Binding and Smiting: One More Merism in Luvian Incantations.

• Author: Melchert, H. Craig

1. INTRODUCTION

   It is a well-known fact that few, if any, cuneiform texts were composed entirely in the Luvian language. The majority of Luvian cuneiform passages available to us belong to esoteric incantations embedded in Hittite religious compositions. Because of their arcane subject matter, discontinuous character, and the fragmentary state of preservation, they are overall less well understood than the Luvian inscriptions in Anatolian hieroglyphic transmission. Recent progress in this domain has, however, been achieved in recent years within the frame-work of the eDiAna and Luwili projects.' An approach to the decipherment of the Luvian incantations that has proved to be particularly fruitful involves the analysis of recurrent formulae, the structure of which furnishes a hint at the semantic relationship between their constituent elements. One example of this line of research is Mouton and Yakubovich 2019, where the concurrent application of combinatorial and etymological method yielded a new merism 'internal and external', used in combination with other merisms, such as 'dead and living' or 'past and future'. (2) In what follows we tackle another Luvian formula, or rather set of formulae, which can likewise occur in combination with merisms but heretofore have resisted interpretation.

   The collocations under discussion feature the forms derived from the stem sa-hu-i-t /sa[[gamma].sup.w]id-/ and the root u-i-t Avid-/, respectively. The former stem was addressed in Laroche 1959: 84, where Luv. Sahuitala- '?' and sahuitara-, supposedly 'regular, normal', are analyzed as cognates of Hitt. sakuwassara-, and Hitt./Luv. sahuihuisuwali- 'legitimate'. (3) Yet the equation of Hitt. sakuwassara- and Luv. sahuitara- was rejected as phonetically irregular in Melchert 1993: 184, while **sahuihuissuwali- was dismissed as a ghost word in Kosak 1996. Thus, for the time being, the derivatives of the stem /sa[[gamma].sup.w]id-/ lack a plausible etymology.

   The Luvian adjectives witantalli(ya/i)- and witatt(a)- are tentatively interpreted as derivatives of the word for 'water' and cognates of wida- 'watery' in Melchert 1993: 271-72. The context suggests, however, that the last lexeme is a noun occurring in a list next to 'mountains', 'rivers', and 'meadows' and thus also denoting a landscape feature (Yakubovich 2017: 11). Furthermore, it cannot be separated from the rare Hittite noun wida-, which may either represent a cognate of Luv. wida- or reflect its borrowing into Hittite. According to Kloek-horst (2008: 1015-16), this noun means 'water', but the meaning 'lake' or 'creek' would suit all the Hittite contexts cited there equally well, simultaneously drawing a bridge to the Luvian form. (4) While witantalli(ya/i)- and witatt(a)- are certainly fitting candidates for formal derivatives of the word for Take' or 'creek', this formal solution does not impose itself, since this is not the only Luvian stem that is fitting as the base of derivation.

   A common feature of the proposals addressed in the previous paragraph is the reliance on the etymological method alone. The likely reason for this state of affairs is the fragmentary character of all the contexts containing the collocation under discussion, which has precluded its consistent identification. The segments /sa[[gamma] (w)id-/ and Avid-/ are directly attested in adjacent positions only in examples (3), (6), and (7) below, which are structurally rather obscure, but even in those contexts they do not co-occur in the same lines. As for the passages (1), (2), and (4), which provide a better starting point for the analysis of the collocation, they have not been adequately restored up to now.

   In fact, our treatment of (2) can be contrasted with the approach of Carruba (1982: 46), where [ti-wa-t]a-an-ta-al-li-ya-an-za u-i-ta-an-ta-al-li-ya-an-za and [ti-wa-t]a-an-ta-an u-i-ta-at-ta-an are restored in 11. 9 and 11, respectively. Carruba realized that the two formulae contain merisms and are related to one another, but his philologically uninformed restoration led him to the translations 'of the day? and the year?' and 'the day? and the year?', respectively. (5) He frankly acknowledged the forced character of his interpretation: wett- / witt- is the Hittite word for 'year', while its Luvian counterpart is /uss(i)-/. More important, however, is the methodological lesson: one should not rush to provide etymologies of obscure forms occurring in fragmentary passages before exhausting the options of their combinatorial analysis.

   The present paper has benefited from the systematic work on restoring fragmentary Luvian incantations that has been conducted as a part of the Luwili project. Our starting point was coping with a surprisingly large number of seemingly unrelated passages where two poorly understood stems emerge as belonging to stable collocations. In what follows we shall endeavor to demonstrate that the formulae featuring the derivatives of/sa[[gamma].sup.w]id-/ and /wid-/ refer to 'binding' and 'smiting' as two distinct effects that hostile witchcraft may have on its victims. In section 2, we intend to present the full dossier of contexts that feature the relevant collocations, ordering them in a way that is conducive to restricting their meanings. Section 3 will be devoted to the etymological analysis of the relevant lexemes: /wid-/ can be linked to the well-known Luvian verbal root denoting strong physical impact, while the analysis of /sa[[gamma].sup.w]id-/ is conducive to fleshing out the Early Proto-Indo-European root *[seh2 'to bind'. In the concluding section 4, we shall argue that 'binding' and 'smiting' were indeed perceived as different functions of witchcraft in Bronze Age Anatolia, and that therefore the formulae under discussion represent true merisms.

