Word-internal plene spelling with and <e> in cuneiform Luwian texts

Rieken Elisabeth

Elisabeth Rieken

Philipp University of Marburg; rieken@staff.uni-marburg.de

Word-internal plene spelling with and <e> in Cuneiform Luwian texts

Melchert's hypothesis that the cuneiform "orthography" of Hittite was transferred to Cuneiform Luwian by the Hittite scribes was tested with regard to the plene spelling with . With our improved knowledge of the historical grammar of Luwian, it could be confirmed. Several cases of plene found a new explanation. In addition, the restricted use of plene was described by a limited set of clear rules.

Keywords: plene spelling, Luwian phonology, Luw. zila

1. Within the syllabic writing systems of Cuneiform scripts, such as Hittite, the syllabograms include signs representing a single vowel (V), combinations of consonant and vowel (CV, VC), and sequences of consonant — vowel — consonant (CVC). Plene spelling is defined as the extra use of an additional unitary vowel sign (V) that either precedes or follows an identical vowel. Thus, in word-internal position, plene spelling follows the patterns of CV1-V1-V1C and CV1-V1-CV as opposed to the non-plene spellings of CV1-V1C and CV1-CV, respectively. In the case of sequences of two vowels, we find plene spellings such as CV2-V1-V1C, CV2-V2-V1C and CV2-V2-V1-V1C vs. non-plene CV2-V1C (or Ci-ya-aC / Cu-wa-aC containing the unified signs <ya> or <wa>).

There are five unitary vowel signs available in the Mesopotamian cuneiform script as adapted for the Anatolian languages: <a>, <e>, , and . This article will focus on the function of the plene spelling with the vowel signs and <e> in word-internal position in Luwian texts transmitted in the cuneiform script.1

While little attention has been paid to the distribution and function of plene spelling in Luwian, the same phenomenon has been frequently discussed in Hittite scholarship. The results of these studies may very well be important for Luwian also because, following Melchert (1994: 27; cf. also Kloekhorst 2008a: 118),

„[t]he documents we have in Hittite, Palaic and Cuneiform Luvian were written by the same scribes working in the same tradition. ... without counterevidence we may and should assume that the principles of orthography are the same for all three languages ..."

Previous assumptions about the function of plene spelling in Hittite are listed in full detail in Kloekhorst 2014: 13-18. According to these, plene spelling is used for

— marking primary and secondary vowel length under the accent (Hrozny 1917: XII; criticized by Gotze 1928: 186 fn. 1; Friedrich 1931: 20; Sturtevant & Hahn 1951: 23),

1 In the course of the work on the "Digitales philologisch-etymologisches Wörterbuch der altanatolischen Kleinkorpussprachen (eDiAna)" funded by the Deutsche Forschungsgemeinschaft (DFG, project RI-1730/7), the topic was brought up by David Sasseville; Miriam Pflugmacher collected the data for i-mutation plene spellings; Zsolt Simon supplemented the data writing first drafts on several lexical items for eDiAna. I am also grateful to an anonymous reviewer who made valuable recommendations. Naturally, all remaining errors are mine.

Journal of Language Relationship • Вопросы языкового родства • 15/1 (2017) • Pp. 19-30 • © The authors, 2017

— marking a glottal stop (Sturtevant 1933: 61-63),

— the disambiguation of Ce/i or e/iC signs with an overgeneralized use in unambiguous cases (Sturtevant 1933: 62f.; followed by Otten & Soucek 1969: 44),

— avoidance of a word consisting of one sign only (except nu, su and ta) (Sturtevant 1933: 64; followed by Otten & Soucek 1969: 49 and generalized for final plene writing in 3-sign words),

— the disambiguation by marking non-dead vowels, the script being on the path to the alphabetic principle (Kronasser 1956: 35 following Pedersen 1938: 5),

— the differentiation of vowel quality in the case of plene <a> (Rosenkranz 1959a: 424),

— vowel length and disambiguation of ambiguous e/i signs (Oettinger 1979: passim),

— marking a secondary effect of the accent on vowel length (Hart 1980: 14f.),

— marking accent (Carruba 1981),

— vowel length including secondary vowel length under the accent, by vowel contraction and compensatory lengthening; also for disambiguation (Georgiev 1983),

— vowel length including secondary vowel length under accent, by vowel contraction and compensatory lengthening (Kimball 1983), but not for disambiguation (Kimball 1983: 7-9),

— vowel length, accent, and disambiguation (Melchert 1984: 83f.),

— marking an initial glottal stop or laryngeal, vowel length in most cases, accent, disambiguation (Weitenberg 1984: 347-350),

— vowel length in most cases, due to accent, but not all accented vowels are lengthened (Melchert 1992),

— only vowel length in most cases (Melchert 1994: 27 and passim),

— for differentiation of <hu> and <ri/tal> in the case of <hu-u> (Kimball 1999: 54-64, 67f.),

— marking an initial glottal stop, reflex of an inherited *hi (Kloekhorst 2006; Kloekhorst 2008: 32; Simon 2013: 12-16 referring to Simon 2010; Kloekhorst 2014: 161-170, 330341, 434-440, 504-508, 529-533; rejected in Weeden 2011: 66f.),

— vowel length due to accent in most cases and the differentiation of vowel quality /o/ and /u/ in the case of plene and (Rieken 2005 based on Held & Schmalstieg 1969: 105-109, Eichner 1980, and Hart 1983: 124-132; cf. also Kloekhorst 2008: 35-60; Kloekhorst 2014: 491-539).2

It is fair to say that the view of plene spelling as a marker of vowel length and, in the case and , as a marker of vowel quality, has gained most supporters and it has also been transferred to Luwian by Melchert (1994: 27 and 2010). However, the latter assumption is not as straightforward as it may seem at first sight. Hittite and Luwian are known to have diverse vowel systems, the main distinction being the lack of /e/ and /e/ in Luwian. Apart from this phonemic difference, we cannot be sure whether or not the phonetic realization of the respective vowels was close enough between the two languages not to cause any insecurity in the perception of the foreign phonemes and their spelling (cf., e. g., the American English pronunciation of Italian pizza with long /1/ in spite of the existence of /1/ in English, due to the realization of Italian /1/ as a vowel higher than English /1/ and identification as the long, tense /1/). In addition, it is a well-known fact that scribes often attempt to spell more accurately when using a foreign language. This may result in the hypercorrect use of extra vowel signs and, as a consequence, as "apparent" plene spellings, so to speak.

