*k̑ō
‘here(to)’
- Language
- Proto-Anatolian
- Class
- Derivative
- Grammar
- adv, grammaticalized case form of a proximal pronoun
- Certainty
- low
- ID
- 201
1. Individual Anatolian Languages
1.1 Lycian A
se
‘and’1.1.1Transmission
The conjunction se is attested across the entire Lycian A corpus.
[D.S.]
1.1.2Forms
| indecl. | se | |
| s(e) | ||
| ṣe | ||
| sẹ | ||
| ṣẹ | ||
| [s]e | ||
| s[e] | ||
| [s]ẹ | ||
| sa | ||
| sẽ |
1.1.3Graphic Features
1.1.3.1Phonological Interpretation
The instance sa (← se) in TL 27 represents a rare case of vowel harmony triggered by the following word (i.e. sa ladi), whose first syllable contains the vowel -a-. However, most instances do not show vowel harmony, cf. TL 107a.1 (se lada), TL 101.2 (se ladã), TL 102.1 (se ladi), TL 123.2 (se : ladi), etc.; cf. Hajnal 1995a:79.
When there is a clitic chain, the conjunction se is always the head of it. When there is no clitic chain, se can be written either together with the following tonic word or apart, cf. TL 102.2 (se tideimis), TL 101.2, TL 76.1 (se : tideimis), TL 57.2, 58.2, 77.3 (se tideime), TL 120.2 (se : tideime). But in most cases, a word divider after se is not written. Therefore, it is possible that the conjunction se occurring in front of a tonic word becomes a proclitic. The vowel harmony (i.e. sa ladi, TL 27) discussed above is taken here as evidence for a proclitic status, which had begun at least at the time of the inscription TL 27.
[D.S.]
1.1.4Meaning
The meaning of Lycian se was assigned by Sharpe 1844a:197 equating it semantically with Greek καί by means of a bilingual inscription of Limyra (TL 117). The conjunction se can join both clauses and noun phrases, cf. Laroche 1957-1958a: 173, Carruba 1969b:64f., DLL:56, Neumann 2007a:311f. In the former function, it is not always translatable. The epexegetic function ‘(ModHG) und zwar’ postulated by Neumann (ibid.) in a few instances is not assured. It depends on the syntactic and contextual interpretation of the passages, which cannot be treated here.
[D.S.]
1.1.5Compounds
The sequence (sebe) occurs twice in the Lycian A corpus (TL 44b.36, 37), but opinions differ as to whether it has unified into sebe (cf. Lyc. B sebe) or not, cf. Houwink Ten Cate 1961a:75, Schürr 1997b:61, DLL:8, Neumann 2007a:312.Pace Schürr (ibid.), the two unique instances of se=be are not enough to posit the existence of a metrical passage in the Lycian A part of the Xanthos stele. New evidence has recently emerged in favor of the non-unified sequence: Due to the discovery of a new fragment (Schürr & Dönmez 2015a), which restores the leftward part of TL 44b.37-43, the particle =be occurs once more as an enclitic, but to a different tonic element, i.e. ñte=be (TL 44b.39); cf. =be.
[D.S.]
1.2 Lycian B
se
‘and’1.2.1Transmission
The conjunction se in Lycian B is attested at least once in the sequence se=de. Some scholars see it also in the sequence (siketesi, TL 44c.35, 57), where they interpret s(e) + *iketesi; cf. DLL:128, Neumann 2007a:311f. This interpretation is not followed here; see under the lemma iketese/i-.
[D.S.]
1.2.2Forms
| indecl. | se | TL 44c.51 (Xanthos) |
1.2.3Graphic Features
The single instance of se in Lycian B is doubtful, but the reading is assured; the picture was shown to me by Birgit Christiansen (Email 2015.11.11.).
[D.S.]
1.2.4Meaning
A functional difference between se and sebe ‘and’ in Lycian B cannot be determined; cf. DLL:128.
[D.S.]
1.2.5Compounds
1.3 Sidetic
śa
‘and’1.3.1Transmission
Attested only in the so-called Longer Strategos Votive Inscription from Side of unknown date: Nollé 2001a:634f. claims that Mansel dated it to the 2nd c. BCE since he connected with the weapon reliefs found nearby. Nollé questioned this dating, since this connection is not assured and in his views a monolingual Sidetic inscription in the 2nd c. BCE is “äußerst unwahrscheinlich”. First, the latter argument is only an unproven assumption. Second, Mansel neither dated the inscription nor connected it with the reliefs. Instead, he dated both the city wall and the reliefs on stylistic and historical reasons into the 2nd c. BCE (Mansel 1968a with detailed discussion). The lack of any connection of the inscription with either of these precludes any dating (although the currently unknown meaning of the inscription may change this). Finally, Eichner 1993a:127 dated it to “etwa 3. Jh. v. Chr.”, without further discussion.
