, 2011). This is of course in dire contrast with the communis opinio, which normally recognises in PSl. kry and Old Irish cru athematic feminine nouns, identical to Avestan xru- and thus representatives par excellence of the PIE animate root noun14 (NB that the respective paradigms of both Slavic and Celtic representatives have too been often equated, and that at least since Pokorny, 1917). The latter view may be deemed problematic especially from the point of view of the implications that such an interpretation has for the gender of the congeners in question. It may not be coincidental that Old Irish Nsg cru is very ambiguous as to its original gender, and much like in Slovene, Čakavian, Slovincian, Polabian and Old Polish, where the outcome of PSl. *kry still serves as the common form for both the nominative and accusative singulars, formally identical with the accusative. For an alternative analysis of the Old Irish Asg cru early application of Stang's Law must be assumed, but this is problematic at best as one would then rather expect *kruuKn (cf. Avestan xrüm if from *kruH2-m), resulting in Old Irish **croi (cf. OIr. cnaoi (LU 7329) = *cnoi to cnu 'nut'): Nsg cru < *krü-s, Gsg crou (> crau ~ crao > cro) < *kruu-os, Dsg cru < *kruu-i,15 Asg cru < *krüm (?)
The Slavic data is less straightforward, owing to the significantly greater variety of the attested forms. These can be grouped together into three more or less dominant inflectional patterns:
a) The feminine long ü-stem *kry, Gsg *krhue, Asg *kr^ub, with the accusative singular *kr^ub usurping the
9	For the alleged reflexes of the so-called rhizodynamic/acrodynamic pattern see Tremblay, 1998; idem 2003, 82.
10	For a resuscitation of this older view see especially Stüber, 1997; eadem 1 998, 53ff.; cf. Pinault, 2003, 162. See, however, Neri, 2005 for a very sound attempt to salvage the more generally accepted idea of an amphidynamic collective *H1 neH-m5(n)/*Hl nH3-mn-' (side by side with its immobile neuter counterpart *H1 neieH-men-, which, nevertheless, seems unlikely precisely because of the combined Anatolian data that so clearly points to *H1 nöm-n-/*Hlnm-en-). There is in fact nothing that can be deemed absolutely fatal to the projection of the second laryngeal in the root (while anlauting *H1 is of course incontestable) and at least Proto-Italic *n5mn, *n5men- (cf. Oscan numn- < *n5men- with the unambiguous reflex of a long vowel) seems to speak rather strongly, if not irreproachably, in its favour (note that PAlb. *ameno- < *anmeno- < *HlnH3men-o- vs. a direct *ameno- < *Hlnm-en-o- does not fare any better than PBSl. oblique *imen- < *inmen- < *Hlnm-en-, dissimilatory loss having to be invoked either way). It is, however, evident that the majority of the comparanda rest on a marginally mobile paradigm *H1 n(ö)(H3)(-)m(-)n-, oblique *H1 n(H)(-)m(-)en-, which certainly is not original (as is made evident by Hittite and Indo-Iranian) but neither can it easily have arisen as a transformation of an amphidynamic pattern. An acrostatic starting point is, so it seems, inevitable.
11	Cf. Stüber, 2002, 195.
12	Cf. s-an- in Lat. sanguis 'blood'.
13	The Greek type nüp, nupö? 'fire' < *pH--ur-ös ^ *pH-un-ös is equally ambiguous. If the proto-form *puH2-r (rather than *puH--r!.) > *pür were based on the strong stem of the originally proterodynamic *peuH2r< *peH2-ur, its Gsg *pH2-un-ös would necessarily reflect a marginally mobile replacement of *pH2-uen-. On the other hand, non-Anatolian IE *p(H-)ür 'fire' (= Umbr. pir, Arm. hur etc.) could easily be seen to reflect a secondary, purely analogical formation based on the oblique stem of the formal amphidynamic collective *pH--un-', based off of an acrostatic neuter.