2 Plene spelling is assumed to have no function at all by Pedersen (1938: 5, 34, 194) and Kammenhuber (1969: 175).

2. If plene is really a marker of vowel length in Luwian cuneiform texts, the sources of long i are of prime importance to the question. According to Melchert (1994: 240f.), Luwian word-internal long i arises from accented short *i in open syllables, from accented long *i, which in turn may be the outcome of *i through compensatory lengthening, of a rising i-diphthong by monophthongization, and of inherited long *e (in contrast with a < *ehi). While most of this is uncontroversial, the view that inherited long *e resulted in i was challenged by Hajnal (1995: 61-64). He tried to show that *e and *ehi merged into a in Luwian. Surprisingly, the question is relevant only for three of the words treated here (ni-i-is 'not', hi-i-ru-u-un 'oath' and ki-i-sa-am-ma 'combed'). In addition to the sources of long i just mentioned, a development of disyllabic -iya- > -i- still observable in the texts seems to be generally accepted (e. g. Plochl 2003: 20; Melchert 2004: 474; Bauer 2014: 30f.). As a consequence, plene may have multiple sources. In each single case, this source is to be determined. If, with our improved knowledge of Luwian phonology and lexicon, we arrive at a coherent picture for all its attestations within the corpus, Melchert's hypothesis of plene as marker of vowel length can be regarded as confirmed.

3.1. The abstract suffix -til- in Luw. pu-u-wa-ti-i-il 'past' is long known to be the equivalent of Hitt. -zzil- (Melchert 1994: 119f. with references). Hitt. -zzil- goes back to the composite suffix *-ti-lo- from abstract *-ti- plus adjectival *-lo- with apocope of the final syllable. The position of the accent of the Hittite derivative can be determined by both the plene spelling in -zz l- and the apocope (cf. Melchert 2001 and Rieken 2008: 246-9). It stands to reason to assume the same for Luwian -til-. Accordingly, -ti-i-il is a case of lengthening an originally short accented *i in open syllable parallel to its Hittite equivalent. Based on the root connection proposed by Ivanov (2002), a back projection *bhweh2-ti-lo- '(entity) belonging to being' seems plausible.

The same sound change can be assumed for cases of reduplication. In general, verbal reduplication syllables were accented, in Hittite (see Dempsey 2015: 333-41). They were lengthened in open syllables and spelled plene (Melchert 1994: 131). The same seems to apply for Luwian. Cases of verbal reduplication syllables with plene are the 2nd sg. imp. ti-i-ta '?' and the 3rd pl. imp. li-i-la-an-du 'let them pacify'. The meaning and etymology of the former are not known, but since the 1st sg. prt. ti-ta-ah-h[a] of the same verb is proof of a stem tita-, the forms lend themselves to such an analysis. The latter is related to the noun lila-/lela-/lila- 'conciliation, pacification' attested in Hittite texts (Melchert 1993: 127). This, in turn, is interpreted as a reduplicated thematic noun derived from the verb la-/l- (root *lehi- 'loosen, release, remove; cf. Kloekhorst 2008: 523f.; Puhvel, HED 5: 77; Tischler, HEG L/M: 56f.). Since inherited *li-lhi-o-would be reflected by **lilla- with assimilation of the laryngeal (Melchert 1994: 79-81), both verbal lila- and nominal lila-/lela-/lila- must be regarded as a late formations. The frequent forms with e in the reduplication syllable either reflect the original Hittite formation (*le-loh1- > le-la-) or are just another example of the "Hittitization" of a Luwian loanword by means of a hypercorrect replacement of i with e in New Hittite, as suggested by Yakubovich (2010: 326333; differently Kloekhorst 2008: 524).

Luw. ti-i-ta-ni dat./loc. sg. 'breast, teat' follows the same pattern, being related to the participle titaimma/i- 'suckling' (differently Kloekhorst 2008: 875-7). Although there is no Old Script attestation of its Hittite cognate teta(n)-, the almost consistent spelling with <te-> and <te-e-> implies that the reduplication vowel e is original, different from the Luwian reduplicant with i. Thus, Hitt. teta(n)- and Luw. t ta(n)- are parallel formations built according to the rules of their respective languages, with different reduplication vowels, which, however, had been lengthened under the accent in the Proto-Anatolian period in either case. The extension with an n-suffix is surely another late development common to both words due to contact.

3.2. The negation ni-i-is can be derived from either *ne+ (cf. Dunkel 2014: II 536, *ne ehi) or *nei+ (cf. Dunkel 2014: II 537f., *ne ihi) depending on what one accepts as the reflex of *e in Luwian. Neither account would be problematic for the assumption that plene represents a synchronic long vowel i.

A better example of *e > i, spelled with plene , is nom./acc. hi-i-ru-u-un, obl. hi-i-ru-u-t° (contra Hajnal 1995: 61-64; see above section 2). Following Watkins (1993: 469-73), the word goes back to *h2er-u- with accented vrddhi grade, enlarged by a suffix *-t-, and can be compared to Gk. dpd 'prayer, imprecation, curse' < *ar-w-a < *h2eru-eh2. According to Eichner's Law, long *e in *h2er-u- was not colored in spite of neighboring *h2, but developed into i in Luwian (Melchert 2004: 471 fn. 1).

