*ópi
‘back (?)’
- Language
- Proto-Luwic
- Grammar
- adv
- ID
- 3087
1. Individual Anatolian Languages
1.1 Hieroglyphic Luwian
/appi/, á-pi
‘back’1.1.1Transmission
This lexeme is attested 31 times in inscriptions from Assur, Commagene, Tabal, Hama, and the Amuq, dating from the mid-9th to the late 8th century BCE. More than 20 of the attestations originate from the ASSUR letters.
[A.P.]
1.1.2Forms
This table offers a representative selection of attestations, excluding broken examples.
[A.P.]
1.1.3Graphic Features
1.1.3.1Phonological Interpretation
The phonological interpretation of this word has been clear ever since Oshiro 1988a:247 connected it with the adverb Hittite appa ‘back’ and thus the entire word families around it in Hittite and its cognates in Luwian. Subsequent recognition of CLuw. appi further supports this analysis, cf. CLL:23.
The spelling of the word is remarkable in so far as it represents one of the words consistently spelt with 〈á〉 (sign *19) instead of 〈a〉 (*450), a choice that has caused much scholarly discussion, not least because 〈á〉 appears to be limited to word-initial position by and large. Kloekhorst 2004a (again Kloekhorst 2006a) argues that 〈á〉 represents a different sound than 〈a〉, viz. a vowel preceded by a glottal stop, i.e. /ʔa/. Differently, Rasmussen 2007a:165 assumes that 〈á〉 is used to reflect /e/, which is unlikelySee Simon 2013a:9f. for a comprehensive critique. since it is generally agreed that Luwian has a three-vowel system /i, a, u/ and distinguishes for length (cf. e.g. Melchert 2003b:185). Yet another hypothesis is put forward by Poetto 2010a: 272, who suggests (following up on an earlier note in Poetto & Salvatori 1981a:19) that 〈á〉 might stand for /ya/, although he can draw on only tenuous onomastic evidence, and particularly the present lexeme appears to be a counter example. Examining word-initial plene spellings across Luwian, Simon 2010b:263f. sees 〈á〉 in the role of representing long /ā/ as well as /ʔa/. Melchert 2010a:153 notes that 〈á〉 and 〈a〉 are no longer complementarily distributed from the middle of the 9th century BCE onwards, arguing against 〈á〉 representing a glottal stop. This view is met with hesitation by Weeden 2011a:67f. in his review of Kloekhorst’s EDHIL. Reviewing the data yet again, Simon 2013a reaches the conclusion that it does represent /ʔa/ after all. No position is taken here.
[AH.B.]
1.1.4Meaning
Even before the word had received a semantic interpretation, it was clear that it did not constitute an ordinary adverb like /appan/ ‘behind’ and can neither function as a postposition nor as a preverb. It was first assigned a meaning by Oshiro 1988a:247, who emphasised the meaning ‘again’ in particular with respect to the contexts in the ASSUR letters and also assigned the meaning ‘then, afterwards’ to /appi/. Examining the ASSUR contexts in detail again, Hawkins 2000a:554f. largely follows his analysis, but is hesitant about the meaning ‘again’. Melchert (pers. comm. apud Hawkins 2000a:550) points to the fact that /appi/ coincides with the introduction of new topics particularly in ASSUR letter f+g. Taking up this suggestion, Waal 2016a:3-5 translates /appi/ in the ASSUR letters contexts as a topicaliser that is used to redirect the letter recipient’s attention back to the sender or vice versa particularly in the introductory clauses of the letters in the exchange of pleasantries, including complaints about the lack of such. Thus e.g. in ASSUR letter a, §3: a-za5-za-ha-wa/i-za á-pi ha-tu-ra+a, approx. /anzanz=ha=wa=anz appi hattura/, translating as ‘(Are you well still?) As for us, we (are) also in life. (Now, I’ve become irrelevant to you at your place.)’ (cf. Waal 2016a:4).
A similar use can be detected in the KIRŞEHİR letter. However, it does not always appear to make sense, and some contexts benefit from a more literal translation as ‘back’. From a development perspective it is likely that the topicaliser function so prevalent in the letters developed out of the adverbial function proper, in so far it was present in the first place.
[AH.B.]
1.1.5Stem
Oshiro 1988a:247 pointed out the etymological connection between /appi/ and the adverb /appa/ ‘back’, mostly relying on Hittite appa ‘back’ due to the obscurity of Luwian /appa/ at the time. The /-i/ ending is usually deemed locatival. Differently, Dunkel 2014a:7 n. 36 hypothesises that at the Proto-Anatolian stage already, the word that would later come out as /appi/ in HLuw. largely replaced /appa/ with a vocalism fashioned after /pari/ ‘before’, imitating the opposition of motion vs. the locatival sense of /appan/ ‘behind’.
[AH.B.]
1.2 Cuneiform Luwian
=appi
‘behind (?)’1.2.1Transmission
The particle occurs almost exclusively in identical clauses in the song incipits from the Ištanuwa festival text.
[E.R.]
