*h₁ē̆́sHr̥/*h₁(e)sH(é)n(´)

The etymological relationship between Hittite ēšḫar/išḫan- n. ‘blood’ and CLuw. āšhar- n. ‘blood’ is generally accepted (EDHIL:256-260, HEG A-K:112-115, HED A, E-I:305-313). The Palaic word for ‘blood’ ešḫur-/ēšḫa(n)-, whose etymological relationship was suggested by Carruba 1970a:52 (followed by Melchert 1994a:196 et passim) is added here.
The Hittite word for ‘blood’ shows a r/n-heteroclitic ablauting paradigm between the strong and weak cases. all possible reconstructions from a phonological perspective regarding ablaut will be given here; i.e. nom./acc. sg. e-eš-ḫar with full accented grade of the root (< *ésHr̥ or *ḗsHr̥) and gen. sg. išḫanāš with either zero-grade of the root and accent on the ending (< *sHanós) or unaccented full grade (< *esHanós) (cf. Rieken 1999a:302-304). The form iš-ḫa-a-na-aš with plene spelling of the suffix has been argued to be either a late analogical formation after other acrokinetic types or an anaptyptic vowel. Kloekhorst (EDHIL:260), on the other hand, takes this form to be old.
The Luwian word for ‘blood’ āšḫar shows an accented full grade of the root (< *ésHr̥). A lengthened grade is impossible because *ē > Luw. i. The oblique form *āšḫan- is preserved in the derivative āšḫanuwant(i)- ‘bloody, having blood’, in which the root ablaut agrees with the one of the strong cases (Rieken 1999a:303f.).
The Palaic cognate has the form ešḫur in the compound ḫinapešḫur, which would go back to a pre-form *ḗsHr̥ or *ésHr̥, in which *r̥ vocalizes to ur (Melchert 1994a:196, 201). The collective form e-eš-ḫa would go back to *ésHōr (Melchert ibid.). An oblique case form e-eš-ḫa-n[a(-x)] is also found showing the n-suffix and the same kind of vowel grade of the root as in the forms of the strong cases. Palaic has clearly generalized the vocalism of the root.
Regarding the type of ablauting paradigm to be reconstructed for Proto-Anatolian, Melchert 1984a:92 reconstructs a lengthened grade of the root for the strong stems (followed by Rieken 1999a:302), which is motivated by the Proto-Indo-European perspective. Kloekhorst (EDHIL:259f.) disagrees and prefers to reconstruct a full e-grade for the strong stems. He points out that this reconstruction squares better with the Luwian ā (< *é) in āšḫar-. Rieken 1999a:303 takes the Luwian vocalism from the weak stems instead. In the end, Kloekhorst (ibid.) reconstructs a proterokinetic paradigm *h₁ésh₂r̥ / *h₁esh₂n´ against the traditional interpretation of an acrokinetic type *h₁ḗsh₂r̥ / *h₁(e)sh₂n-´ (Rieken 1999a:302).
Based on the Anatolian languages only, we can state that there is no compelling argument for reconstructing a lengthened grade, but it is entirely possible. The Palaic collective form ēšḫa, which was often ignored in the literature, should be taken into consideration. The presence of the initial first laryngeal has been argued based on the sign á in HLuw. á-sa-ha+ra/i-mi-sà (see Kloekhorst 2004a:38, Simon 2013a:6), which is still a debated issue, whence *h₁ should be placed between parentheses. Thus, we reconstruct Proto-Anatolian: *h₁ē̆́sHr̥ / *h₁(e)sH(é)n(´), coll. *h₁ésHōr.

[D.S.]

According to EIEC:71, PIE *h₁ē̆́sh₂-r̥/*h₁(e)sh₂-(e)n- n. means ‘(flowing) blood’ while PIE *kreu̯h₂-/*kruh₂- (cf. MIr. crú, W craw, OCS krŭvĭ, Lat. cruor ‘thick blood, gore’) rather denotes ‘blood (outside the body), gore’ (EIEC:71).