3.3. Long i in Luw. kisamma/i- 'combed' is unexpected from the diachronic point of view. This verb shows consistently a participle ki-(i-)sa-am-m° in all of its three attestations, which would fit a root stem as well as a stem in -a-/-ai-(di). The plene spelling in the root with occurs both in Luwian context (twice) and in the Luwian loanword in Hittite k sa(e)-. Since it should be the suffix syllable of kisa(e)- that carries the stress, long i must originate in Luwian ki-(i-)s°. Melchert (1994: 152) assumes for the Luwian word a denominative formation kis-a-/-ai-(di) based on a noun *kis-a- with lengthened grade. This would require a derivational chain such as PIE root *kes- 'to comb' ^ *kes-o- 'comb' ^ *kes-6- 'belonging to the comb, (subst.) comb teeth' ^ *kes-eh2ye/o- 'to treat with comb teeth, to comb', which is not impossible, but a more economical hypothesis would be preferable. Kloekhorst (2008: 482) operates for Hittite with an ablaut-ing stem *kes-ti/ks-enti, which would give *keszi/kis-anzi and with paradigmatic leveling lead to the attested stem kis-. However, this does not work for Luwian. A slight change of this scenario is necessary. Following Ilya Yakubovich (pers. comm.), we may assume paradigmatic leveling of the root verb *kes-ti/ks-enti at an early stage resulting in PAnat. *kes-ti/kes-enti and, by sound change, in PAnat. *kes-ti/kis-enti. While, in Hittite, only the vowel i of the weak stem was extended to the strong stem (^ kis-tsi/kisantsi <ki-is-zi/ki-sa-an-zi>), in Luwian, the root accent of the strong stem of *kas-ti/kis-anti also spread to the plural giving *kasti/kisanti. After the Luwian lengthening of vowels in open syllables (Melchert 1994: 76, 132) had taken place, the plural stem kis- <ki-i-s°> replaced *kas- in the singular. Then, Luw. kis- functioned also as the base for the Hittite loanword kisa(e)-. Thus, Luw. kisamma/i- 'combed' is perfectly in line with the working hypothesis of this article, but, pace Melchert (1994: 152), is not a case of *e > i.

An analogous explanation applies to the spelling [(UZUha'-ap-p)]i-i-sa-a-ti, which is attested alongside eight cases without plene spelling. For its Hittite equivalent happessar, happesn-, Melchert (2013) starts from a paradigm with mobile accent: nom./acc. *h2ep-s with oblique stem *h2ep-s-n-'. In addition to other uncontroversial phonological changes, *h2ep-s receives an anap-tyctic vowel i (*happis-) after the accented syllable in the strong stem. In the weak stem, however, the anaptyctic vowel was inserted before the accented syllable. It then seems to have attracted the accent and developed into e in happes-n-; cf. Hitt. teri-, Luw. tarri- (according to Cop's Law; cf. Cop 1970) < CAnat. *teri- < PIE *tri-. This is followed by paradigmatic leveling resulting in the attested Hittite paradigm happess+ar, happes-n-. In Luwian, the equivalent phonological developments would produce a paradigm nom./acc. happis, obl. *happass-, which with paradigmatic leveling could give the stem *happis- for all case forms.3 Finally, anaptyctic i was lengthened under the accent (happis- as attested). Apart from the analogy of the Hittite

3 The plene spelling of the abl./instr. ending goes back to the contraction of o-o in *-oyodi < *-oyoti (Rieken 2005) and is generalized throughout all stem classes independent of any secondary accent position.

development, confirmation of the proposed historical derivation comes from the fact that the root syllable never shows plene spelling.

3.4. The denominative suffix -i-(di) < *-e- < *-ee- < *-e-ye- (pl. -ai- < *-e-yo-; with inherited accented suffix *-ye/o-) is found in the 1st sg. prt. la-hu-ni-i-ha 'washed' (^ inf. *lahuna according to Mel-chert 1993: 120). A similar explanation probably applies to ]x-i-ta-an?-al-li-i-ti, a 3rd sg. prs. of a denominative verb from a noun in -alla/i- (David Sasseville, pers. comm.), and to the 3rd sg. imp. ar-za-zi-i-du with its variant ar-s[(a-zi-i-du)] because of the complexity of its base. In contrast, tar-si-i-ta 'dried?' contains the homophonous causative-iterative suffix *-eye/o- (David Sasseville, pers. comm.). Other possible examples of these stem suffixes, either denominatives or causative-iteratives, are the 3rd sg. prt. wa-ri-i-ta '?', attested in broken context, and 3rd sg. prt. at-ti-i-da '?'.

3.5. The Luwian adverbs ku-wa-ti-i-in 'as, how' and a-pa-ti-i-i[n 'thus' taken by themselves would be easily explicable as the combination of the pronominal roots *kw- and *obh- with the nominal endings *-oyoti > *-odi > -adi of the ablative plus the ablative particle *-im > -in, as suggested by Goedegebuure (2010: 86f.). The contraction of the two i vowels would naturally result in long as attested. However, we also find semantically identical forms that feature no final nasal but show long i nevertheless: a-pa-ti-i 'thus' and the corresponding form of the proximal stem za-, i. e. za-a-ti-i 'thus, in this way'. The only way to account for them is to assume an accent shift to the final syllable by analogy with the dat./loc. sg. in -i (as attested in, e. g., ta-ti-i 'for the father' or is-sa-ri-i 'in the hand' and paralleled by Lyc. A tdi < *tedi dat./loc. 'who, which' with syncope of the secondarily unstressed root syllable). The same is probably true for the forms underlying ku-wa-ti-i-in 'as, how' and a-pa-ti-i-i[n 'thus' as well, which may have inherited their long i from *kuwati and apati. As a consequence, we cannot be sure whether the long i comes from the contraction of final -i + *-im or from lengthening under accent. In either case, finds an explanation within the known framework.

3.6. As per Melchert (2009: 114), Luw. nom. pl. zi-i-in-z[i 'these' goes back to *koi + -ms- + -oi, i. e. the pronominal nom. pl. *koi (cf. Hitt. ke 'these') recharacterized by the nominal nom. pl. ending -nzi < *-msoi, which, in turn, is the generalized ending *-ms of the acc. pl. recharacterized by the pronominal nom. pl. ending *-oi. The origin of the plene spelled i-vowel is thus the product of the monophthongization of an i-diphthong. The stem zi-i- was also transfered to acc. pl. zi-i-in-za.

Due to its fragmentary context, the meaning of the acc. sg. form zi-i-da-ni-in is not clear. Melchert (1993: 284) tentatively suggests that it is a derivative of zita/i- 'man'. If so, long i in this word goes back to *oi or *ei (cf. zita/i- < *koi-ti- or *kei-ti-, Gusmani 1987/88: 109).

3.7. The suffix of the 3rd sg. prs. verb form a-an-ni-i-ti 'treats' is contracted from -iya- < *-ye/o-(Melchert 2004: 474). Although attested in a fragmentary context, the form a-ri-i-it[- is very probably a form of the verb ari(ya)- 'raise'. Its plene spelled vowel is therefore also likely to be a contraction product of the change -iya- > i. Likewise, ti-i-ih-ha '?', also attested in broken context, can be interpreted as a 1st sg. prt., derived from *tiya-hha.

