*d^hég̑^h-om-/*d^hg̑^h-m-´/*d^hg̑^h-ém

The etymological relationship between Hitt. tēkan-/takn- n. ‘earth’, CLuw. tiyamm(i)- c. ‘earth’ and HLuw. dat. sg. ta-ka-mi-i has been commonly accepted in the literature (cf. EDHIL:858-862, HEG T,D:292-300 with further literature). Between the Hittite and Luwian cognates, there is a disagreement vis-à-vis the gender of the noun in question. From a Luwian perspective, we expect the word for ‘earth’ to be originally common gender as seen in CLuw. tiyamm(i)- c., because Luwian neuter stems do not undergo i-mutation. Thus, the neuter gender of the Hittite noun is to be explained as secondary due to the identical shape of the nominative and accusative te-e-kán and their similarity with neuter n-stems (Melchert 2003c:150, EDHIL:860).

The pre-form of the Cuneiform Luwian word tiyamm(i)- c. can be best explained as follows. The vowel -a- in front of the geminate -m- must reflect an accented *é, which would have triggered Čop’s Law and geminated the nasal to -mm-. The sequence -iy- arose from a voiced velar, which was regularly lost in Luwian yielding a glide -y- before front vowels. Therefore, the pre-form must contain at least the following elements *T(E)ĝ^(h)ém- (Kimball 1994c:78, Melchert 2003c:151). Differently, Steer 2013b:65f. suggests that the pre-form was *d^hég̑^hom assuming a development of the sequence *-ég̑^(h)o- > /-iya-/ and an analogical gemination based on the locative form. However, this is contradicted by HLuw. ta-ka-mi-i as discussed below, which demonstrates that the sequence *-ég̑^(h)o- must trigger Čop’s Law, before *ĝ turns into a glide. Moreover, the only parallel adduced for this sound change is anninniyam(i)- ‘female cousin’, whose morphological interpretation, though attractive, is not assured.

Regarding the Hieroglyphic Luwian word ta-ka-mi-i, Čop 1970a:91 argues that the preservation of the velar is due to the sound law of *-éC- > *-éCC- (named Čop’s Law), which would have rendered the voiced velar into a fortis velar. Thus, he interprets HLuw. dat. sg. ta-ka-mi-i phonologically as /takkami/ (followed by Melchert 2003c:150; not accepted in EDHIL:861, because Kloekhorst does not expect the paradigm for ‘earth’ to have had a full accented e-grade in the root with an o-grade in the suffix such as *d^hég̑^h-om). This is the most persuasive explanation. Pace EDHIL ibid. and Steer 2013b:67, it is not convincing to derive Luw. ta-ka-mi-i from *d^hg̑^hṓmi, because there is no evidence in Luwian for the development of an initial cluster *#d^hg̑^h- to /tag-/, nor for the presence of a synchronic phoneme /-g-/ (< *-g-/-g̑-) either in Luwian or in Lycian.

Regarding the univerbated CLuw. īnzagan-, Melchert 2003c suggests convincingly that it goes back to a cluster with an initial thorn, i.e. *ÞK-, realized in Luwian as a dental fricative, which would have devoiced the following stop, i.e. [intskan-]. The original reflex of the thorn cluster would have been preserved here due to the univerbation, in consequence of which it would no longer have been in word initial position. However, the meaning of the Cuneiform Luwian lexeme īnzagan- is debated; see under lemma.

It is clear from the Luwian synchronic data that we are facing an original ablauting paradigm, in which one of the stem variants has been generalized. The dialect of Kizzuwatna Luwian has generalized one stem variant ending phonologically in tiyamm(i)- c., while the Empire Luwian dialect has generalized another yielding phonologically /takkami/ (ta-ka-mi-i). Therefore, we can posit that the Proto-Luwian stage still had an ablauting paradigm with an accented root vowel *Tákkam- in the strong stem and with the accent on the second syllable at least in the locative case, i.e. *Tii̯ámm-.