14	Cf. the highly archaic OIr. ru 'red colour' < *H1 rudh- (for the declension see GOI §323).
15	See Joseph, 1988, 181ff.; Uhlich, 1 995, 22, 28.
Luka REPANŠEK: DIAL.	^k^ve^ ,fND THE ACC[^NT^UAL I^ISTORY C^^ f^RO^C^-SLAVG*k^)'', 639-M6
plare 06 the	*k:ry uukide	C^i^V^E^vOn,
S^c^H^icN^i^i^, ^ola t^ian adid Old Polish:
Nsg *kry ^ *kr'bub (< Asg): Sln. kri,Č^I<. kri,
krä, ^. k—)räi, (^F^oI.At); vSeC^c^;5I. Kp+hlih (Sta.
<	*kru-s
Gsg *krbue (OCSI. KpiiBe etc.) < *kru-u-es
Asg *krtub,virtually<*krM-u-m.
The attested accentual pattern is fully parallel to that of the hysterodynamic PSl. *d-bti 'daughter' < *duk-'te, Gsg *dbtere < *duk-te'-re, Asg *dht'erb = Lith. dukteri 'id.' < *duk-'ter-im and the amphidynamic type *(o)bry, Gsg *(o)brbue < *-uu-es, Asg *(o)bri>ub (^ *dbrbub) 'eye-brow' < *-'üu-im,1^ both marginally accented accusative singulars being best explained by the older, already Balto-Slavic rule of accent retraction known as Pedersen's Law.
b)	The more recent feminine /-stem pattern:
Nsg *kry ~ *krhub (secondarily replaced by the Asg; on the process cf. PSl. *kamy 'stone' ^ *kamenb
<	Asg.)
Gsg *krhui (as in Russ. Kpoeu, Sln. krvi etc. for original *krbue by the following chronology of analogical remodelling: *krbue ^ *krhue ^ *kriui)^''
Asg *krhub
c)	An originally neuter sigmatic stem, preserved by a number of Western Slovenian dialects as has now been clearly established and aptly explained by Furlan, 2011:'8
WSln. Nsg *kri, Gsg *krvesa/i1'9
As no convincing model has been found which would successfully account for the secondary creation of such neuter sigmatic forms (see Furlan op. c/t., p. 13), this particular archaic and peripheral pattern should from the point of view of PIE (that is if one does not want to reconstruct an athematic feminine stem bes/de a sigmatic neuter for Slavic, which incidentally is not as likely as it is unnecessary) be easily proved to be the probable starting point of all the later productive patterns. In Furlan, 2011 (esp. 14-16) a plausible scenario has already been put forward which would account for the reshiftings that took place in the argu-
ably originally proterodynamic sigmatic declension after it began to acquire the characteristics of, on the one hand, the inherited baritone non-ablauting sigmatic stems (i.e. the *nebo/*nebese type) and, on the other, those of the feminine long w-stems. The relative chronology can, however, be further refined through careful consideration of the previously neglected ac-centological data.
