Dark Matter The Root *√k̑u̯el ‘Dark, Black’

The assumption of a root *√k̑u̯el ‘dark, black’ offers new possible etymologies for Arm. šaɫax , Gk. πηλός, Toch. b kwele , Hitt. kuu̯aliu -, Gk. κύλα, Lat. culex , and Lat. color , whose derivational background will be dealt with in the course of this paper.


1
The Root *√kȗ̯ el In the Indogermanisches etymologisches Wörterbuch (p. 629), J. Pokorny mentions a root *kȗ̯ el-'schlammig' ('muddy') with a question mark, the evidence for which is claimed to be limited to Armenian (šaɫem 'to moisten, to mix the mortar' , šaɫax 'clay, mud, mortar'), and Lithuanian (švelnùs 'soft, smooth'). It is not easy to derive the latter from a proposed meaning 'muddy' , so we will have to leave it aside for now.
As a matter of fact, it seems to have remained unnoticed that there is an exact equivalent of Arm. *šaɫ elsewhere in the ie languages: The very same pre-form *kȗ̯ l ̥s/no-does indeed also account for Gk. πηλός, (Dor. πᾱλός) m. 'clay, earth, mud, mire' [Hdt.+].6 The semantic match is impeccable, and also phonologically the outcome πηλός, πᾱλός is precisely what we would expect according to Greek sound laws.7 A neat Graeco-Armenian isogloss is in itself certainly a good thing and definitely an asset, but to round out the argument we would of course want to adduce further material in favor of setting up a root *√kȗ̯ el. It would therefore be advisable to look for other potential cognates within the Indo-European languages. For that reason, it will first be necessary to ascertain a possible basic 1 The derivational function of -ax is not clear. It appears in words of different semantic fields, some of which are thought to be of foreign origin; cf. Martirosyan 2010:220, 660, 761. 2 For more examples cf. Klingenschmitt 1982:140. 3 Cf. Klingenschmitt 1982Olsen 1999:133 f.;Martirosyan 2010s.v. 4 Cf. Klingenschmitt 1982Olsen 1999:40;Martirosyan 2010s.v. 5 Cf. Klingenschmitt 1982Olsen 1999:51. 6 "Without a convincing etymology. Pre-Greek?" (Beekes 2010 s.v.). 7 *pāló-< *kȗ̯ l ̥no-or *kȗ̯ l ̥so-with *kȗ̯ -(~*k u̯ -) > Gk. p-before -al-< *-l ̥-as in πάλιν from *k u̯ l ̥h1i-. For the development of *(-)kȗ̯ -in Greek cf. Balles 2002:13ff. As G. Kostopoulos (Vienna; p.c.) points out to me, we cannot exclude a priori a parallel development of an anlauting *kȗ̯to Gk. k-, similar to the development in internal position (*-kȗ̯ -> Gk. -pp-and -kk-[ἵππος, ικκος (onomast.), πέλεκκον, etc.]), for which the exact dialectal (or otherwise conditioned) distribution has yet to be defined (cf. also Peters 2000:38857). In any case, if we may assume a likewise legitimate outcome *kāló-< *kȗ̯ l ̥no-or *kȗ̯ l ̥so-it becomes possible to account for the substantive κηλίς ῖδος f. 'stain, spot, defilement; stain, blemish; naevus' [trag., Antipho, x etc.11 Again, this account seems to be in line with the apparent development of *kȗ̯ -sequences (as opposed to *k u̯ ) in Tocharian, though the minutiae are still disputed. It is impossible to recapitulate the different doctrines here, but see Fellner 2005 for an exhaustive overview. It will suffice for our purposes to compare the supposed development of *kȗ̯ ol-o-> b kwele to *kȗ̯ on-m̥ > b obl. kweṃ 'dog'12 or to *h1ekȗ̯ os > b yakwe 'horse' (as opposed to a "real" labiovelar in similar positions: *tek u̯ os n. > b cake 'river' , *k u̯ olo-> b kele 'navel').13

11
Cf. nil:580f. 12 Maybe an inherited biphonematic sequence *Ku̯ -in antevocalic anlaut position generally developed a "pseudo-Lindeman" realization with an anaptyctic vowel *Kuu̯ -in the prehistory of Tocharian? Cf. in any case also Pinault 2008:445ff. 13 This view-viz. that inherited *K u̯ and *Ku̯ behaved differently in Tocharian b-is essentially the one outlined by Ringe (1996:42) and also Kim (1999:149ff. and 158). 14 For the process cf. Hesychius' gloss of κύλα reads τὰ ὑποκάτω τῶν βλεφάρων κοιλώματα. τὰ ὑπὸ τοὺς ὀφθαλμοὺς μῆλα. τὰ ὑπώπια. 'the cavities underneath the eyes, the μῆλα [apples?] under the eyes, the parts under the eyes' . K. Latte (1966:545) emendates μῆλα 'apples' to κοῖλα, but this is unnecessary. In lsj s.v. μῆλον (b) 'apple' , we find a separate definition (ii.3.) for the plural μῆλα as 'swellings under the eyes' exclusively for this Hesychius gloss (which, however, is emended s.v. κοῖλα to κοῖλα 'hollows, cavities'). In all the other instances where μῆλα is used in a metaphorical sense, it stands for things that do in fact have a closer resemblance to apples than a potential swelling under the eye, such as cheeks, tonsils, or a girl's breasts.