PIE *h₁ē̆́sh₂-r̥/*h₁(e)sh₂-(e)n- ‘blood’ is an r/n-heteroclite. Less clear, however, is which inflectional class the noun belongs to: it is usually either attributed to the ḗ/é-acrostatic inflectional class (*h₁ḗsh₂-r̥/*h₁ésh₂-n-: cf., e.g., Schindler 1975a:6) or to a kinetic ablaut class (proterokinetic *h₁ésh₂-r̥/*h₁sh₂-én-: cf., e.g., Schaffner 2001a:92, 576 n. 283; or hysterokinetic *h₁ésh₂-r̥/*h₁sh₂-n-´: cf., e.g., Balles 1999a:4).

The reconstruction of an acrostatic paradigm is especially based on Gk. ἦαρ ‘blood’ (Hsch. η 8 Latte), which might go back to *h₁ḗsh₂-r̥ and point to an inherited lengthened grade, but metrical lengthening cannot be excluded either (cf. GEW I:432). An acrostatic paradigm would account for the amphikinetic collective *h₁ésh₂-ōr (> TochA ysār, TochB yasar), which could have been derived from the weak acrostatic stem *h₁ésh₂-n-; cf., e.g., PIE coll. *u̯éd-ōr ← (acrostatic) *u̯ód-r̥/*u̯éd-n- n. ‘water’ (cf. Steer 2013a:190f.).

There are, however, good reasons for the reconstruction of a proterokinetic paradigm *h₁ésh₂-r̥/*h₁sh₂-én- as well. On the one hand, Gk. ἔαρ, Hitt. ēšḫar, CLuw. āšḫar[ša], and Ved. ásr̥-k point to a nom./acc. sg. *h₁ésh₂-r̥ (cf. Balles 1999a:4); on the other hand, a proterokinetic weak stem *h₁sh₂-én- can be reconstructed as the derivational base of forms such as Lat. saniēs ‘ulcer, wound matter’ < *h₁sh₂-en-ih₂- (as per Schaffner 2001a:576 n. 283, EDLIL:538) and perhaps also Lat. sanguis, sanguinis ‘blood’ (if san- < *h₁sh₂-en-, cf. EDLIL:537f.). The Hittite forms point to an underlying proterokinetic stem as well: “The oldest attestations show that the paradigm originally was nom.-acc.sg. e-eš-ḫar, gen.sg. iš-ḫa-na-a-aš” (EDHIL:258) and that these forms “must reflect (virtual) *h₁ésh₂r, *h₁esh₂enós, from PIE *h₁ésh₂r, *h₁sh₂éns” (EDHIL:260). There is also once a gen. sg. form iš-ḫa-a-na-aš = /ishánas/ (KUB 17.18 ii 29), which supports the assumption of an original proterokinetic weak stem /ishán-/ < *h₁sh₂-én-. Furthermore, a proterokinetic paradigm could have been the derivational base of a hysterokinetic collective *h₁sh₂-ḗr, which seems to be continued by Lat. assyr ‘blood’ (aser, ascer in glosses), which is generally interpreted as /aser/.Cf. Klingenschmitt 1992a:118, Balles 1999a:5f. with n. 8, EDLIL:58. Schrijver 1991a:29 argues that Lat. /aser/ should be left aside because of the missing rhotacism and the vacillating forms. Yet his reservations are unfounded, because assyr, aser and other forms that lack rhotacism (e.g., Lat. fēsiae, maiōsibus) and are quoted by Roman grammarians and lexicographers certainly stem from documents or inscriptions antedating the Latin rhotacism (i.e., before ca. 350 BC).