The paradigm underlying the Western Slovene set of sigmatic forms can be securely reconstructed as Common Slovene *kri, *krvesa/i, basing the reconstructed accent pattern on the admittedly sparse but reliable set of telling attestations such as Ter ot kbrv^s^0 (prepositional genitive) = ksrve:s^^ < *krvesi and, e.g., Log pod Mangartom k(3)rbiese/a < *krvesi/a.22 In theory, Common Slovene *kri, *krvesa/i could point to a defectively mobile paradigm *kry, Gsg *krbu-es-e, Dsg *krbu-es-i of the *uxo, *ušese (< *ušese) type, assigned to accentual paradigm (AP) d. It should be noted in passing, however, that the reconstructed pattern of the so-called AP d is altogether problematic. It is true that an original *uxo, Gsg *ušese 'ear' > PSl. *ušese (NApl *ušesa > *ušesa) better accounts for the Slovene type uho 'id.', Gsg = NApl ušesa than *uxo, *ušese vs. NApl *ušesa, i.e. a mobile paradigm, resulting in dial. Sln. uho, Gsg ušesa : NApl ušesa, which however cannot do without analogical modification of the NApl *ušesa to *ušesa (on the latter still rather poss/ble development see Snoj, 1996, 292). But it is also true that this kind of analogy needs to be taken into account in the case of all other AP c sigmatic neuters anyway and that Gsg = NApl ušesa is not in fact the dominant pattern in nonstandard Slovene. Neither is there any immediately obvious reason why this particular s-stem should have behaved differently than any other inherited neuter of a comparable morphological shape. Note that even if a reflex of an originally proterodynamic paradigm is assumed for Slavic, the marginally mobile *H2eus-os, Gsg *H2us-es-(o)s (cf. Stüber, 2002, 193-194)23 would again result in Proto-Slavic AP c *uxo, Gsg *Wšese > *ušese. In the end it all boils down to the question of whether one wants to favour a two-part analogy to account for the type Gsg = NApl ušesa in AP c over the possibility that the eccentric pattern Gsg = NApl ušesa in AP d was liable to attraction to the predominant pattern with Gsg e : NApl e in other structurally comparable sigmatic neuters. Either way analogical levelling has to be invoked.24
16	Cf. Snoj, 1994, 491-493, 514-515, 526.
17	Thus convincingly explained by Snoj, 1994, 492.
18	A resumptive presentation of the dialectal material is given in Orel, 2015.
19	For a comprehensive list of relevant attestations and their detailed interpretation see Furlan, 2011, 9-13.
20	De Courtenay, Glossar/o del d/aletto del Torre. See Spinozzi Monai, 2009, s.v.
21	For w as a graphic representation of the reflex of *e compare zvbcwr 'in the evening' (op. c/t., s.v.).
22	Exactly like ukü, učiesa 'eye'; uxü, usiesa 'ear' etc. (SLA pre-printed dictionary slips), with 'le for e.
23	I.e. as in Hitt. iššdš 'mouth' < *HH-es-ds, which must imply an older proterodynamic *HlH3-es-(o)s. Note also the highly archaic NAdu *H^(e)us-iH1'ears' < *H2(e)us-s-iHl.
24	For an analogical explanation of the Gsg = NApl ušesa type see Snoj, 1996, 293.
Luka REPANŠEK: DIAL. SLOVENE »kfves- AND THE ACCENTUAL HISTORY OF PROTO-SLAVIC »kry 'BLOOD', 639-646
However that may be, a hypothetical *kry, »krhuese, »krhuesi would of course only on surface level and only synchronically behave as a possibly AP d noun, obviously due to the monosyllabicity of the NAsg form, which was regularly assigned a circumflex intonation. Since in the case of *kri no alternative by-form **krvesa appears beside the normal *krvesa, however, the neoacute-resembling mezostatic accent »krhuese, »krhuesi (virtually AP d) would at some point have to be ousted on analogy with the reflex of the predominantly mobile type. The end result being then fundamentally the same, and this is significant, as under the assumption of an originally mobile pattern: *kry, Gsg »kriu-es-e, Dsg *krhu-es-i (i.e. exactly like »oko, »očese ...).
It is not, however, just the principle of Occam's razor that in fact militates against the former starting point but, more importantly, the uncomfortable assumption that should we want to start from an originally immobile (= virtual AP d) pattern viz. NAsg »krü, Gsg »krüu-e, Dsg »krüu-es-i, the underlying proterodynamic neuter NAsg »kreuH^-s-ß ^ »krüH^-s-ß, Gsg (»kruH^-es-s >) »kruH^-es,^5 Dsg »kruH^-es-ei would be expected to show a generalised baritone accent throughout the paradigm: NAsg »krüH^-s-ß, Gsg »krüH^-es, Dsg »krüH^-es-ei. This particular type of accent regulation in an originally proterodynamic neuter, however, is only to be expected as a direct consequence of the synchronous generalisation
of the full-grade stem26 (cf. the type observable in PSl. »uert-men > »uer-m^ 'time, weather', Gsg »uert-men-e > »uer-men-e).