It would thus be far more plausible if μῆλα referred to the cheeks in this case. As G. Kostopoulos (Vienna; p.c.) kindly points out to me, the use of μῆλον for 'cheekbone' is well-attested in Roman and Byzantine times18 and also in Modern Greek, where μήλο means both 'apple' and 'cheekbone' . M. Divjak-Mirwald (Vienna; p.c.) very cleverly adduced the typological parallel from French, where pommette (diminutive of pomme 'apple') likewise means 'cheekbone' . The gloss could then be understood as '(either) the cavities underneath the eyes, (or) the cheekbones under the eyes, (in general) the whole area under the eyes' . 16 Cf., e.g., lsj s.v. κοῖλος. 17 For the forms cf. Frisk s.v.; also Beekes 2010 s.v. ("rather […] Pre-Greek"). 18 Cf. Sophocles 1900:757 s.v. μῆλον "cheek-bone, the part that blushes".

Indo-European Linguistics 3 (2015) 24-41
But at this point it might be advisable to have a look at another related gloss, namely κύλλια, which is explained as ὑπώπια μελανά19 'black areas under the eyes ' [Hsch.]. From this, we may conclude that both κύλα and κύλλια denoted only the dark area underneath the eyes, viz. dark circles under the eyes, and not necessarily also the cheekbones. For it now becomes possible to read μέλανα (or μελανά) for μῆλα in the other gloss as well, so τὰ ὑπὸ τοὺς ὀφθαλμοὺς + μέλανα 'the black areas or spots under the eyes' .20 But of course, such a conjecture is not a trivial emendation of the actual text, even if it may be semantically justified.
Another option would therefore be to take μῆλα 'dark parts' at face value and try to find a morphological explanation for it.21 One would not even have to give up the connection with the root *√mel 'black, dirty' of Gk. μέλας,22 but only see in it a formation with a lengthened grade, viz. *mēl-o-n. 'dark spot' that is attested in this very same meaning in Germanic (Goth. mel*, ohg māl, Germ. Mal; cf. Kluge/Seebold 200224:592). It could be regarded as an insurmountable obstacle, though, that this would be the only attestation of a presumed μῆλα 'dark parts' in the whole of Greek literature.
In this respect it might also be worth clarifying that the word ὑπώπιον of both glosses κύλα and κύλλια, which appears again in a third relevant gloss23 κύλ- Taken together, the evidence therefore strongly suggests that the Hesychian gloss of κύλα is to be understood as 'the cavities underneath the eyes, the dark areas under the eyes, black eyes' . Accordingly, we can surmise that κύλα was most probably the Greek term for 'dark circles under the eyes' and/or 'black eyes' and continues a substantivization of an adjective *kȗl-o-'dark, black' .
The very same adjective was perhaps also the basis for a marginal Latin formation. Within Latin, or rather the prehistory of Latin, a suffix containing a velar24 could apparently be used as a prototypical enlargement of color adjectives, denoting 'such-and-such a color' or 'such-and-such a colored thing' .25 For instance, we need to assume a form like *albex, gen. *albicis 'whitishness, white color' (albēre 'to be white') to account for the adjective albicāns 'whitish' [Catull.+], subalbicāns [Varro]. The latter can then be explained as a possessive denominative formation in -ant-,26 literally 'having *albex, having whiteness' that was only secondarily reinterpreted as a present participle, which gave rise to the verb albicāre 'to have a whitish tinge' [Varro+].27 For the formation compare gemmāns in a possessive sense 'decorated with jewels' from gemma f. 'jewel; bud' , which likewise cannot be a present participle, since the verb gemmāre means 'to come into buds' . According to this model, we can nicely account for the analogical creation of other color verbs in°icāre ( A third case might be *uirex, *uiricis 'greenery, green color' (uirēre 'to be green') → uirectum n. 'area of greenery' [Dirae+], parallel to arbor oris f. 'tree' → arbustum n. 'area of trees; wood, copse' ,31 unless, of course, this is an analogical formation with a complex suffix°ectum taken from frutectum 'thicket of shrubs' [Col.+] ← frutex 'shrub' (thus Weiss 2009:294141).