A clear case of -iya- > i is mi-i-sa-an-za 'flesh', which has a by-form mi-ya-sa-an-za.

If Kloekhorst's (2008: 1033, 1036-8) reconstruction of a root *tyehi- 'end' (underlying Hitt. zinne/a- 'stop, finish' < *ti-ne-hi-/ti-n-hi-, ze- 'cook (intrans.), be cooked' < *tyehi-, zanu- 'cook (trans.)' < *tihi-neu-) is correct, a root connection with Luw. zila 'subsequently, thereupon' becomes attractive. The Luwian word could then be easily explained as a dat./loc. of a stem zila-< *ziyala- < *tyehi-lo- 'end' vel sim., and the meaning 'towards (its) end' would refer to the state

of affairs mentioned in the previous sentence, whence the grammaticalized function as an adverbial connector. This would provide a new example of -iya- > .

Yet another environment of -iya- > can be found in the suffix -id-. The sheer number of instances of plene spellings in this suffix shows that they do not occur randomly:

wa-ni-i-ti-i-is nom. sg. of wani/it-iya/i- '(made) of rock(?)'; cf. the base word

uwaniya/i- 'of a rock-face, cliff', and uwaniyant(i)-'?' on the same fragment

NA4u-wa-ni-i-ta-im-ma-an nom./acc. sg. n. of part. NA4uwanit-ai-mma/i- 'petrified' wa-ar-hi-i-ta-ti-is nom. sg. of warhi/it-ant(i)-'?'

ma-al-li-i-ta-a-ti abl. of malli/i(t)- 'honey'

[(ti-)]i-ti-i-ta-a-ti abl. of titi/i(t)- 'pupil (of the eye)'

[d]a-a-ni-i-ta nom./acc. pl. n. of dani/i(t)- 'stele'

The accepted analysis of wa-ni-i-ti-i-is is that of an iya/i-adjective of appurtenance derived from an id-stem /wanid-/ 'stele, rock-face' (Starke 1990: 187). In addition to the unexpected plene spelling of the suffix in Cuneiform-Luwian, Hieroglyphic-Luwian features the strange neuter acc. sg. form ("STELE")wa/i-ni-za 'stele' (e. g. in TiLSEVET §5) with the allomorph -za (instead of -sa) in the secondary ending. The nom./acc. with -za occurs in CEKKE side by side with the oblique stem in /-d-/ (§ 3 acc. STELE-zi?! /waniyanza/ and § 22 dat./loc. sg. STELE-ri+i /waniri/ with rhotacism). It is obvious that what we are facing here is a new example of the analogical spread of the thematic ending -an (+ -za), which was first recognized by Melchert (2004) for neuter consonant stems and stems in -u(d)-. Parallel to the extension of nom./acc. sg. *hiru (with regular loss of final -d-) ^ *hiruwan > hirun (with syncope), we may assume that nom./acc. sg. *wani developed into waniyan (+ -za) and, with syncope, further into wanin (+ -za). Moreover, the spread of the long suffix vowel of the nom./acc. hirun to the oblique stem hirud-^ hirud- by paradigmatic leveling, which was suggested by Melchert (2004), is also present in the oblique stem /wanid-/, as shown by its derivatives wa-ni-i-ti-i-is 'of rock(?)' and NA4u-wa-ni-i-ta-im-ma-an 'petrified'.4 Within Hieroglyphic-Luwian, this development is paralleled by the nom./acc. sg. n. forms sanawi-sa (neuter stem in -i(d)-) vs. sanawiyan-za (neuter stem in -iya-) as demonstrated by Yakubovich (2016: 465). The same explanation probably also applies to most of the other cases of Cuneiform Luwian /- d-/, although it cannot be entirely excluded that in some cases -id- was taken over only in the oblique stem while the nom./acc. sg. n. in -i remained intact as attested for nom./acc. sg. n. malli, abl./inst. mallid- 'honey'.

In one out of the five cases, i. e. ("STELE")wa/i-ni-za 'stele', the unexpected case form in -inza indicates that a specific morphological development must be underlying. In another case, malli(t)- 'honey' < *melit-, we expect the stress to be on the first syllable because of the application of Cop's Rule (Cop 1970). In two more cases, the plene spelling of both the first and second syllable (titi(t)- 'pupil of the eye' and dani(t)- 'a cult object'), makes necessary an explanation for the plene spelling of at least one syllable in terms of something other than stress. The combined morphological and phonological explanation for the emergence of the suffix - d-suggested above provides exactly this. For one case only, there is no independent evidence for the origin of long - d- proposed here.

3.8. Following Carruba (1982) and Melchert (1990: 199-201), nine words with plene in the ending were convincingly interpreted as nom. or acc. forms of adjectives with the appurtenance suffix -iya/i-, i. e. with *-iyis, *-iyin and *-iyinzi respectively (see Melchert 1993: s.vv.). Cf.:

4 This makes unnecessary Melchert's (2004: 472 fn. 2) own forced explanation invoking a restoration of the stem final dental wanid- parallel to zarza 'heart' < *zard-sa, which has a completely different phonological shape.

— [hu-]i-it-wa-a-li-i-is KUB 35.49 iv 13' 'alive'

— lu-u-la-h]i-i-is KUB 35.49 iv 15' 'of the mountain-dwellers'

— ta-a-ti-i-is KUB 35.49 iv 14', ta-ti-i-is KUB 35.46, 6', da-a-ti-i-in-zi KBo 29.55 i 6 'paternal'

— AMA-i-is KUB 35.46, 6', AMA-i-in KBo 13.260 ii 30 'maternal'

— GEME-i-is KUB 35.46, 6' 'of a female servant/slave'

— [m]u-ut-ti-i-is KBo 7.68 (+) 69 ii 13' 'having power'

Most of them (with the exception of AMA-i-in and [m]u-ut-ti-i-is) are attested in the second and third ritual of the MUNUSSU.GI fKuwatalla. In the same group of texts we also find five attestations of unexpected plene spellings with of endings with i-mutation:

— [i-is-sa-r]i-i-is 'hand' (KUB 35.46 iv! 2')

— [ku-wa]-an-na-ni-i-in 'eyebrow' (KBo 29.10 ii 6')

— [ma-as-sa-na-al-li]-i-in 'divine' (KBo 29.10 ii 7')

— du-u-pa-im-mi-i-is 'struck' (KUB 32.8(+) 5 iii 28')

— la-al-pi-i-in= 'eyelash' (KUB 32.8(+) 5 iii 14')5

Two pairs of them come from the same fragment and can be clearly regarded as mistakes in terms of postulated function of the plene spelling since they occur beside multiple examples without plene attested in parallel contexts of the same ritual. What we observe in the first group of examples is an earlier attempt to render faithfully the reflexes of *-iyis, *-iyin and *-iyinzi, which contain either a disyllabic sequence or, more probably, vowel length.