The Proto-Slovene paradigm »kri, »krvesa/i, »krvesu/i therefore rather clearly points to AP c, which expectedly translates the original mobility of the inherited pattern in a proterodynamic neuter (NAsg »kreuH^-s-ß ^ »krüH^-s-ß, Gsg »kruH^-es (< »-es-s), Dsg »kruH^-es-ei). This pattern is fully parallel to the one plainly observable in PSl. »ibm^ 'name' < »nmen (with »-en for »«, which can easily be analogical),27 Gsg »ibmene < »ibm-en-e < »H^nm-en-s, again with leftward accent shift from the penult by Pedersen's Law.28
Subsequent development is easily envisaged. The morphologically highly abnormal and hence unstable inherited Proto-Slavic paradigm NAsg »krü^,29 Gsg »kru-u-e, Dsg »kru-u-es-i30 etc. would immediately be regularised in favour of the oblique stem. The transference of the suffixal morpheme »-es- from the rest of the oblique cases into the anomalous genitival form resulted in the mobile sigmatic paradigm (Gsg »kru-u-e ^ »kruu-es-e > »kriu-es-e, by Pedersen's Law), which still enjoys limited productivity in the Western dialects of Slovene (I). Simultaneously, however, a reinterpretation of the inherited genitive »kru-u-e as a combination of the stem »kruu- and desinential »-es triggered the generalisation of the asigmatic »kruu- in the oblique
25	Proto-Indo-European seems not to have tolerated any heteromorphemic geminates (consider the likes of PIE »g^öm 'cattle' (Asg) < »g'öm-m by assimilation from an older »g'ou-m, descriptively a "Stang's Law" development). It is important to note, however, that the loss of a segment as the immediate result of simplification in a *C,C, cluster did not result in compensatory lengthening (i.e. mora-transfer) of the preceding vowel if the lost segment was a fricative: PIE »H^ssi > »H^esi 'you are', PIE *H^us-s-es (Gsg) > *H2us-es, as preserved by Ved. usäs 'dawn', very similarly PIE *H2us-s-^H1 (NAdu) > *H^us-^H1, cf. Av. uši, OPers. (u-š-i-y> = *uši 'ears, intelligence'. In much the same fashion one can envisage a straightforward development from a proterodynamic *kruH--es-s to a descriptively amphidynamic *kruH--es. It is very likely that it was the very alomorphy thus created in the orginally proterodynamic pattern that exerted enough pressure to restore transparency in favour of the stabilised variant »kreuH's- (^ Gsg *kreuH's-os etc.) in Indo-Iranian and Greek (see above).
26	To be added to Snoj's insightful observation (1993, 240, reiterated in idem 1994, 526) that »die [neutrale] Akzentparadigmen, die anfangsbetonte Formen enthielten, diese auf das ganze (singularische) Paradigma verallgemeinert haben."
27	*-(m)n-, oblique *-(m)en—> *-(m)en, oblique *-(m)en-, on analogy with the masculine hysterodynamic type in < *-en + -s (for PIE *-en#). There is no need to assume (contrary to Schindler, 1975b, 9; cf Nussbaum, 1986, 119; Snoj, 1 993, 231; Neri, 2005, 219 but cf p. 222!) that the ending goes back to a hysterodynamic singular neuter collective (i.e. semantically the Gr. ü5rop/PSl. *uoda 'water' (vs. Hitt. uitär beside uätar 'id.') type), the evidence for which, at least in the case of the PIE proterodynamic neuter men-stems (or acrostatic «-stems for that matter), is in fact vanishingly small. Neither would a hysterodynamic internal derivative be expected in an originally immobile neuter, even if it did quite naturally secondarily acquire a proterodynamic pattern. But nor is there any solid proof for an amphidynamic collective. The Proto-Germanic neuter *namö 'name', which at face value does seem to be exactly that, does not in fact unambiguously point to original *-(m)ö < *-(m)on-H^ of the collective, since in Germanic proterodynamic neuter «-stems generally acquire the ending of amphidynamic masculines, i.e. *-mö^ *-(m)ö < *(m)on-s (subject to subsequent remodelling and non-unitary split into *-ö and *-ön, the latter coalescing with the old *-ön > *-ö > *-a). There are no survivals of an overtly immobile type (note that cases such as Goth. hliuma (m.) 'hearing, audience' = 'ear' (cf. Cor. I 12:17) < *kleu-mö ~ pl. hliumans 'ears' seem to reflect the possessive derivative, which was masculine from the start; OHG sämö 'seed', however, is a good candidate for an actual collective reinterpreted as masculine singular). The fact that there arose the need to create new, analogical plurals such as PCelt. *anman-ä, PSl. *ibmen-a, Goth. namn-a etc., compared to old, inherited collectives, directly continued by Av. nämqm = Ved. nämä < *(m)on-H, is not a definite sign of a singulativisation of the inherited plural since such archaic internal derivatives could simply have been (and generally were) ousted by more productive morphology.