If we consider the Hittite word for ‘earth’, we see an ablauting paradigm between the forms of the strong cases, i.e. nom./acc. sg. n. te-e-kán, the ones of the weak cases, i.e. gen. sg. ták-na-a-aš, dat./loc. sg. ták-ni-i, and the directional adverb ta-ga-a-an ‘onto/into the earth/ground’ (wrongly identified as a locative in the literature; see Melchert 2003c:151). Without drawing data from the other non-Anatolian Indo-European languages, we can compare Hitt. te-e-kán with HLuw. ta-ka-mi-i /takkami/ and reconstruct a pre-form with an accented e-grade in the root, which would account for the shape of both forms (Melchert ibid., contra EDHIL:861). Overall, we can reconstruct a Proto-Anatolian paradigm of the following shape: for the strong cases *d^héĝ^h-om- (differently Steer 2013b:71, who suggests a secondary anaptyptic vowel with retraction of the accent), for the weak cases *d^hg̑^h-m-´ and a loc. sg. *d^hg̑^h-ém.

The reconstruction offered here is from a Hittite and Luwian perspective only.

[D.S.]

PIE *d^hég̑^hom- means ‘earth’, cf. EIEC:174, NIL:86.

The PIE word for ‘earth’ can be reconstructed as an amphikinetic feminine m-stem noun: nom. sg. *d^hég̑^h-ōm (< */d^hég̑^h-om-s/), acc. sg. *d^hég̑^h-ōm (< */d^hég̑^h-om-m/), gen. sg. *d^hg̑^h-m-és, dat. sg. *d^hg̑^h-m-éi̯, loc. sg. *d^hg̑^h-ém(-i), etc. (NIL:88f. n. 1 with lit.).For an alternative reconstruction of the original paradigm, cf. Beekes 1985a:45f. and EDHIL:859: nom. sg. *d^hég̑^h‑m̥, acc. sg. *d^hg̑^h-ém-m, gen. sg. *d^hg̑^h-m-ós.For a similar m-stem noun, cf. perhaps PIE *g̑^héi̯-om-/*g̑^hi-m-´, loc. sg. *g̑^hi̯-ém(-i) ‘winter’ (Friedman 2003a, NIL:162–169, Steer 2013b:73–87). In the case of PIE *d^hég̑^h-om-/*d^hg̑^h-m-´, the reconstruction of the amphikinetic ablaut pattern is primarily based on the Hittite and Luwian forms: e.g., Hitt. nom./acc. sg. n. tēkan, gen. sg. taknāš, dat./loc. sg. taknī (cf. above sub PAnat. Reconstruction). These forms seem to preserve the ablaut and accent shift of the inherited paradigm. The nom./acc. sg. Hitt. tēkan can thus regularly go back to PIE *d^hég̑^h-ōm.Differently Steer 2013b, who discusses the possibility of explaining the e-vowel in Hitt. tēkan as an inner-Anatolian anaptyxis. In the weak cases, an anaptyctic vowel (schwa secundum) was inserted into the word-initial consonant cluster *d^hg̑^h-m- in the prehistory of Anatolian, with gen. sg. *d^h_eg̑^h-m-és, dat. sg. *d^h_eg̑^h-m-éi̯ > Hitt. gen. sg. taknāš [tagnā́s], dat./loc. sg. taknī [tagnī́] (cf. Mayrhofer 1986:152, NIL:89 n. 2). The non-Anatolian branches of IE seem to reflect an innovated late-IE paradigm with generalization of the zero-grade root. The resulting strong stem *d^hg̑^h-om- and weak stem *d^hg̑^h-m- first showed metathesis to *g̑^hđ^h-om- and *g̑^hđ^h-m-, with an allophone [đ^h] arising from /d^h/ in this position.According to Mayrhofer 1986:152 (following Schindler 1977a), the simplification of the word-initial group *d^hg̑^h- occurred in Sievers-Lindeman variants of weak case forms such as gen. sg. *d^hg̑^hm̥-és > *g̑^hm̥-és, in which a rule #TK > #K / __N̥ was valid (cf. PIE *dk̑m̥tóm > *k̑m̥tóm ‘100’). But syllabifications like *jam- in weak case forms of the Vedic paradigm are not attested (NIL:90 n. 6).This allophone [đ^h] regularly disappeared in the weak cases: *g̑^hđ^h-m- > *g̑^h-m- (cf. Lühr 2000a:306, Klingenschmitt [personal communication]). If this is correct, we have to reconstruct the following late-IE paradigm: nom. sg. *g̑^hđ^h-ṓm (< *d^hg̑^h-ṓm), acc. sg. *g̑^hđ^h-ṓm (< *d^hg̑^h-ṓm), gen. sg. *g̑^h-m-és (< *g̑^hđ^h-m-és < *d^hg̑^h-m-és); cf. nom. sg. Gk. χϑών < *g̑^hđ^hṓm (Ved. kṣā́ḥ, Av. zā̊, PCelt. *(g)dū, Alb. dhē < *g̑^hđ^hṓ-s are [independent] analogical formations on the basis of the reanalyzed acc. sg. *g̑^hđ^hṓ-m), acc. sg. Ved. kṣā́m, Av. ząm (also PCelt. *(g)dōm, cf. Ringe 1996a:42 n. 1) < *g̑^hđ^hṓm, gen. sg. Av. zǝmō, Ved. jmás (also gmás with dissimilation gm° ← jm° in the formula diváś ca gmáś caCf. EWAia I:425.) < *g̑^hmés.