1.2.2Forms
| indecl. | (=)a-ap-⸢pí⸣ | KUB 35.139 i 9’ | (CTH 773, MS) |
| (=)a-[ap-pí] | KUB 35.139 i 10’ | (CTH 773, MS) | |
| (=)a-ap-pí | KUB 35.139 i 14’ | (CTH 773, MS) | |
| (=)a-a[p-pí] | KUB 35.139 i 17’ | (CTH 773, MS) | |
| (=)a-ap-pí | KBo 14.121 ii 7’ | (CTH 773, MS?) | |
| (=)a-ap-p]í | KBo 14.121 ii 11’ | (CTH 773, MS?) | |
| (=)a⸣-ap-pí | KBo 14.121 ii 14’ | (CTH 773, MS?) | |
| (=)a-ap-pí | KBo 14.121 iii 2’ | (CTH 773, MS?) | |
| (=)a-ap-p[í] | KBo 14.121 iii 4’ | (CTH 773, MS?) | |
| (=)a-ap-pí | KBo 14.121 iii 7’ | (CTH 773, MS?) | |
| (=)a-ap-p[í] | KBo 14.121 iii 11’ | (CTH 773, MS?) | |
| (=)a-a]p-pí | KBo 40.256, 5’ | (CTH 773, MS?) |
The doubts expressed in CLL:23 about the segmentation of the particle chain seem unnecessary. The existence of āppa in Cuneiform Luwian does not speak against the assumption that the Ištanuwa dialect features =appi as an enclitic particle with a specialized or grammaticalized function.
[E.R.]
1.2.3Graphic Features
Checked against the photographs in the Mainzer Fotoarchiv.
[E.R.]
1.2.4Meaning
The particle =appi occurs in enclitic position after the 2sg./3sg. dat. personal pronoun(?) and reflexive particle =ti attached to the first word of the clause. If the order is comparable to that of the other Luwian dialects and Hittite, =appi must represent an “Ortsbezugspartikel” in the last position of the particle chain (Carruba 1968a:15 n. 3). An etymological relationship with the adverb/postposition āppa ‘back, again’ is most likely. The context repeated several times in the ritual runs as follows: lalluwyawa tapāla duwayai kuiš=ti=appi wašša ‘he reveres the Storm clouds belonging to Lallupiya, who is behind thee, rain (?)’ (for the translation, see under duwaya-(i); only the context at KUB 35.139 i 17’ is different, but provides no further information). The particle =appi is translated here tentatively with ‘behind (?)’ on etymological grounds (as per CLL:23; cf. CLuw. āppa ‘back, again’).
[E.R.]
1.2.5Stem
CLuw. =appi is yet another example of an adverb that reflects several frozen case endings: directional -a in āppa, locatival -i in =appi.
[E.R.]
1.3 Lycian A
epi
‘(preverb)’1.3.1Transmission
This preverb is well attested throughout the Lycian corpus.
[D.S.]
1.3.2Forms
Despite being the last word of N 325, Bousquet 1992a:186 suggests that ]epide represents the « conjunction » epi=de ‘désormais’ (following Laroche 1979a:351), which is only possible, if the text continues on the next face, i.e. N 324, although the opposite is actually presumed by Bousquet. However, contra Laroche 1974a:119 (and Laroche 1979b:351, DLL:14), epide must not be taken as a full conjunction meaning ‘in the future (vel sim.)’, but rather as the accented preverb epi followed by the particle =de (thus also Neumann 2007a:60). The analysis by Laroche is followed by Cau 1999a:27 n. 18 and Melchert 2002c:249 n. 36 for the passage of TL 44a.41, which is not necessary there either; for the syntactic analysis of this passage, see Sasseville 2021c.
[D.S.]
1.3.3Graphic Features
Neumann 2007a:59 claims that ep[i] tuwete (TL 27.1) is rather to be restored as ep[ñ], which is not supported by the parallels tuwete (e)pi (TL 44a.20) and me=t(e)=epi=tuwẽti (N 320.32-33). Though possible, the phrase epñ tuwe-(ti) ‘to set back up’ would be rare for a dedicatory inscription.
[D.S.]
1.3.4Meaning
Pedersen 1904a:198f. argues cogently that Lyc. epi is a preverb, though of obscure semantics. This is followed by Meriggi 1936b:276 who equates its function with the one of the Greek cognate ἐπί ‘on top of’, which is hardly accurate. Laroche 1979b:350f. compares Lyc. epi with Hitt. appa and appan, restoring a CLuw. appi, which turns out to be attested in the Luwian dialect of Ištanuwa as well as in Hieroglyphic Luwian (the former is pointed out by Starke 1990a:606). Melchert (DLL:14) sees in example N 320.32-33 (me=t(e)=epi=tuwẽti : mara : ebeija ‘They establish these laws here’) a “telicizing” effect, which he claims would work well for other examples. Although his claim is entirely possible, it cannot be considered proven. Overall, most scholars agree that Lyc. epi is a preverb but leave it untranslated (Laroche 1979b:351, Neumann 1984b:93, DLL:14 “local adverb with varying force”; contra Neumann 2007a:60 “semantically comparable to Lyc. epñ ‘after, behind’”Neumann’s claim is etymologically motivated and it is clear that epi has synchronically acquired a completely different function that still eludes us..
[D.S.]
1.3.5Compounds
2. Reconstruction
2.1Reconstruction ‘back (vel sim.)’
Comparing Lycian epi ‘(preverb)’ with HLuw. /appi/ ‘back’ and CLuw. (Ištanuwa) =appi ‘behind (?)’, we may reconstruct a Proto-Luwic adverb *ópi ‘back (vel sim.)’ (thus also Melchert 2017a:530, less likely *ápi, see LIPP II:67) with semantics similar to its counterpart found in Hitt., CLuw. and HLuw. /appa/ (The form *ópi is preferred by LIPP II:72, cf. Gk. ὀπίσσω ‘backward’, although *ápi is also possible; Greek ἐπί ‘on top of’ is separated etymologically by LIPP II:244f., so that a reconstruction *épi, albeit phonologically possible, is less attractive here). Contra Melchert 2009c, Hitt. appāi-/-i(ya)-(ḫi) ‘to end’ does not require the existence of a Pre-Hittite *appi, since the derivational process of the suffix -āi-/-i(ya)-(ḫi) is hardly restricted to adverbs ending in -i (see Sasseville 2021a:437f.).
[D.S.]