Later copyists became confused and, by hypercorrection, added plene in the mutation syllable of the words of the second group. This was probably prompted by the other plene spellings in the same text that the scribes may not have understood properly and interpreted as an attempt to represent the vowel quality instead of quantity. Zsolt Simon (pers. comm.) kindly draws my attention to the fact that the extra in [i-is-sa-r]i-i-is 'hand' is rather compressed compared to the surrounding signs and seems to have been squeezed in secondarily. This may also be interpreted as an indication of the lack of confidence on the side of the scribe.

Confirmation for this hypothesis can be found in cases of unexpected plene spelling of the i-vowel within Hittite contexts. Instances of these are:6

— :ku-ra-a-im-mi-i-is 'cut' (KUB 51.27 i 7')

— si-wa-an-ni-e-es '?' (KUB 35.146 iii 8')

— [№nda]7u-wa-am-me-i-en (a bread) (KUB 25.50 ii 8)

— [NlNDA]la-at-ta-ri-i-en (KUB 25.50 ii 9)

— ku-wa-ra-am-mi-e-es (KBo 30.168 Rs. 9)

If the Hittite scribes wished to render the words of the Luwian texts, which belong to a foreign language, with special accuracy (see section 1), this could easily result in a hypercorrect spelling by means of an extra vowel sign. On the phonological level, it was not meant to represent a long vowel, but a "proper" /i/, while on the morphological level, plene was used to emphasize the correct Luwian form with i-mutation. This explanation would be valid for the attestations of both the second group of words in this section and :ku-ra-a-im-mi-i-is in Hittite context.

In contrast, we may associate the other four spellings of the latter group containing e-signs with Yakubovich's (2010: 326-33) scenario of the hypercorrect spread of /e/ in Luwian loanwords in order to avoid non-standard "Luwian" i-vocalism in Hittite texts. Thus, the diffusion

5 Two more words stem from the same corpus (KUB 35.13, 20' and KBo 9.41 i 5' and ibid. 6'), but since neither meaning nor context are known, nothing can be drawn from them.

6 Courtesy Zsolt Simon.

of /e/ does not only affect lexemes as Yakubovich (2010: 326-333) had suggested, but also grammatical morphemes.

If this is correct, both types of unexpected plene spelling, with and <e>, are due to hyper-correction, the former in avoidance of wrong Luwian, the latter by shunning putative Luwianisms.

The adjectival interpretation of three items is not assured due to their fragmentary contexts, but there is also no counter-evidence. Three more are fragmentary themselves:

— tar-ga-as-sa-na-al-li-i-i[s 'of(?) a muleteer'

— [ ]x-zu-ú-wa-ni-i-is '?'

— wa-ni-i-ti-i-is 'of rock(?)'

— za-si-i-in[(-) '?'

— pa-ri-i-it[- '?'

— pa-ri-T-[ '?'

3.9. Two more contexts of plene in mutation syllables need consideration. They are due to specific phonological and graphic conditions, respectively.

While the following words are to be classified as iya/i-adjectives as well, their bases end in a vowel (-a- or -u-). With Melchert (1990: 201f.), it is safe to assume that, in this specific environment, the extra sign does not represent vowel length at all, but rather the glide between the first vowel and the mutation vowel (/-ayi-/ or /-uyi-/):

— a-ar-ra-i-in-zi 'long'

— ku-um-ma-i-in-zi 'pure'

— pár-la-i-in 'front (?)'

— wa-ar-pa-i-in-zi 'of enclosure'

— wa-as-ha-i-is 'sacred'

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

— i-wa-ru-ú-i-es 'of iwaru-'

The same phenomenon is observable in other morphological contexts as well, e. g. in the 3rd pl. imp. ap-pa-ra-i-in-du '?'.

Another group of unexpected plene spellings with is found in the position after and <ú> representing either the labial glide or the labial element of /kw/ or /hw/. For Hittite, Kloekhorst (2014: 134-61) was able to show that plene spellings with <e> such as <Cu-(u)-e-eC> and <Ci-(i)-e-eC> in this position are used to make up for the lack of signs <we> and <ye> (as opposed to existent <wa> and <ya>). Thus, the spellings <Cu-(u)-e-eC> and <Ci-(i)-e-eC> are functional equivalent to <Cu-wa-aC> and <Ci-ya-aC>, respectively. Although Kloekhorst (2004: 430-4) seems to be hesitant to transfer the principle to , it does apply to it as well (in spite of the existence of <wis>). As a matter of fact, it offers a convincing explanation for the following spellings with where no long vowel is to be expected:

— ku-i-is, ku-i-in 'who'

— ha-a-u-e-es 'sheep'

— ha-a-ú-i-is 'sheep'

— [ha-a]-ú-i-is 'sheep'

— da-a-u-i-is 'eye'

— da-ak-ku-ú-i-is, [da]- rak/an-ku-ú-i-in 'dark'

— [hu-]i-it-wa-a-li-i-is 'alive' (first plene )

Accordingly, the adverb za-(a-)ú-i-in 'here' is ambiguous. It may be just another example of the last group, but also a case of the contraction -(i)ya- > , if it is a combination of za(-a)-ú-i

'here' + -an 'in' (cf. pariyan 'beyond' from pari 'forth, away' + -an 'in'; for -an < *en(i) see Yaku-bovich: forthcoming). Also in the verb forms u-i-it-t[a-ri and ta-ra-a-u-i-it-ta the plene spelling with after the signs and , respectively, does not allow for any conclusions on the length of the vowel /i/.

An important repercussion of the evidence scrutinized in the last two sections is that there is no reason to assume that the i-mutation vowel was long (contra Melchert 2003: 187f. and Rieken 2005: 171). The plene spellings attested in the mutated endings find various, but plausible and coherent explanations in their respective contexts.