This being the case, we could, in principle, project a substantive *culex, gen. *culicis 'dark color' or 'dark colored thing' . And indeed, this is what seems to be attested in Pliny's Naturalis Historia, book 25, in a description of remedies for diseases of the eyes: Oculorum aciem centaurio maiore putant adiuuari, si addita aqua foueantur, suco uero minoris cum melle culices, nubeculas, obscuritates discuti, cicatrices extenuari, albugines quidem etiam iumentorum sideritide. plin. Nat. 25.142 It is believed that the greater centaury strengthens the eyesight if the eyes are treated with a mixture of water; with the juice of the smaller kind, mixed with honey, one can dispel culices, nubeculas and obscuritates and diminish scars; the ironwort, it is believed, heals albugo in beasts of burden.
The textual evidence itself-in the guise of an evident tricolon or hendiatrissuggests that culices, here, is to be regarded as a synonym of nubeculas and obscuritates, i.e., 'small clouds' and 'spots of darkness' that appear in the field of view, perhaps as a result of a macular degeneration. The mentioning of albūgō 29 Either exhibiting the so-called "Hoffmann suffix" or the substantivizing/individualizing n-suffix attached to a possessive adjective *rubicus 'red ( inis f. 'a white opaque spot in the eye' in the following sentence strengthens the view that culices has to be interpreted as something like 'dark spots' in the present case. For want of other obvious instances of this word, however, one usually regards culices as the plural of culex m. (f.) 'gnat, midge, or a similar insect' [Plaut.+]. The evident difference in meaning has either been tentatively bridged via the assumption of a metonymical use of the word 'gnats' in this case (e.g. "culices (sc. imaginarios)" according to the tll s.v. culex; the old s.v. culex recognizes a special meaning "Moving spots appearing before the eyes" exclusively for this attestation), or one has simply taken it literally and translated "gnats are removed"32 or even "small flies [sc. may be] removed which have got into the eye".33 Apart from the fact that one might easily think of better methods for removing gnats that have gotten into one's eye than using the juice of a plant (why not use a finger?), both strategies seem difficult to justify from the text itself and may be regarded as obsolete in view of the possibility that there might have been two homophonous words culex.34 In any case, we might as well want to question the commonly accepted35 etymology of culex 'gnat' and its Celtic equivalents (i-stem OIr. cuil 'fly' , Welsh pl. cylion 'flies' , OBret. colaenn, Corn. kelionen [singulative]) as related to Ved. śula-m.n. 'spit, skewer' [rv+]. This connection is semantically attractive-at first sight at least-since culices are sometimes described as stinging insects. Cf. Hor. Sat. 1.5.14: mali culices ranaeque palustres auertunt somnos "evil mosquitos and the frogs in the swamps avert my dreams", or Eug. Tolet. Carm. 38.3: mordaces culices "biting mosquitos".
Morphologically, one would have to accept the following scenario: As is clear from Indo-Iranian, Ved. śula-has to be segmented śu-la-and belongs to Ved. *śū-ci->> sūci-f. 'needle EWAia ii:739). Both iel reviewers very kindly point out to me that eye floaters (threadlike specks in the visual field) are commonly known as "mouches volantes" ("flying flies") in several languages, so basically the idea of "culices (imaginarios)" does after all remain a serious possibility. actual basis of śula-, culex, PCelt. *kuli-, and perhaps also Arm. slakʿ 'arrow, javelin, pike' ,36 would then have to be an otherwise unattested adjective *kūlo-'pointy, spiky' . This is, of course, not impossible, but proves somewhat more difficult on the phonological side, for one would have to account for the short ŭ of Latin and Celtic as opposed to the long ū of Vedic. One could certainly invoke shortening of the ū in accordance with Dybo's Law,37 but since there is no external evidence for the position of the accent in the predecessors of Lat. culex and OIr. cuil38 and the operation of Dybo's Law in Italic and/or Celtic is quite disputed,39 we might as well remain doubtful about this etymology altogether.
On the other hand, assuming a connection with *kȗl-o-'dark, black' seems gratuitous at first sight as well, since it is only idle conjecture that the gnat or the fly was named after it being dark or black. However, coincidentally a similar formation can perhaps be seen in another word for an insect, i.e. Lat. cīmex icis m. 'bed-bug' [Liv. Andr.+], which seems40 to be a derivative exhibiting the same velar suffix from an underlying color adjective *kȋ̯ eh1-mó-'dark' that can be reconstructed on the basis of Vedic śyāmá-'dark gray, black' , Lith. šeṁas, šeṁas 'bluish gray' (cf. also *kȋ̯ eh1-u̯ ó-above under 2).41 Note that the appearance of *kȋ̯ eh1-mó-in the zero grade *kȋh1-mo-(→ cīmex) is exactly in line with what we assumed above for *rubex, *uirex, and, for that matter, *culex.
