*h₂óns-/h₂n̥s-´

The Hittite verb ḫāš-/ḫašš-^(ḫi) ‘to procreate, give birth’ can be directly reconstructed as a PAnat. root verb *h₂óns-/h₂n̥s-´ of the ḫi-conjugation (pace EDHIL:319f. *h₂éms-, HED:217; cf. Kimball 1999a:331 with *h_2/3Ems-, although *h₃ is made less likely by the retention of the laryngeal in the compounded Hittite names in -ḫšu-See under PAnat. *h₂énsu- and cf. Melchert 1994a:164 and Melchert 2012c:174 for the unconvincing view of seeing two roots *h₂ens- and *h₂ems- with the same semantics, where only *h₂ is retained beside a neighbouring /-s-/.). The HLuw. verb attested as 3pl. impv. /has(s)antu/ is taken by default as a root verb of the ḫi-conjugation with expected theme vowel, although its synchronic verbal stem class is not assured due to the lack of unambiguous forms. Likewise, it is unclear whether Luwian would have generalized the strong stem with a single /-s-/ (remodelled after the ḫi-verbs of the corresponding Hittite class in -āC₁i, -aC₁C₁anzi) or the weak stem with the geminate /-ss-/, although the assimilation of the *-ns- > /-s(s)-/ sequence speaks a priori for a generalization of the strong stem; for the rules, see PAnat. *h₂énsu- as well as under the PIE lemma.

For a summary of the largely outdated literature on Indo-European etymologies of the root, see HEG H:192-194.

[D.S.]

The verbs Hittite ḫāš-/ḫašš-^(ḫi) ‘to give birth, beget, procreate’ and Luwian /has(s)a-⁽ⁱ⁾/, ha-sa- ‘to give birth, beget’ suggest a basic root meaning ‘to generate, bring into being’. Anatolian is the only branch that preserves a verb derived from this root. In Core Indo-European, it seems to have been replaced by the semantically synonymous *g̑enh₁- except for in a few lexicalized nominal formations. Because of the parallel semantic profile of these two roots, derivatives of *g̑enh₁- will be used throughout the lemma to illustrate the morphosemantic prehistory of the formations under discussion.

A basic meaning ‘to generate, bring into being’ aligns well with the words for ‘descendant, offspring, grandchild’ in Anatolian. A little bit more obscure is the path that led to Hittite ḫaššu- ‘king’, Avestan ahu-, aŋhu- m. ‘master’, and Proto-Germanic *ansuz ‘god, deity’. One possibility is to separate the u-stems and explain ḫaššu- ‘king’ via the meaning of the verb ‘to procreate’, which has a striking parallel in the word king itself, from Proto-Germanic *kuningaz. As pointed out in EDHIL:328 s.v. ḫaššu-, this word ultimately goes back to *g̑enh₁- ‘to beget, procreate’, probably as a derivative in *-inga- ‘belonging to, descended from’ from *kunja- ‘kin, family’ (cf. EDPG:311). The exact semantic motivation behind it is unclear, however. Was the *kuningaz originally someone ‘descended from [sc. noble] kin’, or a more modest ‘leader of a clan’ (cf. Gothic kindins ‘governor’, derivative of *kinþiz ‘kind, race’ < *g̑enh₁-ti-cf. also Gothic þiudans ‘king’, derivative of þiuda ‘people’)? In any event, the morphosemantic prehistory of Hittite ḫaššu- is probably not easily comparable with that of *kuningaz, as a seemingly primary u-stem does not normally carry the meanings necessary for either of the two interpretations of the Proto-Germanic word just presented.

There are three remaining options to reconcile the evidence. One way would be to depart from a meaning ‘(grand)son’ for PAnat. *h₂ensu- (cf. the compound names with -ḫšu- as well as CLuw. ḫamšu-kkalla- c. ‘great-grandson’, Lycian xahba- c. ‘grandchild’) and assume that the son of the king and heir apparent was referred to by *h₂ensu- as a title (quasi the ‘son par excellence’) and that he then carried this title over when he became king himself, ultimately resulting in a semantic shift from ‘son’ to ‘king’. Another way would be to assume an original meaning of an agent noun ‘(sc. great) begetter’ for ḫaššu- (compare the 30 sons and 30 daughters of the Queen of Zalpa), though it seems strange that the u-stem could mean both ‘begetter’ (> ḫaššu- ‘king’) and ‘thing begotten’ (> PAnat. *h₂ensu- ‘offspring, (grand)son’).