28	Contrary to Snoj, 1993, 233; Neri, 2005, 211 and pass., cf Pronk, 2009 pass. Note here, however, that set-root *H;neH--mn (^ »HflH-mn), Gsg *HflH3-men-s may through regular neutralisation of the accent in a mobile paradigm in the case of *HflH-men have unproblematically resulted in the same Proto-Slavic accentual pattern.
29	Circumflex intonation in a monosyllable needs no special explanation. It is not, however, strictly speaking the result of accent neutralisation in a mobile paradigm, although the ultimate result is the same.
30	Note that the non-colouration of PSl. *-e- in the suffix is to be regularly assumed in a grammatical morpheme (Repanšek, 2010, 166; Furlan, 2011, 19).
Luka REPANŠEK: DIAL. SLOVENE »kfves- AND THE ACCENTUAL HISTORY OF PROTO-SLAVIC »kry 'BLOOD', 639-646
cases and then logically led to the creation of a new dative (etc.) singular form *krhu-i ... (II):31
Nsg = Asg *kru.s
N = A *krus
Gsg *kruu-es < *-es-s G *kruu-es-e
'x
Dsg »kruu-es-ei	D *kruu-es-ei
N = A *krus
I ^ G *kruu-e ^ D *kruu-l ^
What is plainly obvious is that in both its accentual and morphological pattern the asigmatic paradigm came fatally close to the group of inherited mobile feminine stems in long -ü of the *(o)bry, *(o)brhue type.32 The correlation was sufficient to afford a successful means to renovate the non-systemic inanimate paradigm of *kry, »krhue and adapt it fully to the pattern displayed by the feminine long ü-stem nouns (II b). Note that the adaptation must also have involved a replication of the animate accusative singular form *krhub. This further caused
partial identification (especially through the accusative, dative, locative and the instrumental) with the pattern observable in the group of feminine short /-stems of the *ndtb (< *nöku-ti- 'night') type and subsequent logical introduction of the by now dominant /-stem paradigm (II c) *kry, *kr-bui, *krhui (possibly), *krhub (with subsequent and typologically expected generalisation of the accusative *krbub in the nominative singular):
II ^ II a
*kry *kr^ue
II a ^ II b = AP c
*kry *kr^ue *kryui (?)
*krbu^
II b ^ II c = AP c *kry *krbui *krbui *krbub'
+ L '*kr^ui I *kr^ubiq
31	If on the evidence of Slavic, the Old Irish data is reconsidered, it can easily be envisaged that the very same process that triggered the creation of the Proto-Slavic dative singular form *krbui could also have been responsible for the generalisation of the oblique stem *kruu-in the Old Irish paradigm. In fact, this seems to be an Insular Celtic innovation, as is demonstrated by Late British *krou (MW creu 'blood, gore' etc.), which in my opinion best represents a straightforward and purely formal thematisation of the oblique stem (as in, e.g., Goth. triu 'tree' < *dr-eu-o-):
NAsg cru < *krü-s, *kruH2-es > *kruu-a.s/e.s ^ *kruu-os (> crou)
^ oblique *kruu- (for the British forms cf. already Cowgill, 1 985, 23) t Dsg *kruu-i etc.