4. There are several cases of plene spelling with that have not been discussed in the previous sections, because nothing can be said about their origin due to the lack of a convincing etymology or morphological analysis. It may, however, be emphasized that in each single case the plene spelling is found in open syllable and no other plene spelling occurs in the word. Therefore, nothing forbids us to assume that the position of the stress was on the plene written syllable, which would then be regularly lengthened. Cf.:

— hal-li-i-na-i, [hal-]li-i-na-i, hal-li-i-n[a-i] (factitive suffix -ina-(i))

— par-ta-ri-i-na-li-ti

— mi-i-lu-us-ga-an

— GIShi-i-e[l-lu-wa, GlShi-i-lu[-wa

— GISti-i-ra-na

— n]a-di-i-en-ta

— hi-i-sa-a[l-, hi-i-sa[(-a)l-

5. The second and much shorter part of this article deals with the plene spelling with the vowel sign <e>. The use of the sign is surprising given that the phonological system of Luwian as established does not contain a vowel /e/. However, there are 20 instances of <e> in the corpus. Interestingly, 10 out of these occur in the context of the sequence <Ci-e-ya>7 instead of expected <Ci-ya> or <Ci-i-ya>, which normally represents the adjectival or verbal derivational suffixes of the shape *-ye/o- (e. g. na-a-ni-e-ya 'of the brother' and a-ni-e-ya-an-t[i(-) part. of an(n)i(ya)- 'carry out, treat'). In a single case we find <Ci-e-a> with omission of the palatal glide (wa-ri-e-a). If the choice of <e> is more than just a spelling convention, it should mark an allo-phone of /i/ that may have arisen by a kind of dissimilation process next to the palatal glide.

An allophonic interpretation is probable also for the attestations of <e> in the neighborhood of /h/, which is a typical lowering context;8 cf. si-e-hu-wa-en-zi se-e-wa, hu-u-e-hu-u-i-ya, and hi-e-ru-un. Obviously, lowering of /i/ precedes the regular loss of /h/ before the labiovelar glide in se-e-wa (for the loss see Melchert 1994: 258).

Finally, <e> occurs three times in da-a-i-e-ni '?' and once in si-wa-an-ni-e-es '?'. The status as Luwian is dubious in either case (cf. Melchert 1993: 201 and Rieken 1999: 37 contra Melchert 1993: 195). Perhaps the spelling with <e> in this phonetic context, which is unusual for Luwian, indicates a Hittite origin of the two words.

Based on the plene spellings with <e>, we may thus conclude that Luwian /i/ had an allo-phone [e], which occurred within the diphthong /iya/ [eya] and before and after /h/.

7 After the <e> in lu-ri-e-x[ (KBo 8.130 iii 2') two horizontal wedges that may belong to a <ya> can still be seen.

8 It also causes the lowering of /u/ to [o] in Luwian (Melchert 2010) and to /o/ in Hittite (Rieken 2005 and Kloekhorst 2014, with references). A second option for hi-e-ru-un would be the lowering of /i/ before /r/, which is a lowering context for /u/ in Hittite as well.

6 Summary. More than 20 years ago, Melchert (1994: 27) brought forward the hypothesis that the scribes in the Hittite capital used the same spelling rules for Luwian as they did for Hittite. In the previous sections, this assumption has been put to a test by scrutinizing all data available for plene in Luwian texts. It was shown that, with our improved knowledge of spelling rules in Hittite and Luwian historical phonology and grammar, Melchert's hypothesis was confirmed and can be used as a reliable basis for the investigation of plene spellings with other vowel signs. In general, plene in Luwian marks a long vowel /1/, which arises from original or anaptyctic /i/ under the accent (attested in open syllables; section 3.1 and 3.3), by sound change *e > 1 (section 3.2), by contraction of two e-vowels via *e > 1 (section 3.4) and contraction of two i-vowels (sections 3.5 and 3.8), through monophthongization of the diphthongs *oi and probably *ei (see sections 3.6 and 3.2), and by syncope of iya > i (section 3.7). In support of the hypothesis., new etymologies (e. g. zila) and morphological analyses (suffixes -id- and -til-, ending -in) were proposed. However, in addition to the phonological and morphological contexts listed above, spelling peculiarities after vowels (section 3.9), the relevance of origin and transmission, and other sociolinguistic factors for the use of plene and <e> were taken into account (section 3.8). As a result, it became clear that the vowel /i/ in the i-mutated endings is short. Plene spellings in this morphological context find various other explanations that allow us to attribute them to the derivation with the suffix of appurtenance *-ye/o-, to certain phonological conditions and to hypercorrection.

Plene spelling with <e> indicates an allophone [e] of the phoneme /i/ that is limited to the context of the diphthong /iya/ and the position before and after /h/ (section 4).

The rules for the occurrence of spellings with a plene and <e> extrapolated from the corpus are not always applied with the same degree of strictness, but the evidence for a meaningful usage of this graphic device is clear enough not to discard it hastily.

Bibliography

Bauer, Anna H. 2014. Morphosyntax of the noun phrase in Hieroglyphic Luwian (Brill's Studies in Indo-European Languages & Linguistics 12). Leiden / Boston: Brill. Brosch, Cyril. 2011[2012]. Zum hethitischen Lautgesetz *Two > Ta. Historische Sprachforschung 124: 59-65. Carruba, Onofrio. 1981. Pleneschreibung und Betonung im Hethitischen. Zeitschrift für vergleichende Sprachforschung 95: 232-248.

Cop, Bojan. 1970. Eine luwische orthographisch-phonetische Regel. Indogermanische Forschungen 75: 85-96. Dempsey, Timothy R. 2015. Verbal reduplication in Anatolian. PhD dissertation UCLA. URL:

http://www.escholarship.org/help_copyright.html#reuse. Dunkel, George E. 2014. Lexikon der indogermanischen Partikeln und Pronominalstämme. Vol. 2: Lexikon. Heidelberg: Winter.