The more convincing account might be to see ḫaššu- ‘king’ as a direct cognate of Avestan ahu-, aŋhu- m. ‘master’ and Proto-Germanic *ansuz ‘god, deity’ that originally meant ‘superior being (vel sim.)’. The u-stem could be seen as a possessive internal derivative of another u-stem from the same root, perhaps attested through Vedic ásu- m. ‘life’, Avestan ahu-, aŋhu- m. ‘life’. While the precise semantics must remain open for now (but see below), the morphosemantic side of this account is corroborated by the fact that the underlying u-stem was seemingly also the basis of an external possessive derivative by means of *-ro-, attested in Vedic ásura- ‘divine, spiritual’, m. ‘good spirit; evil spirit; chief of the evil spirits’ and Avestan ahura- m. ‘god, master’, which come conspicuously close in meaning to Hittite ḫaššu- ‘king’, Avestan ahu-, aŋhu- m. ‘master’, and Proto-Germanic *ansuz ‘god, deity’.

The root is variously reconstructed as *h₂ens- (e.g., LIV²:269) or *h₂ems- (e.g., EDHIL:319). The reconstruction of *n or *m largely hinges on the precise interpretation of the Anatolian sound laws involving sequences of (accented or unaccented) vowel plus *m or *n as well as syllabic *m̥ or *n̥ before *s. The presence of /-ms-/ in the Luwian cognates seems to speak in favor of *m as the root-internal nasal. It should be noted, however, that Gothic keeps inherited *-ns- and *-ms- apart (compare, e.g., acc. pl. *h₁om(h_x)sōns > Goth. amsans ‘shoulders’). If Jordanes’s rendering of the Gothic word for half-gods, ansis, is to be trusted – he is said to have had Gothic ancestry himself – and does not constitute a concession to Latin phonology (where /ms/ was disallowed), this would support the reconstruction of the root as *h₂ens-.

It is relatively undisputed that both *-m̥s- and *-n̥s- gave *-anz- in Hittite and Luwian (cf. *h₂m̥s(o)no- > Hitt. ḫanzana- ‘black’; *n̥s- > Hitt. anz-, Luw. /anza-/ ‘we; us’; probably *h₂n̥s-osi̯o- > Hitt. ḫanzāšša- ‘offspring’see also Melchert 1994a:121). Posttonic *-V́ms- seems to show partial assimilation to *-V́ns- in Hittite (cf. *h₁ó/áms- > Hitt. ānš-^(ḫi) (and ānš-^(mi)) ‘to wipe’; *h₁óm(h_x)so- ‘shoulder’ > Hitt. anašša- ‘part of the back’?see also Melchert 1994a:164), but stays (or becomes?) *-V́ms- in Luwian (cf. CLuw. āmmašš-^(ti) ‘to wipe’see also Melchert 1994a:269) and Palaic (cf. Pal. āmša-⁽ⁱ⁾ ‘to wipe’see also Melchert 1994a:221). This seems to preclude the possibility that Hittite ḫā̆š(š)- in the cognate set under discussion (i.e., ḫāš-/ḫašš-^(ḫi) ‘to give birth, beget, procreate’, ḫāšša- c. ‘descendant’, ḫaššu- ‘king’) goes back to *h₂(ó/é)ms-. Instead, the full assimilation, mirrored by Luwian /has(s)a-⁽ⁱ⁾/, HLuw. (NEPOS)ha-su- ‘family’, Lycian xahba- c. ‘grand-child’, is better seen as the regular outcome of *-V́ns-see also Melchert 1994a:135, 163, 270. The *-m- in Luwian *-ms- in /hams(i)-/ ‘grandson’, Luw. /hamsukkalla-/ c. ‘great-grandson’ could then be seen as the product of a secondary dissimilation of pretonic *-nsV́-. This implies that CLuw. ḫamš(i)- ‘grandchild’ and HLuw. /hams(i)-/ ‘grandson’ continue *h₂onsó-, a substantivized τομός-type noun meaning ‘thing begotten; child, descendant’ or the weak stem of a root noun *h₂ens-´ ‘thing begotten; child, descendant’, while Hittite ḫāšša- c. ‘progeny, issue, offspring, descendant’ goes back to a τόμος-type noun *h₂ónso- ‘offspring’, parallel to Greek γόνος ‘birth, child, offspring’ (Il.+) < *g̑ónh₁o- (see Nussbaum 2017a:235). On /hamsukkalla-/ c. ‘great-grandson’, see below.