t Late Proto-British *krou (MW creu, MCo. crow) < *kruu-o/ä-
32	Femine stems of the *suekry/*suekr^.ue ('mother-in-law') type are a much less likely source of analogy (pace Furlan, 2011, 15), since originally (i.e. before the operation of Ivsic-Stang's Law) these had a mezostatic columnal accent: *suek-ry, *suek-ri-u-e, *suek-ri-u-b < *-ruH-- (on the reconstruction cf. also Snoj, 1994, 498-499, with a different interpretation of the accentual history, however). Note that simple affinity between the morphological patterns of the asigmatic *kry/*kr^u-e and a random disyllabic feminine uH2-stem should not be assumed to have been sufficiently strong to motivate complete integration of the former, rather eccentric pattern into the latter.
0
II
—>
—>
—»
Luka REPANŠEK: DIAL. SLOVENE »kfves- AND THE ACCENTUAL HISTORY OF PROTO-SLAVIC »kry 'BLOOD', 639-646
NAREČNOSLOVENSKO *krves- IN NAGLASNI RAZVOJ PRASLOVANSKEGA *kry 'KRI'
Luka REPANŠEK
Univerza v Ljubljani, Filozofska fakulteta, Oddelek za primerjalno in splošno jezikoslovje, Aškerčeva cesta 2, 1000 Ljubljana
e-mail: luka.repansek@ff.uni-lj.si
POVZETEK
Proterokinetični (prednjepremični) naglasno-prevojni sklanjatveni vzorec, ki se ga teoretično smiselno predpostavlja za praindoevropsko (pide.) sigmatsko osnovo srednjega spola *kreuH2-es- (prvotni pomen je težko precizno rekonstruirati, saj ni nujno, da je bila tvorjenka funkcijsko in s tem semantično gledano nomen abstractum), se preurejen zrcali v različnih posplošitvah prevojnih stopenj nekdaj premične paradigme v grščini in stari indijščini na eni strani (pide. *kreuH2-s-) in praslovanščini ter otoški keltščini na drugi (pide. *kruH2-s-). Tak, s primerjalno metodo podprt in s tem popolnoma upravičen sklep pa vendarle ogroža enako legitimna možnost, da je podoba starogrškega in staroindijskega kontinuanta predpostavljenega prajezičnega izhodišča pravzaprav rezultat paralelnega internega analoškega preoblikovanja, medtem ko je izhodišče slovansko-keltski izoglosi *krüs 'kri' (led) mogoče interpretirati tudi kot v ničti prevojni stopnji posplošeno brezpriponsko izglagolsko tvorjenko *kruH.^-s (led) ženskega spola. Prav narečno slovensko gradivo, ki opozarja na periferni soobstoj praslovanske stranskosklonske sigmatske osnove *krhu-es- ob očitnem neologizmu *krhu-, pa je tisto, ki sklep o obstoju prajezične proterokinetič-ne sigmatske osnove *kreuH2-s-0, *kruH.^-es- bistveno utrjuje in hkrati omogoča v slednji prepoznati izhodišče vsaj za praslovanski, verjetno pa tudi otoškokeltski samostalnik. V prispevku se možnost, da je izvorno praslovansko paradigmo *kry, *krhues- < *kruues- (tu rekonstruirano kot osnovo s premičnim naglasnim mestom tipa psl. *ibm-^ 'ime', Red *ibmen- < *i(n)m-en-) mogoče osmisliti kot verjetno izhodišče vsem produktivnim in arhaičnim slovanskim sklanjatvenim vzorcem samostalnika s pomenom 'kri', preverja z zgodovinskonaglasoslovnega zornega kota.