[Zs.S.]
1.3.2Forms
| indecl. | śa | S3.1 |
Unclear case:
| indecl. | śa | S3.3 |
Attestations following Pérez Orozco 2008a:136, checked against the published photographs.
The inscription is written in scriptio continua, thus any segmentation is matter of conjecture. In case of the first line (jewaseabaseiśtratagejaśśajewasemasarase) the repeated string jewase and the improbability of a geminated <śś> makes the existence of śa relatively assured, as proposed already by Neumann 1968a:92, Neumann 1978b:874 (though without arguments) and followed by practically everyone since then (Ševoroškin 1975a:162; Faucounau 1980b:648; Eichner 1993a:127; Pérez Orozco 2003a:107, Pérez Orozco 2008a:138). Nevertheless, Nollé 2001a:636 maintains that we do not know if ejaś śa represent one or two words.
The case of the third line (pebijaseñnebarśakaśaakośadanaϑemataś) is much more complicated, however. While a segmentation at °aa° is probable (pebijaseñnebarśakaśa akośadanaϑemataś), it is not clear if śa can be segmented from the end of the first string (as per Faucounau 1980b:648 and Pérez Orozco 2003a:107, Pérez Orozco 2008a:138) or rather belongs to the previous word (as per Neumann 1968a:89–91, Neumann 1978b:874; Ševoroškin 1975a:162; Eichner 1993a:127). Due to these problems Nollé 2001a:636 does not segment this phrase.
[Zs.S.]
1.3.3Meaning ‘and’
There is a general agreement regarding the meaning (Ševoroškin 1975a:164; Faucounau 1980b:648; Eichner 1993a:127; Pérez Orozco 2008a:136), which is made plausible by the parallel structure of “jewase abase (…) śa jewase masarase (…)”.
[Zs.S.]
1.4 Hittite
kā
‘here(to)’1.4.1Literature / Comments
HED K:10, Hoffner; Melchert 2008a:289, Goedegebuure 2014a:290f.
[D.S.]
2. Reconstruction
2.1Reconstruction ‘hither’
Adiego 1995a:31f. (Adiego 2007a:411) compares Lyc.A se, Lyc.B se, sebe, Car. sb and Venet. ke ‘and’ and reconstructs a pre-form PIE *k̑e being either the particle PIE *k̑e ‘here’ or, less likely, a completely new lexeme – this is the communis opinio. We take over this comparison, but we leave out the non-Anatolian language Venetic and add Sidetic śa (Faucounau 1980b:648, Pérez Orozco 2008a:136). Thus, we reconstruct a Proto-Luwic pre-form PLuw. *k̂V ‘and’. The etymological assignment of Sid. śa to this lemma requires that *k̂ develops to Sid. <ś>, for which this would be the first piece of evidence. The absence of the particle *k̂V as a conjunction in Luwian could be attributed to two different reasons; either it was lost and replaced in Pre-Luwian or it is a south-west areal development, in which Luwian did not take part. We also equate etymologically the Hittite spatial adverb kā ‘here(to)’ with the Luwic conjunction. Pace Adiego 1995a:31f., we reconstruct a Proto-Anatolian grammaticalized allative of the demonstrative proximate pronoun PIE *k̑o-/k̑i- ‘this one here’ (see Part 2: Proto-Indo-European Etymology), i.e. PA *k̑ō, based on the arguments described in the following paragraph. The current assignment requires that *ō yields /e/ in Lycian, but /a/ in Sidetic. For different proposals cf. Kronasser 1956a:153. He equates Lyc. A se with Hitt. šu and reconstructs the pronominal stem *so-. Carruba 1969b:64f. criticizes this suggestion by pointing out that an etymological *s before vowels yields Lycian /h/, but admits that conjunctions and particles do not always follow the regular sounds laws. This is not quite exact. Particles and conjunction rather show greater erosion. It is thus not expected that the sound /s/ is retained, where it should have weakened to /h/ and a comparison with the pronominal stem *so- cannot be maintained on phonological grounds, pace Neumann 2007a:312. Carruba 1969b:64f. suggests a univerbation of *s(o)- + *=Ho (CLuw. =ḫa) and compares it to Hitt. šu, which he reconstructs as *so- + deictic element *u (Eichner 1971a:36). This explanation of the conjunction šu is no longer followed; cf. HEG S:1124, LIPP II:299 n. 1 with further literature. The Lycian cluster reduction -sH- regularly yields -s-, so that Carruba’s suggestion would be phonologically proper for Lycian. However, whether the same sound change applies for Carian and Sidetic is uncertain. Bader 1982a:105, 140 derives Lyc. se from *ĝhse and compares the initial sequence with Gk. ξύν, cf. LIPP II:393. The phonological development may be supported by the derivation of Lyc. tese- ‘oath’, if from *dók̂so- (Melchert 1994a:314), but cf. under lemma. Bader´s etymology is possible, but in view of the Anatolian comparanda, it is less attractive than our new etymology suggested above. Houwink ten Cate (Houwink Ten Cate 1961a:74) connects Lycian se etymologically with the Lycian preverb ese. This suggestion had already been made by Torp (Torp 1901a:24 n. 1); cf. Neumann 2007a:312 with further literature. Dunkel (LIPP II: 393 n. 11) accepts it and explains it as e- (< anaphoric *é) + se. The etymology is not accepted here; see under ese.