Eichner, Heiner. 1980. Phonetik und Lautgesetze des Hethitischen — ein Weg zu ihrer Entschlüsselung. In: Mayrhofer, Manfred / Peters, Martin / Pfeiffer, Oskar E. (ed.). Lautgeschichte und Etymologie. Akten der VI. Fachtagung der Indogermanischen Gesellschaft. Wien, 24.-29. September 1978: 120-165. Wiesbaden: Reichert. Friedrich, Johannes. 1931. Hethitisch und „kleinasiatische" Sprachen. Geschichte der indogermanischen Sprachwissenschaft, 2. Teil, 5. Band, Lief. 1. Berlin: de Gruyter. Georgiev, Vladimir I. 1983. Hethitica II. Balkansko Ezikoznanie / Linguistique Balkanique 26: 5-28. Götze, Albrecht. 1928. Madduwattas. Hethitische Texte 3. Leipzig: Hinrichs.

Gusmani, Roberto. 1987/88. Recenti apporti alla questione delle forme "satem". Incontri Linguistici 12: 105-110. Hart, Gillian R. 1980. Some observations on plene-writing in Hittite. Bulletin of the School of Oriental and African Studies 43: 1-17.

Hart, Gillian R. 1983. Problems of writing and phonology in Cuneiform Hittite. In: Transactions and Proceedings of the American Philological Association 1983: 100-154.

Held, Warren / Schmalstieg, William R. 1969. Some comments on the Hittite phonological system. In: General Linguistics 9: 93-110.

Hrozny, Bedrich. 1917. Die Sprache der Hethiter: ihr Bau und ihre Zugehörigkeit zum indogermanischen Sprachstamm. Ein Entzifferungsversuch. Boghazköi-Studien 1. Leipzig: Hinrichs.

Ivanov, Vyacheslav V. 2002. Indo-European *bhuH- in Luwian and the prehistory of past and perfect". In: Proceedings of the Twelfth Annual UCLA Indo-European Conference, Los Angeles, May 26-28, 2000: 70-106. (Journal of Indo-European Studies Monograph 40). Washington, DC: Institute for the Study of Man.

Kammenhuber, Annelies. 1969. Hethitisch, Palaisch, Luwisch und Hieroglyphenluwisch. In: Spuler, Bertold (ed.). Altkleinasiatische Sprachen (HbOr. 1. Abt. II. Bd., 1. u. 2. Abschnitt, Lieferung 2). Köln / Leiden: Brill.

Kimball, Sara E. 1983. Hittite Plene Writing. Philadelphia: University of Pennsylvania.

Kimball, Sara E. 1999. Hittite historical phonology (Innsbrucker Beiträge zur Sprachwissenschaft 95). Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck.

Kloekhorst, Alwin. 2008. Etymological dictionary of the Hittite inherited lexicon (Leiden Indo-European Etymological Dictionary Series 5). Leiden / Boston: Brill.

Kloekhorst, Alwin. 2008a. Studies in Lycian and Carian phonology and morphology. Kadmos 47: 117-146.

Kloekhorst, Alwin. 2006. Initial laryngeals in Anatolian. In: Historische Sprachforschung 119: 77-108.

Kloekhorst, Alwin. 2014. Accent in Hittite. A study in plene spelling, consonant gradation, clitics, and metrics (Studien zu den Bogazköy-Texten 56). Wiesbaden: Harrassowitz.

Kronasser, Heinz. 1956. Vergleichende Laut- und Formenlehre des Hethitischen. Heidelberg: Winter.

LIV2 = Kümmel, Martin and Helmut Rix, eds. 2001. Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstammbildungen. 2nd ed. Wiesbaden: Reichert.

Melchert, H. Craig. 1984. Studies in Hittite historical phonology (Ergänzungshefte zur Zeitschrift für Vergleichende Sprachforschung 32). Göttingen: Vandenhoeck & Ruprecht.

Melchert, H. Craig. 1990. Adjectives in *-iyo- in Anatolian. Historische Sprachforschung 103: 198-207.

Melchert, H. Craig. 1992. Hittite vocalism. In: Carruba, Onofrio (ed.). Per una grammatica ittita. Towards a Hittite Grammar (Studia Mediterranea 7): 182-196. Pavia: Iuculano.

Melchert, H. Craig. 1994. Anatolian historical phonology (Leiden Studies in Indo-European 3). Amsterdam / Atlanta: Rodopi.

Melchert, H. Craig. 2003. Language. In: Melchert, H. Craig (ed.). The Luwians (Handbuch der Orientalistik 68): 170210. Leiden / Boston: Brill.

Melchert, H. Craig. 2004. The inflection of some irregular Luvian neuter nouns. In: Groddek, Detlev / Rößle, Sylvester (ed.). Sarnikzel. Hethitologische Studien zum Gedenken an Emil Orgetorix Forrer (19.02.189410.01.1986) (Dresdner Beiträge zur Hethitologie 10): 471-475. Dresden: Verlag der TU Dresden.

Melchert, H. Craig. 2009. The animate nominative plural in Luvian and Lycian. In: Nedoma, Robert / Stifter, David (ed.). *h2nr. Festschrift für Heiner Eichner (Sprache 48): 112-117. Wiesbaden: Harrassowitz.

Melchert, H. Craig. 2010. The verbal prefix "u-" and vs. <w> spellings in Anatolian cuneiform. Paper presented at the East Coast Indo-European Conference 29, June 19, 2010, Ithaca.

Melchert, H. Craig. 2013. Hittite "heteroclite" s-stems. In: Cooper, Adam I. / Rau, Jeremy / Weiss, Michael (ed.). Multi nominis grammaticus. Studies in Classical and Indo-European linguistics in honor of Alan J. Nussbaum on the occasion of his sixty-fifth birthday: 175-184. Ann Arbor / New York: Beech Stave.

Oettinger, Norbert. 1979. Die Stammbildung des hethitischen Verbums (Erlanger Beiträge zur Sprach- und Kunstwissenschaft 64). Erlangen / Nürnberg: Carl.

Oettinger, Norbert. 2009. Die Indo-Hittite-Hypothese und die Ausgliederung des Anatolischen. In: Paper at the "Arbeitstagung der Indogermanischen Gesellschaft", Würzburg, 24.-26. September 2009.

Oettinger, Norbert. 2015. Hethitisch partipartiske- 'laufen' (*sperdh-) und mutmutali- 'Schweinewühlplatz'. In: Münchener Studien zur Sprachwissenschaft 69: 269-279.