2. CONTEXTUAL ANALYSIS

   The evidence for the merism comes from the Hittite-Luvian ritual tradition associated with the temple attendant Kuwattalla and the Old Woman Silalluhi (CTH 759-763). Two passages featuring the lexemes under discussion demonstrably occur in the texts of substitution rites involving a white sheep (UDU BABBAR), either a live animal or a figurine. It is appropriate to begin with passage (1), where the adjectives sahuitantalli(ya/i)- and witantalli(ya/i)- are juxtaposed with the pair ulantalli(ya/i)- 'dead' and huiduwali(ya/i)- 'living', which form a natural merism and belong to several chains of merisms identified in Hittite-Luvian incantations and listed in Mouton and Yakubovich 2019: 211-12. Not all the bipartite merisms included in these chains denote naturally and thus quasi-universally complementary notions; some of them, like 'of mountain-dwellers and desert-dwellers' are culturally determined (see further examples in Watkins 1970, such as 'cattle and sheep' for domestic livestock). As we shall show, 'binding and smiting' is of the second type. If the restoration of [sa-hu-i-da-d]n-ta-al-li-ya-an is correct, this is a strong argument for identifying sahuitantalli(ya/i)- and witantalli(ya/i)- as yet another bipartite merism. Naturally, this tells us next to nothing about the lexical meaning of both adjectives beyond their denoting complementary notions. (6)

   (l)KUB 35.43+ ii 1-3, NS, CTH 761 1. [ma-a-an sa-hu-i-da-a]n-ta-al-li-ya-an ku-i-ha ma-a-an u-i-ta-an-ta- al-li[-a]n 2. [ku-i-ha ma-a-am-p]a u-la-an-ta-al-li-ya-an ku-i-ha 3. [ma-a-am-pa hu-i-]du-wa{-la}-li-ya-an ku-i-ha [Whether] anything (is) of [bindjing, whether [anything] (is) of smiting, [whethe]r anything (is) of the dead, [whether] anything (is) of the [li]ving. More clarity can be reached once we turn to the continuation of the same incantation introducing the white sheep as a ritual substitute. (7) The objects that it is expected to carry away from the ritual patron are called sahuitant(a)- (n.) and witatt(a)- (n.), the ostensive cognates of sahuitantalli(ya/i)- and witantalli(ya/i)-, respectively. (8) Since the function of substitute animals in ritual contexts consists in absorbing the miasma or other malign phenomena that plague the ritual patron, it stands to reason that sahuitant(a)- and witatt(a)- have negative denotations and/or connotations.

   (2) KUB 35.43+ ii 7-11, NS, CTH 761 7. [ma-a-an t]a-a-u-wa-ti ku-i-ha ma-a-na-ta {erasure} i-is-sa-ra-ti ku-i-ha 8. [a-a-ya-t]a ma-a-an ta-wa-as-sa-ti ku-i-ha tu-um-ma-an-te-et-ta 9. [sa-hu-i-t]a-an-ta-al-li-ya-an-za u-i-ta-an-ta-al-li-ya-an-za 10. [na-nu-um-p]a za-as pa-ra-ad-du a-ar-ra-az-za ha-a-u-i-is 11. [sa-hu-i-t]a-an-ta-an u-i-ta-at-ta-an [Whether] (he = the ritual patron) saw anything with (his) eyes (or) [mad]e anything with (his) hands, whether (he) heard anything with (his) ears, that of [bin]ding or that of smiting, [now] may this white sheep carry (away) the [bo]nd and the blow! While witatt(a)- and its cognate witantalli(ya/i)- are well attested in (1) and (2), sahuitant(a)- and its derivative sahuitantalli(ya/i)- represent a product of restoration in KUB 35.43. The parallel that is conducive to fleshing out the stem /sa[[gamma].sup.w]id-/ in the merisms under discussion belongs to a different manuscript but also reflects a white sheep substitution rite embedded in the Kuwattalla tradition.

   The relevance of the fragmentary incantation (3) was obfuscated for a long time by the natural but erroneous phonetic reading lu-ut instead of the Sumerogram UDU BABBAR (e.g., Starke 1985: 101). The equally fragmentary Hittite narrative preceding the incantation refers to the sheep fat (KUB 35.74: 2' [.sup.UZU]I.UDU), presumably the raw material for fashioning the figurine of a substitute white sheep. Crucial for our purposes is the presence of sahuitant(a)- and witatt(a)- in the adjacent lines of the incantation, which indicates that the white sheep rite was expected to absorb both kinds of malign phenomena.

   (3) KUB 35.74 5'-8', NS, CTH 762 5'. [za-u-i-na-as] u-i-si-ta UDU BABBAR [...] 6'. [sa-hu-i-ta-a]n-tal-li-is '[-i-ta-an-tal-li-is] 7'. [la?-a?-la?-at-t]a...