Phonologically speaking, a reconstruction of the instrumental case of a pronominal pre-form PIE *k̑o-h1 ‘with/through this’ (cf. García Ramón 1997a) is possible (cf. Old Hittite sentence introductory particle ta, if it comes from *to-h1, Rieken 1999b:86, cf. LIPP II:775, but differently Part 2: Proto-Indo-European Etymology). However, the instrumental case in Hittite has no spatial function (Brosch 2014b:76), which diminishes the credibility of a putative PIE *k̑o-h1 ‘with/through this’ developing to ‘here(to)’ in Hittite; cf. Goedegebuure 2014a:288. Goedegebuure (ibid.) suggests an endingless locative for the reconstruction of Hitt. kā, which is not conventional for a Proto-Indo-European thematic o-stem. Besides, no Indo-European parallel can be found, in order to support this hypothesis; cf. Part 2: Proto-Indo-European Etymology.
The functional development of the allative (Brosch 2014b:42-47, Goedegebuure 2014a:284–286, 290f.) on the other hand is closer to the semantics of the spatial adverbs kā ‘here(to)’ and its grammaticalization finds more support from the early Indo-European perspective (see Part 2: Proto-Indo-European Etymology). Although Hitt. kā is more often attested with the locative meaning ‘here’, it does not interfere with an older allative sense. The development from the allative sense to the locative one is unproblematic (cf. Melchert 2017a). Hittite kā may thus share the same etymological ending as the Hittite local adverbs such as anda, arḫa, p(a)rā and š(a)rā, which are taken as old allatives; cf. Starke 1977a:132, Hoffner; Melchert 2008a:76, 290, Melchert 2017a.
The grammaticalization of a Proto-Anatolian allative *k̑ō ‘(in addition) to this, hereto’ to PLuw. *k̂ō, which functions as a conjunction coordinating clauses and lexical items, is not without parallels, cf. Gk. =δε (allative marker) and δέ (additive and adversative conjunction) (LIPP I:148f., 128-131), CLuw. =ḫa. In view of the allative interpretation of PAnat. *k̑ō the Old Hittite sentence connective ta- can be reconstructed the same way, viz. < PIE all. *tō ‘(in addition) to that’.
[D.S.] [A.O.]
3. Proto-Indo-European
3.1Semantic Reconstruction
Syntactically, syntagmatically and phonologically as well as semantically, the conjunctions PLuw. *k̑V ‘and’ and Venet. ke ‘id.’ are directly comparable. Based on these equations, the reconstructable proto-form would be identical with the PIE proximate particle (or endingless locative) *k̑e ‘here’ (LIPP II:396f.). However, since PIE *k̑e ‘here’ is primarily locatival, the semantic change to a copulative conjunction is unlikely. According to LIPP II:692 n. 17, there is no evidence in PIE for coordinating copulative conjunctions that go back to unextended deictic particles.
An alternative solution, which is opted for here, is the connection with Hitt. kā ‘here(to)’,Cf. on the correlation between proximal deictic pronouns and conjunctive coordination Viti 2008a:39f.which originally was used as an allative, near-deictic local adverb. Such adverbs are a plausible source for copulative conjunctions. Similar formations are reconstructable for PIE itselfCf. PIE all. *i̯-ó ‘and’, which is a conjunction with coordinative sense, cf. LIPP I:157, EIEC:20 s.v. AND.as well as for individual IE languages.Cf., e.g., PIE *éti ‘to this, in addition to that’ > Lat. et ‘and’, cf. LIPP II:260–264, WH I:421f., EIEC:215 s.v. FURTHER. Differently de Vaan 2008:195 (from PIE *h1eti- ‘to go beyond’, cf. also WH I:422). PIE *éti itself originates probably in an ablatival particle (*é-ti ‘thence’), but switched path semantics (cf. on this process Hackstein 2007a:138–142).