There is also the possibility that the word for ‘blood’ had both inflectional patterns, the acrostatic inflection being primary and the proterokinetic resulting from a secondary reshaping of the paradigm. As already described by Schindler 1975a:6f., acrostatic nouns often acquire a secondary kinetic inflectional pattern; cf., e.g., PIE (ó/é-acrostatic) *u̯ód-r̥/*u̯éd-n- n. ‘water’ (Hitt. u̯ātar, instr. u̯edanda < *u̯éd-n̥-° [?]) reshaped to (proterokinetic) *u̯ód-r̥/*u̯ed-én- (Hitt. u̯ātar/u̯itēn-; cf. also Steer 2013a:192). It is, therefore, possible that the word for ‘blood’ was originally an acrostatic stem *h₁ḗsh₂-r̥/*h₁ésh₂-n- (cf. Gk. ἦαρ), which was secondarily remodeled to a proterokinetic stem *h₁ē̆́sh₂-r̥/*h₁esh₂-én- (whence, e.g., Lat. saniēs, etc.). In Vedic, the paradigm ásr̥-k/asn-´ shows a hysterokinetic weak stem as-n-´ < *h₁esh₂-n-´In Ved. gen. sg. asnáḥ and instr. sg. asnā́ < pre-Ved. *h₁es[h₂]nés and *h₁es[h₂]néh₁ the laryngeal disappeared regularly in a sequence *-sHn-, according to Peters 1980a:172 n. 124, Mayrhofer 1986:150; cf., e.g., YAv. gen. pl. hazasnąm (Y 61.3, Yt 6.4) < *ség̑^hes-h₃n-ō̆m : YAv. hazaŋhan- m. ‘robber’ < *ség̑^hes-h₃on- ‘provided with violence’., which is either the result of a remodeling of a primary acrostatic weak stem as well or of a combination of the strong stem *h₁ésh₂-r̥- and the weak stem *h₁(e)sh₂-n-´ of the collective formations *h₁ésh₂-ōr/*h₁sh₂-n-´ and *h₁sh₂-ḗr/*h₁sh₂-n-´, respectively.

The exact morphological segmentation and root etymology of PIE *h₁ē̆́sh₂-r̥-/*h₁(e)sh₂-(e)n- are unknown (*h₁ē̆́sh₂-r̥- [i.e., root *h₁esh₂ + suffix -r/n-?] or *h₁ē̆́s-h₂-r̥- [root *h₁es + suffix -h₂- + suffix -r/n-?]).

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

Ved. ásr̥j-/asn-´ n. ‘blood’ (RV [nom. sg. ásr̥-k 1.64.4]+, AV [gen./abl. sg. asn-ás, instr. sg. asn-ā́]+; only ep. instr. sg. asr̥j-ā and the like); also in compounds, cf. asr̥k-pā́van- ‘blood drinking’ (AV), ásr̥ṅ-mukha- ‘having a bloody face’ (AV)Cf., e.g., EWAia I:149, with lit. 

Gk. ἔαρ beside εἶαρ (presumably with metrical lengthening), also ἦαρ (Hsch. η 8 Latte) n. ‘blood’, fig. ‘juice’ (Call., Euph., Nic. et al.Cf. LSJ s.v.); cf. also the compound εἰαρο-πότης ‘drinking blood’ (Hsch. ε 719 Latte)Cf., e.g., GEW I:432, EDG:366. 

Arm. ariwn (an-st.: gen./dat. sg. arean, abl. sg. y-aren-ē, instr. sg. aream-b, gen./dat. pl. arean-c‘) ‘blood’ (Bible+)Cf., e.g., Olsen 1999a:490f., EDAIL:138. Cf. also Balles 1999a:4f. 

Archaic Lat. assyr ‘blood’ (Paul. ex Festo; aser, ascer in glosses), generally interpreted as /aser/Cf., e.g., Schaffner 2001a:92, 576 n. 283.; cf. also assaratum ‘mixture of wine and blood’ (Paul. ex Festo)Cf., e.g., Schaffner 2001a:92, 576 n. 283, EDLIL:58, each with further lit. 

Latv. asins f. (i-st.) ‘blood’ (nom. pl. asinis) with variants (asins m. [i-st.]; asnis f. [i-st.])Cf., e.g., EDBIL:526, according to whom “[a] Lithuanian counterpart asinas seems to occur exclusively in the writings of Daukantas.” Cf. also Balles 1999a:4f. 

TochA ysār, TochB yasar ‘blood’Cf., e.g., DTB²:525, Hackstein 2017a:1310.