Ključne besede: praslovansko *kry, slovensko narečno gradivo, staroirsko crü, sigmatske osnove, naglasoslovje,
naglasna mobilnost, analogija
ABBREVIATIONS
A = accusative;
AP = accentual paradigm;
Arm. = Armenian;
(Y)Av. = (Young) Avestan;
Čak. = Čakavian;
D = dative;
dial. = dialectal;
du = dual;
G = genitive;
Goth. = Gothic;
Gr. = Old Greek;
Hitt. = Hittite;
Lat. = Latin;
Lith. = Lithuanian;
MCo. = Middle Cornish;
MW = Middle Welsh;
N = nominative;
OCSl. = Old Church Slavonic;
OHG = Old High German;
OInd. = Old Indic;
OIr. = Old Irish;
OPers. = Old Persian;
PAlb. = Proto-Albanian;
PBSl. = Proto-Balto-Slavic;
PGmc. = Proto-Germanic;
PIE = Proto-Indo-European;
pl = plural;
Plb. = Polabian;
PSl. = Proto-Slavic;
Russ. = Russian;
sg = singular;
Sln. = Slovene;
Slov. = Slovincian;
Umbr. = Umbrian; Ved. = Vedic
Luka REPANŠEK: DIAL. SLOVENE »kfves- AND THE ACCENTUAL HISTORY OF PROTO-SLAVIC »kry 'BLOOD', 639-646
BIBLIOGRAPHY
Cowgill, W. (1985): PIE duuo '2' in Germanic and Celtic, and the nom.-acc. dual of non-neuter o-stems. Münchener Studien zur Sprachwissenschaft, 46, 13-28.
De Vaan, M. (2003): Avestan Vowels. Amsterdam -New York, Rodopi.
Dybo, V. A. (2000): M0p^0H0n0rH30BaHHHe napa-gHrMaTHHecKHe aK^eHTHHe CHCTeMH. tom I: THnonorna h reHe3HC. MocKBa, m3hkh pyccKOH KyntTypH.
Furlan, M. (2002/2003): Členitev (pra)indoevrop-ščine. Hetitščina, izroček k predavanjem (šolsko leto 2002/2003).
Furlan, M. (2011): O slovenskem narečnem kri, kr(i)vesa 'kri, krvi' ali o prvotnem sklanjatvenem vzorcu praslovanskega *kry 'kri'. Jezikoslovni zapiski, 17, 1, 7-22.
GOI = Thurneysen, R. (1946), ^ Grammar of Old Irish. Dublin, Dublin Institute for Advanced Studies (reprinted with supplement 1975).
Hamp, E. P. (1977): Indo-European *kreuH. Indogermanische Forschungen, 82, 75-76.
Joseph, L. S. (1988): The inflexion of OIr. cru-. Eriu, 39, 169-187.
Neri, S. (2005): Riflessioni sull'apofonia radicale di proto-germanico *nam5" 'nome'. Historische Sprachforschung - Historical Linguistics, 118, 201-250.
Nikolaev, A. (2010): HccnegoBaHHa no npaHHgoeBponeftcKOH HMeHHOH Mop^onorHH. CaHKT-neTep6ypr, HayKa.
NIL = Wodtko, D. S., Irslinger, B., Schneider, C. (2008): Nomina im indogermanischen Lexikon. Heidelberg, C. Winter Verlag.
Nussbaum, A. J. (1986): Head and Horn in Indo-European. Berlin, Walter de Gruyter.
Nussbaum, A. J. (1999): »Jocidus: an account of the Latin adjectives in -idus. In: Eichner, H. & H. C. Luschützky (eds.): Compositiones Indogermanicae in memoriam Jochem Schindler. Praha, enigma corporation, 377-419.
Orel, I. (2015): Cy^ecTBHTentHoe KpH "KpoBt" no BaTpocnaBy 06naKy h b cnoBeHCKHX Hape^Hax. In: KepeMHgqeBa, C. (ed.): CiBpeMeHHHTe H3MepeHHa Ha egHo Hay^Ho npo3peHHe: goK^agH ot Me^gyHapogHaTa Hay^Ha KOH^epeH^Ha b necT Ha 150-TaTa rognmHHHa ot po^geHHeTo Ha BaTpocnaB 06naK (1864-1896). Co^Ha: Multiprint, 45-60.