For more PIE and einzelsprachlich examples cf. LIPP II:131.Therefore, it is plausible that the Luwic conjunction continues PIE *k̑o-o, an original allative formation of the proximal pronoun PIE *k̑o-/k̑i- ‘this one here’.On this pronominal stem, cf. PAnat. *k̑o-/k̑i-/k̑e- ‘this’.
Also Gk. ὧ(=δε) ‘hither’ > ‘here’ can be reconstructed as PIE all. *so-o(=de) ‘thither’. Besides, Gk. ὧ(=δε) ‘this way, so’ goes back to PIE instr. *so-h1(=de).
On semantic grounds, the reconstruction as an instrumental case form PIE *k̑o-h1 has to be dismissed, because the instrumental ending PIE *-h1 is reconstructed as “instrumentalisch, modal” (LIPP I:127).Cf., e.g., the pronominal forms PIE *nó-h1, *só-h1 ‘this way’ (LIPP I:130).Even typologically, the development from ‘instrumental’ to ‘N-conjunction’ is merely impossible, cf. the semantic map by Haspelmath 2004a:24 (fig. 11).Also the sentence connective Hitt. ta- should not be reconstructed as instr. *to-h1 ‘so, this way’ (LIPP I:131, LIPP II:775 n. 1), but rather as all. *to-o ‘(in addition) to that’.
3.2Morphological Reconstruction
In view of the notes on the semantic reconstruction given above, it seems to be more practicable to give up the equation of PLuw. *k̑V ‘and’ and Venet. ke ‘id.’.There are various other etymologies for Venet. ke ‘and’: < PIE *k̑m̥ (Hamp 1993a, similarly Lejeune 1952a), < *kē̆C (on the phonology, cf. Lejeune 1974:108, 148), < so called inverse PIE 1.*=ku̯e ‘and’ (cf. LIPP II:699, Dunkel 1982a:129f. n. 2), ← loan word Gr. καί /kē/ ‘and’ (LIPP II:692 n. 17f.; on the borrowing of coordinators and the reasons for that, cf. Mithun 1988a:351)It is more promising to connect PLuw. *k̑V ‘and’ with Hitt. kā ‘here(to)’ and trace it back to PIE all. *k̑ó-o > *k̑ō ‘hereto’.This is Dunkel’s (LIPP I:157) reconstruction of Hitt. kā. He does not, however, connect the non-Hittite Anatolian material.
On the vowel quality of the ending, cf. LIPP I:154, 155 n. 2, 160f. n. 23. Zeilfelder 2001a:23–39, 101–140 offers a semantic and syntactical study of this case.Lat. hō-vorsum ‘hereto’ (< all. *gho-o, cf. Dunkel 1994a:22)In LIPP II:285 Dunkel reconstructs Lat. hō- as instrumental *gho-h1, which does not make sense semantically.parallels this reconstruction morphologically and semantically.On this pronominal stem, cf. PAnat. *k̑o-/k̑i-/k̑e- ‘this’.
The phonologically and morphologically possible reconstruction as an instrumental case form, viz. PIE *k̑o-h1 is unlikely for semantic reasons, cf. above.
[A.O.]
4. Transmission
NPhryg. -ke, -κε conj. ‘and’ (Haas 1966a:No. 5.1+)This conjunction most likely goes back to enclitic PIE 1.*=ku̯e ‘and’, cf. LIPP II:691 with n. 15. On syntactical and syntagmatic considerations cf. Lubotsky 1989b:149–151.
Messap. †ke conj. ‘and’ MLM 22 Cae, 3rd c. BCE; MLM 48 Al, 3rd c. BCE; MLM 6 Me, 3rd c. BCE)This alleged conjunction (cf., e.g., LIPP II:692 with n. 18) is a ghost word. The grapheme sequences belong to personal names, as is put forward by MLM II:202 s.v. kebeirišoas, 204 s.vv. keošeteos, keošorrihi.
Venet. conj. ke ‘and’ (Lejeune 1974:No. 2Not No. “3” as the index (ibid. 332) tells’, 12 a, 26, (?)63, 203).Pace Lejeune 1974:332, the conjunction is not “proclitique” but merely postpositive to the first coordinated constituent.
On the semantic difference between Venet. ke ‘and’ and kve ‘id.’, cf. Poccetti 2010a:510.
Lejeune 1974:332