In line with what was said above at the end of Semantic Reconstruction, the morphological prehistory of Hitt. ḫaššu- ‘king’ might be elucidated as follows: an acrostatic u-stem *h₂ónsu-, *h₂énsu- served as the basis of an internally derived proterokinetic derivative with possessive meaning, *h₂énsu-, *h₂n̥séu̯- ‘having *h₂ónsu-’, parallel to examples proposed by, e.g., Nussbaum 1998b:147. This would have been substantivized to ‘superior being (vel sim.)’ already in PIE. Hittite ḫaššu- ‘king’ and Germanic *ansuz ‘deity’ continue its full grade, while Avestan ahu-, aŋhu- m. ‘master’ (OAv. nom. pl. aŋhauuas(cā)) preserves the zero grade. This scenario works in theory, but it is difficult to see the acrostatic derivational base *h₂ónsu-, *h₂énsu- in Vedic ásu- m. ‘life’ and Avestan ahu-, aŋhu- m. ‘life’, as these could only continue a zero-grade *h₂n̥s-u-. Perhaps it is best to separate Vedic ásu-, Avestan ahu-, aŋhu- (qua *h₁ésu- ‘existence’ from *h₁es- ‘to be’?) from Av. ahu-, aŋhu- m. ‘master’, Hitt. ḫaššu-, PGmc. *ansuz, and Ved. ásura-, Av. ahura- (thus EWAia I:147f. s.v. ásura-). For these, a derivational base *h₂ónsu-, *h₂énsu- ‘generative power (vel sim.)’ remains an option, with Hitt. ḫaššu-, Av. ahu-, aŋhu-, PGmc. *ansuz continuing an internally derived *h₂énsu-, *h₂n̥séu̯- ‘having generative power (vel sim.)’, and Ved. ásura-, Av. ahura- going back to an externally derived *h₂n̥su-ró- of the same meaning. For a semantic parallel from *g̑enh₁- compare Latin genius ‘the tutelar deity or genius (of a person, place, etc.)’, interpreted by NIL:150 note 68 as the personification of the generative power “Ursprünglich die personifizierte Zeugungskraft”.

The other apparent u-stem attested in HLuw. (NEPOS)ha-su- ‘family’, and Lycian xahba- c. ‘grand-child’, presupposes a semantically different u-stem, namely with a passive-resultative meaning ‘thing begotten (vel sim.)’. Since it is somewhat difficult to imagine that a word could mean both ‘king’ and ‘(non-royal) offspring’ at the same time, one might entertain the idea that the latter is actually the continuant of the acrostatic stem *h₂ónsu-, *h₂énsu- ‘generative power (vel sim.)’ reconstructed above, but in a concretized sense, comparable to Greek γονή which means ‘act of generation’ (Pi., Aeschin.), ‘that which engenders, seed’ (Hes., Hdt., Pi., Hp.), and ‘offspring, family’ (Il., Od., A., S., E., etc.). CLuw. ḫamšukkalla- c. ‘great-grandchild’, HLuw. /hamsukkalla-/ c. ‘great-grandson’ might then reflect the unaccented allomorph (quasi *h₂ensu-ké-leh₂-). At least for the a-stem Lycian xahba- c. ‘grand-child’, one could also envisage a substantivization *h₂énsu̯eh₂- based on an underlying verbal adjective *h₂n̥s-u̯o- ‘begotten’, formally comparable with Ved. pak-vá- ‘cooked’ or Lat. (g)nāvus ‘active, busy, diligent’ (*g̑n̥h_1/3-u̯ó-; see NIL:150f. note 84).Cf. Lyc. A xahba- under Stem for an alternative.

[S.H.]

Hittite ḫāš-/ḫašš-^(ḫi) ‘to give birth, beget, procreate’, ḫāšša- c. ‘descendant’, ḫanzāšša- c. ‘offspring’, ḫaššu- ‘king’, Hieroglyphic Luwian /has(s)a-⁽ⁱ⁾/, ha-sa- ‘to give birth, beget’, Cuneiform Luwian ḫamš(i)- ‘grandchild’, Hieroglyphic Luwian /hams(i)-/ ‘grandson’, Cuneiform Luwian ḫamšukkalla- c. ‘great-grandchild’, Hieroglyphic Luwian /hamsukkalla-/ c. ‘great-grandson’, HLuw. (NEPOS)ha-su- ‘family’, Lycian xahba- c. ‘grand-child’.See above, and EDHIL:319–321 s.v. ḫāš-ⁱ / ḫašš-, 323f. s.v. ḫāšša-, 293 s.v. ḫanzāšša-, 327f. s.v. ḫaššu-.

Vedic ásu- m. ‘life’, ásura- ‘divine, spiritual’, m. ‘good spirit; evil spirit; chief of the evil spirits’.Cf., e.g., EWAia I:147 s.v. ásu-, 147f. s.v. ásura-

(Old and Young) Avestan ahu-, aŋhu- m. ‘life’, ahu-, aŋhu- m. ‘master’, ahura- m. ‘god, master’.Cf., e.g., EWAia I:147 s.v. ásu-, 147f. s.v. ásura-

Proto-Germanic *ansuz ‘god, deity’ (Old Norse áss, óss, pl. æsir, Old English ōs, Old Saxon ās, ōs, Gothic [in Late Latin transmission] ansis ‘half-gods’ Jordanes Getica 13: semideos, id est ansis vocaverunt ‘they called them half-gods, that is: ansis’)Cf., e.g., EDPG:30 s.v. *ansu-