Pinault, G.-J. (2003): Sur les themes indo-europeens en *-u-: derivation et etymologie. In: Tichy, E., Wodtko, D. S. & B. Irslinger (eds.): Indogermanisches Nomen. Derivation, Flexion und Ablaut: Akten der Arbeitstagung der Indogermanischen Gesellschaft. Bremen, Hempen Verlag, 153-188.
Pokorny, J. (1917): Streitfragen zur altirischen Grammatik. Zeitschrift für celtische Philologie, 11, 2-29.
Pronk, T. (2009): The accentuation of the Slavic n-stems. In: Olander, T. & J. H. Larsson (eds.): Stressing the Past. Papers on Baltic and Slavic Accentology. Amsterdam - New York, Rodopi, 101-114.
Repanšek, L. (2010): Slovansko-staroirska vzporednica praslovansko *kry in staroirsko cru 'kri'. Jezikoslovni zapiski, 16, 1, 163-170.
Scarlata, S. (1999): Die Wurzelkomposita im Rg-Veda. Wiesbaden, Reichert Verlag.
Schindler, J. (1975a): Zum Ablaut der neutralen s-Stämme des Indogermanischen. In: Rix, H. (ed.): Flexion und Wortbildung: Akten der V. Fachtagung der Indogermanischen Gesellschaft. Wiesbaden, Reichert Verlag, 259-267.
Schindler, J. (1975b): L'apophonie des themes indo-europeens en -r/n. Bulletin de la Societe de Linguistique de Paris, 70, 1-10.
SLA = The Slovenski lingvistični atlas pre-printed slips database. Ljubljana, Inštitut za slovenski jezik Frana Ramovša ZRC SAZU.
Snoj, M. (1993): Der Akzentuierung der urslawischen neutralen -men-Stämme. Linguistica, 33, 227-241.
Snoj, M. (1994): Naglaševanje praslovanskih -y/hu-osnov ženskega spola. Slavistična revija, 42, 4, 491-528.
Snoj, M. (1996): Zur Akzentuierung der urslawischen sigmatischen Nominalstämme. Razprave SAZU, Razred 2, Razred za filološke in literarne vede/Dissertationes, Classis 2, Philologia et litterae, 15, 289-299.
Spinozzi Monai, L. (2009): Il Glossario del dialetto del Torre di Jan Baudouin de Courtenay. Udine, Consor-zio universitario del Friuli - Sankt-Peterburg, St. Petersburg Branch of the Archive of the Russian Academy of Sciences - Ljubljana, ZRC SAZU, Inštitut za slovenski jezik Frana Ramovša.
Stüber, K. (1997): Urindogermanisch »h^nömn 'Name', *h2öngun 'Salbe' und der Ablaut der neutralen «-Stämme. Die Sprache, 39, 74-88.
Stüber, K. (1998): The historical morphology of n-stems in Celtic. Maynooth, The Department of Old Irish, National University of Ireland.
Stüber, K. (2002): Die primären s-Stämme des Indogermanischen. Wiesbaden, Reichert Verlag.
Tremblay, X. (1998): Sur parsui du Farhang-i-Öim, ratu-, psrstu-, pitu- et quelques autres themes avestiques en -u. Studia Iranica, 27, 187-204.
Tremblay, X. (2003): La declinaison des noms de parente indo-europeens en -ter-. Innsbruck, Institut für Sprachen und Literaturen der Universität Innsbruck.
Uhlich, J. (1995): On the fate of intervocalic *-u- in Old Irish, especially between neutral vowels. Eire, 46, 11-48.
Van Nooten, B.A., Holland, G.B. (1994): Rig Veda. A metrically restored text with an introduction and notes. Cambridge (Massachusetts), London, Harvard University Press.