For poetic formulas in Indo-Iranian and Greek (Ved. kṣā́m … pr̥thivī́m, YAv. ząm pǝrǝϑβīm ‘broad earth’ and Ved. kṣā́m … urvī́m ~ Gk. εὐρεῖα χϑών), cf. Schmitt 1967a:181f., KEWA III:686, with older lit.

[S.S.], [T.S.]

Ved. kṣám- f. ‘earth; soil’ (RV[+]): e.g., nom. sg. kṣā́ḥ, acc. sg. kṣā́m, abl. sg. kṣmáḥ, abl./gen. sg. gmáḥ, jmáḥ, instr. sg. jmā́, loc. sg. kṣámi, kṣā́man(i), kṣmayā́, jmán; nom./acc. du. kṣā́mā; acc. pl. kṣā́ḥ, loc. pl. kṣā́su. – Av. zam- f. ‘earth’: nom. sg. YAv. zā̊, acc. sg. OAv. ząm, YAv. ząm, instr. sg. YAv. zǝmā, gen. sg. YAv. zǝmō, zǝmas-ca, etc.Cf., e.g., EWAia I:424f., ALF:140f., NIL, esp. 87, 90f. nn. 6–7, Gotō 2013a:18f.

Gk. χϑών, gen. sg. χϑονός f. ‘earth, ground, land’ (Il.+), χϑαμαλός adj. ‘near the ground, on the ground, low’ (Il.+), χαμαί adv. ‘on the ground, to earth’ (Il.+), etc.Cf., e.g., GEW II:1071 s.v. χαμαί, 1098f. s.v. χθών, EDG:1612f. s.v. χαμαί, 1632f. s.v. χθών, NIL:87, 91 n. 8 and (for χ[ϑ]αμ°) 87f., 94 n. 30, 97 n. 43.

Lat. humus f. ‘earth, ground’ (Plaut.+; loc. sg. humī)Cf., e.g., WH II:664f., EDLIL:292, NIL:87, 94f. n. 31, 95 n. 33.

OIr. dú f.^? ‘earth, place’ (gen./acc. sg. don, dat. sg. dú, don)Cf., e.g., Schumacher 2000a:32 with n. 13, NIL:87, 91 n. 9, EDPC:155f.

OPr. semmē, same f. ‘earth, land’, Lith. žẽmė f. ‘id.’, Latv. zeme f. ‘id.’Cf., e.g., LEW II:1299, PKEŽ 2013 IV:58–60, NIL:88, 96 n. 41, ALEW II:1296–1298, EDBIL:516.

OCS zemlja f. ‘earth, land’, Russ. zemljá, SCr. zèmlja, etc.Cf., e.g., Vasmer III:93, EDSIL:542, NIL:88, 96 n. 41.

Alb. dhē m. ‘earth, soil, land’ (pl. Tosk dhéra, Gheg dhéna)Cf., e.g., Demiraj 1997a:155f., AED:80f., Matzinger 2006a:179, NIL:87, 91 n. 11.

TochA tkaṃ, TochB keṃ ‘earth, ground’Cf., e.g., Ringe 1996a:42 n. 1, NIL:87, 91 n. 12, Adams 2013a:205.