Oettinger, Norbert. 2015a. Der Flexionstyp idg. *seh2ur 'saure Flüssigkeit', *nebh-s n. 'Gewölk'. In: Indogermanische Forschungen 120: 255-267.

Otten, Heinrich / Soucek, Vaclav. 1969. Ein althethitisches Ritual für das Königspaar (Studien zu den Bogazköy-Texten 8). Wiesbaden: Harrassowitz.

Pedersen, Holger. 1938. Hittitisch und die anderen indo-europäischen Sprachen (Historisk-filologiske meddelelser 25). Kobenhavn: Munksgaard.

Plöchl, Reinhold. 2003. Einführung ins Hieroglyphen-Luwische (Dresdner Beiträge zur Hethitologie 8). Dresden: Verlag der TU Dresden.

Rieken, Elisabeth. 1999. Untersuchungen zur nominalen Stammbildung des Hethitischen (Studien zu den Bogazköy-Texten 44). Wiesbaden: Harrassowitz

Rieken, Elisabeth. 2005. Zur Wiedergabe von hethitisch /о/. In: Meiser, Gerhard (ed.). Sprachkontakt und Sprachwandel. Akten der XI. Fachtagung der Indogermanischen Gesellschaft in Halle/S., 17.-23. September 2000: 537549. Wiesbaden: Reichert.

Rieken, Elisabeth. 2005b. Neues zum Ursprung der anatolischen i-Mutation. Historische Sprachforschung 118: 48-74.

Rieken, Elisabeth. 2010. Das Zeichen <sà> im Hieroglyphen-Luwischen. In: Aygül Süel (ed.), Acts of the VIIth International Congress of Hittitology, Çorum, August 25-31, 2008: 651-660. Ankara: n/a.

Rieken, Elisabeth. 2013. Sekundäre denominal u-Stämme im Hethitischen. In: Cooper, Adam I. / Rau, Jeremy / Weiss, Michael (ed.). Multi nominis grammaticus. Studies in Classical and Indo-European linguistics in honor of Alan J. Nussbaum on the occasion of his sixty-fifth birthday: 274-284. Ann Arbor / New York: Beech Stave.

Rieken, Elisabeth. 2016. Zum luwischen Ursprung von heth. ^ta/uhukanti- 'Kronprinz'. In: Henning Marquardt / Silio Reichmuth / José Virgilio Garcia Trabazo (ed.). Anatolica et Indogermanica. Studia linguistica in honorem Johannis Tischler septuagenarii dedicate (Innsbrucker Beiträge zur Sprachwissenschaft 155): 267-277. Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck.

Rosenkranz, Bernhard. 1959. Zur hethitischen Orthographie und Lautlehre. In: von Kienle, Richard / Moortgat, Anton / Otten, Heinrich / von Schuler, Einar / Zaumseil, Walter (ed.). Festschrift Johannes Friedrich zum 65. Geburtstag am 27. August 1958 gewidmet: 417-426. Heidelberg: Winter.

Simon, Zsolt. 2010. Das Problem der phonetischen Interpretation der anlautenden scriptio plena im Keilschrift-luwischen. Babel und Bibel 4: 249-265.

Simon, Zsolt. 2013. Once again on the Hieroglyphic Luwian sign *19 <я>. Indogermanische Forschungen 118: 1-21.

Simon, Zsolt. 2016. Hinweise auf einen luwischen Lautwandel. N.A.B.U 2016/1: 40f.

Starke, Frank. 1990. Untersuchung zur Stammbildung des keilschrift-luwischen Nomens (Studien zu den Bogazköy-Texten 31). Wiesbaden: Harrassowitz

Sturtevant, Edgar H. 1933. A comparative grammar of the Hittite language (William Dwight Whitney Linguistic Series 1). Philadelphia: Linguistic Society of America.

Sturtevant, Edgar H. / E. Adelaide Hahn. 1951. A comparative grammar of the Hittite language. Revised edition. New Haven: Yale University Press.

Watkins, Calvert. 1993. Some Anatolian words and forms. In: Meiser, Gerhard (ed.). Indogermanica et Italica. Festschrift für Helmut Rix zum 65. Geburtstag: 469-478. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck.

Weeden, Mark. 2011. Spelling, phonology and etymology in Hittite historical linguistics. Review of Kloekhorst 2008. Bulletin of the School of Oriental and African Studies 74: 59-76.

Weitenberg, Joseph J. S. 1984. Die hethitischen u-Stämme (Amsterdamer Publikationen zur Sprache und Literatur 52). Amsterdam: Rodopi.

Yakubovich, Ilya. 2008. The Luvian enemy. In: Kadmos. Zeitschrift für vor- und frühgriechische Epigraphik 47: 1-19.

Yakubovich, Ilya. 2016. A Luwian welcome. In: Sarka Velharticka (ed.). Audias fabulas veteres. Anatolian studies in honor of Jana Souckova-Siegelova: 463-484. Leiden / Bosten: Brill.

Yakubovich, Ilya. Forthcoming. Hittite local adverbs in comparative perspective. In: Rieken, Elisabeth (ed.). 100 Jahre Entzifferung des Hethitischen: Morphosyntaktische Kategorien in Sprachgeschichte und Forschung. Akten der Arbeitstagung der Indogermanischen Gesellschaft in Marburg, 21.-23. September 2015. Wiesbaden: Reichert.

Элизабет Рикен. Инлаутные написания р1епе с <г> и <е> в клинописных лувийских текстах

Гипотеза Мелчерта о том, что клинописная «орфография» хеттского была передана клинописному лувийскому хеттскими писцами, была проверена на написаниях р1епе с <г>. Уточненные представления об исторической грамматике лувийского подтверждают данную гипотезу. Новые объяснения даются ряду случаев написания р1епе с <г>. Кроме того, ограниченность применения р1епе описывается при помощи набора четких правил.

Ключевые слова: написания plene, лувийская фонология, лувийское zila

Word-internal plene spelling with <i> and <e> in cuneiform Luwian texts Текст научной статьи по специальности «Языкознание и литературоведение»

Аннотация научной статьи по языкознанию и литературоведению, автор научной работы — Rieken Elisabeth

Похожие темы научных работ по языкознанию и литературоведению , автор научной работы — Rieken Elisabeth

Текст научной работы на тему «Word-internal plene spelling with <i> and <e> in cuneiform Luwian texts»