*skanda-

Rieken 2010b:657 connects the root /skand-/ of the verb /skandal(l)issa-/ ‘to furnish with applications, to make gleam’ etymologically with PIE *skend- ‘to gleam’. The velar k in the Luwian root /skand-/ must go back to the Proto-Indo-European velar obstruent *k, because it cannot reflect the palatal *k̂, which would have yielded Luwian -z(z)-; cf. Melchert 1994a:251f., differently Melchert 2012a with further literature. The alveolar stop after the nasal in the root /skand-/ can go back to either an original *-t- or *-d-, since both were neutralized to /d/ after n; cf. Rieken 2010a:306.

Based on the semantics and the phonological similarity of the root, the two Cuneiform Luwian lexemes šakantama/i- ‘decorated with appliqué’ and šak(k)antat(t)ar ‘appliqué’ are surely etymologically related to HLuw. /skandal(l)issa-/ (as suggested by Starke 1990a:527f.). The highly frequent spelling of the word initial sequence as ša-kán- in both Cuneiform Luwian cognates with the one exception, i.e. ša-ag-ga-an-ta-at-ta-˹ra^?˺ (KBo 13.61 rev. 5’), speaks for the interpretation of -a- in <ša> as an empty vowel and, therefore, the beginning of these words should be phonologically read as /skand-/ (pace Starke 1990a:517 n. 1910). The exception mention above can be explained as an anaptyptic vowel, i.e. /sakk-/.On anaptyxis in Hittite see Melchert 1994a:108. Since the Hittite language does not tolerate the initial cluster /sk-/, one should assume that the anaptyxis come from a Hittite speaker.The fate of an etymological *sk- cluster in Hittite is /isk-/, for which see Kimball 1999a:429-431. For the fate of the initial cluster *sk- in Luwian however, see the lemma HLuw. /skandal(l)issa-/ under 3.1 Phonological Interpretation.

Differently, Melchert (CLL:185) connects these lexemes to the root *sek- ‘to cut’, which is semantically plausible. However, both the cuneiform and hieroglyphic orthography speak in favor of an initial cluster /sk-/ independently from one another. For the hieroglyphic Luwian one see under HLuw. /skandal(l)issa-/.

All three attested lexemes belonging to this root, i.e. CLuw. šakantama/i-,  šak(k)antat(t)ar  and HLuw. /skandal(l)issa-/, are derivatives from either a verbal or nominal base */skanda-/, which can be reconstructed for Proto-Luwian. From the Anatolian perspective alone, nothing can be said about the ablaut grade of the root or about the exact nature of the derivational base, i.e. Proto-Luwian *skanda- < Proto-Anatolian *skVnd(o/eh₂)-. From the Proto-Indo-European perspective, an action noun *(s)kn̥d-ó- is posited, for which see Part 2 below.

[D.S.]

Both PLuw. *skanda- ‘(gleaming) decoration’ and PCelt. *kando- ‘bright’ descend from the deverbal adjective PIE *(s)kn̥dó- ‘gleaming’. Semantically, this association is trivial: ‘gleaming’ > PCelt. ‘bright’ (> MBret. ‘full moon’), ‘gleaming, bright thing/material’ > PLuw. ‘decoration’.For the fact, that frequently words for ‘light, full of light’ are also used for bright surfaces and the objects described as such, cf. Buck 1949a:1046–1050 s.vv. shine, bright.
On the word for ,moon‘ < *‘gleaming, shining’, cf. Syntactic Reconstruction and Phraseology.

The action noun PIE *(s)kn̥dó- ‘gleaming’ has to be reconstructed with s mobile, because in Luwian (*skanda-) it shows anlauting *s- and in Celtic (*kando-) the s-less variant is continued. In view of the Celtic root vowel *a it is recommended to reconstruct zero-grade *(s)kn̥do-, which may be the predecessor of PLuw. *skanda-, too.

All the actually attested Anatolian continuants of the root 2.*(s)kend- have to be taken as inner-Anatolian derivatives from PLuw. *skanda- ‘(gleaming, bright) decoration’.Cf. CLuw. šakantama/i-, CLuw. šak(k)antat(t)ar, HLuw. /skandal(l)issa-/. Most of the word formational morphemes of the relevant lexemes are productive in Anatolian (cf. PAnat. reconstruction) and, except for PCelt. *kando-, there are no Core IE cognates of the Anatolian material.

In three Core IE languages a word for ‘moon’ is derived from or connected with the root PIE 2.*(s)kend- ‘to (start to) gleam’: Ved. divine name candrá-mās- (m.) ‘(gleaming) moon’,Cf. also MInd.(+) canda- ‘moon’ (< *candda-) < candrá- ‘gleaming’, cf. EWAia I:529 and, for the Middle Indic phonology, Jamison 2008a:37f.MBret. cann ‘full moon’ and Alb. hãnë ‘moon’. Since these formations are morphologically not identical and, at least in Indic (epithet ‘gleaming (moon)’) and Middle Breton (< OBret. cant ‘white’), semantically secondary, such a lexeme cannot be reconstructed for the proto-language. However, the word for ‘moon’ frequently is derived from roots or lexemes meaning ‘to shine, gleam’, cf. Lat. lūna ‘moon’ < *leu̯k-s-neh₂- ‘the shining one’.Aside from that, the moon is also commonly named after its use for time measurement, viz. derived from words meaning ‘to measure’, cf. PIE *meh₁-n- vel sim. ‘moon’ from PIE *meh₁- ‘to measure (out)’ (LIV²:424f.).
For an overview of words for ,moon‘, cf. Buck 1949a:54f., EIEC:385.

[A.O.]

PCelt. *kand-o- ‘bright’, cf. W cann ‘white, bright’, OBret. cant ‘(Lat.) canus, white, hoary’, MBret. cann ‘full moon’For personal and devine names containing this etymon, cf. Delamarre 2007a:215, Delamarre 2009a.

EDPC:—, Henry 1900a:53 s.v. 1 kann

Vedic
Ved. (ś)candrá- ‘gleaming, bright; (sbst., m.) moon’ (RV+), candrá-mā- m. ,moon‘
EWAia I:529, KEWA I:373, Grassmann 1996a:1414
Ved. intens. part. nom. sg. (m.) cániścadat ‘gleaming’ (RV.5.43.4, hapax, Dhātup.)
EWAia I:528f., KEWA I:372f., Grassmann 1996a:436

(Young Avestan)
YAv. čaēčasta- ‘(name of a lake)’
Bartholomae 1979a:575This lexeme probably does not belong here, cf. EWAia I:529 with lit.

(?)Greek
Gk. κάνδαρος ‘ἄνθραξ, coal’ (Hsch. κ 644 Latte)
GEW II:776, DELG:491, EDG:634, LSJ:874Cf. also Schrijver 1991a:428.

Latin
Lat. ac-cendere ‘to set on fire, kindle (Andr.+); to make bright (Cic.+)’
EM:92, WH I:151f., EDLIL:106f., OLD:18f.
Lat. candēre ‘to shine, gleam’ (Enn.+)
EM:91f., WH I:152f., EDLIL:87, OLD:264f.

Albanian
PAlb. *skand(n)ā, cf. Alb. hãnë, Tosk hënë ‘moon’
Delamarre 2009a:23, Orel 2000a:16, AED:146f.