Further support for this interpretation might be obtained if the Greek adjective πελλός 'dark-colored, dusky (mostly of animals)' (or πέλλος) [S., Arist., Theoc.] could be regarded as the "Aeolic" continuant of a full-grade *-noadjective *kȗ̯ el-nó-43 that would bear relation to substantivized zero-grade *kȗ̯ l ̥nó-in the same way as the above-mentioned περκνός 'dark, dusky' does to 42 Note, however, that one could hypothesize a pre-einzelsprachlich i-stem *kȗli-'the black one' > 'gnat, fly' (directly attested by the continuants of PCelt. *kuli-'id.') which was enlarged only within Latin by the velar suffix as *kȗli-k-'id.' (parallel to *pusli->> *puslik-'flea' , see previous note On the other hand, there are also arguments in favor of the second possibility, viz. *kȗ̯ l ̥só-, which would have to be interpreted as a thematic deriva-tive44 from an s-stem basis. In principle, derivatives of this kind are widespread among virtually all the Indo-European languages, though often well-hidden as obscure or unexplained formations due to einzelsprachlich phonetic developments that make an immediate identification quite difficult. The idea of thematic derivatives from s-stem bases with a double zero grade in the root and the suffix of the base word goes back to J. Schmid and P. Persson, who as early as 1893 both published their very similar analysis of Lat. russus 'red, red-haired' [Enn.+] as a derivative of an s-stem attested in Gk. ἔρευθος n. 'redness' (cf. Schmid 1893:387 and Persson 1893:270), in a modernized fashion: *h1réu̯ dh-es-'redness' → *h1rudh-s-ó-'having redness' . For the sake of convenience, and because russus is actually a very iconic example, I would like to use the term "russus adjectives" as some kind of abbreviation for "thematic possessive derivatives from neuter s-stems with a double zero grade in root and suffix of the base word".
An interpretation along these lines for Arm. šaɫ* and Gk. πηλός could, of course, only be granted a certain amount of cogency if we were able to establish 44 For possessive adjectives with the suffix *-ó-cf. Schindler 1984;Peters 1999:453;Widmer 2004:32ff.;Meier-Brügger 2006:120;Nussbaum 2014b:245. 45 Since possessive derivatives of verbal nouns could have either an active or passive meaning (cf. Nussbaum 2014a:244), the same adjective *s ǝ k-s-o-'having cutting' might be continued by Gmc. *sahsa-'knife' < 'having cutting = the cutter' as well. Note that for *sahsaan o-grade preform *sok-s-o-is equally possible. ] 'color' is a seemingly isolated formation within Latin. It is therefore conceivably one of the cases where a masculine s-stem in -ōs, younger -or, -ōris continues an inherited formation, namely an internally derived masculine from an underlying neuter s-stem, as opposed to the vast majority of stems in -or, -ōris that were productively formed within an inner-Latin Caland system alongside verbs in -ēre and -ēscere, and adjectives in -idus (cf. Rau 2009:124 f.;Stüber 2002:57ff.). The commonly accepted etymology46 of Lat. color connects it to the root *√kȇl 'to hide, conceal' . This is phonologically possible, and one usually refers to Ved. várṇa-'color' from the root √var 'to cover' (via:235 f.) for the semantics. Admittedly, this account is not impossible, but aside from the fact that an sstem *kȇl-os n. is not securely attested anywhere (cf. Stüber loc.cit.), it might be regarded as an additional obstacle that its presumable semantics should have been something like 'hiding, concealment' , which would at first have had to develop into 'something that provides concealment in the form of a covering' , whence 'cover' and then ultimately 'color' .
Indo-European Linguistics 3 (2015) 24-41 The alleged basic meaning of color as 'dark color' might perhaps still be seen in the denominative verb colōrāre, which basically means 'to give something a darker color' as in Cato's De agri cultura in a description of how to make a certain kind of pastry: item unguito coloratoque caldum ne nimium (Cato Agr. 80) "spread it [sc. the cake] in the same way (with honey) and let it become brown, but at a moderate heat." Though, of course, an interpretation as 'to give something a different color' is possible here as well.
Therefore, an original meaning 'dark color' of color is all the more plausible, and a derivation from the proposed root *√kȗ̯ el 'dark, black' might, after all, be the preferable option.