*ne
‘not’
- Language
- Proto-Anatolian
- Grammar
- neg
- ID
- 3145
1. Individual Anatolian Languages
1.1 Hieroglyphic Luwian
/na/, NEG
‘not’1.1.1Transmission
This lexeme is attested 68 times in inscriptions from all regions except Emirgazi, Boğazköy, Malatya and Aleppo, dating from the 13th to the early 7th century BCE.
[A.P.]
1.1.2Forms
The above table offers a representative selection, omitting broken attestations.
Uncomplemented NEG2 is commonly assumed to represent /na/ in 1st millennium BCE texts; the earlier attested /nawa/ mostly relies on NEG instead.
[A.P.] [AH.B.]
1.1.3Graphic Features
1.1.3.1Phonological Interpretation
The spelling of this particle caused confusion among scholars for considerable time. In only very few cases, the negation is spelt 〈na〉 (sign *35), whereas the logograms NEG (*332a) and NEG2 (*332b, also read 〈ná〉) are used much more frequently. Confounding the matter further, the latter were long thought to have the same meaning as REL (*329, also read 〈kwa/i〉), although the signs themselves were recognised to differ from one another, see e.g. Laroche 1960c:170-176.This conflation was based on an assumption made by Forrer 1932a:42, who falsely considered NEG2 to represent a ligature of REL and 〈ia〉 because of the double bars at the bottom of sign *332b. It was only Hawkins 1975a:120f. who shed light on the matter when realising that NEG2 could sometimes be used to represent the value /na/ or /n/, and NEG3 (*332c) without the double bars at the bottom was used for /ni/.See Hawkins 1975a also for a history of previous scholarship on the signs.
[AH.B.]
1.1.4Meaning
The meaning of this word as a negation was clear from early on (see Hrozný 1933a:98), although many occurrences were not identified due to the misinterpretation of NEG2, cf. the discussion above. Like its precursor /nawa/ ‘not’, shorter /na/ represents a factual negation, as opposed to the prohibitive negations /nis/ and /ni/, see the treatments by Hawkins 1975a, Morpurgo Davies 1975a, and Hawkins & Morpurgo Davies 2010a.
A passage well-known for its clear illustration of the meaning of /na/ is KARKAMIŠ A23, §§4-5: *a-wa/i-mu mi-zi *a-la/i-ni-zi pi-pa-sa-ta *a-mu-pa-wa/i *a-la/i-na-za NEG2 pi-ia-ta, approx. /a=wa=mu minzi alaninzi pibassatta ammu=ba=wa alananz na piyatta/, translating as ‘She (scil. the goddess Kubaba) kept giving my enemies to me, but she did not give me to the enemies’ (cf. Hawkins 2000a:119, Morpurgo Davies 2011a:209). It can also be found in curse formulae such as KARKAMIŠ A31+, §14: ní-pa-wa/i za-ia (“LOQUI”)mara/i-ta kwa/i-sa NEG2 AUDIRE+MI-ti-ti, approx. /niba=wa zaya maratta kwis na tummantitti/, meaning ‘Or if he (lit.: who) does not hear these words, (him the goddess Kubaba of the … shall oppose!)’ (cf. Hawkins 2000a:142). The negation is typically found immediately before the verb as in the examples cited above. It can also be emphasised with the addition of an adverb in the phrase /na manuhha/ ‘by no means’, see e.g. KARKAMIŠ A5a, §3: a-wa/i FRONS.LA/I/U-˹li?˺ FRATER.LA-˹ni?˺ NEG2-a ma-nú-ha INFRA ARHA pa+ra/i-há, approx. /a=wa hantili …-ni na manuhha zanta ahha parrahha/, translating as ‘I became by no means irrelevant to my eldest brother’ (cf. Hawkins 2000a:182).
Rarely, a double negative occurs, in which case a logical positive seems to be the outcome instead of a negative (as is often the case in other languages), as noted by Hawkins 1975a:145, cf. also Melchert 2003b:206. Thus e.g. ASSUR letter d, §10, combining prohibitive /ni/ and factual /na/: ni-wa/i-mu-i á-pi NEG2-a VIA-wa/i-ni-si, approx. /ni=wa=mu appi na harwanni/, likely translating as ‘(Dispatch [some military equipment?] for the master!) Don’t not dispatch it to me!’ (i.e. ‘Don’t fail to dispatch it!’; cf. Hawkins 2000a:535).
[AH.B.]
1.1.5Compounds
1.2 Cuneiform Luwian
nā
‘not (?)’1.2.1Transmission
This word is attested once in a song of Ištanuwa.
[D.S.]
1.2.2Forms
| indecl. | na-a | KUB 35.137 rev.? 5’ | (CTH 773, NS) |
1.2.3Graphic Features
Checked against the photographs in the Konkordanz (Mainzer Fotoarchiv).
[D.S.]
1.2.4Meaning
Despite the fragmentary context, Meriggi 1957a:203 n. 28 identified this word potentially as the negation nā without the addition of the morpheme -wa. His analysis is tentatively followed by CLL:152, but not by Laroche 1959a:72, who refrains from assigning any interpretation. In any case, Meriggi’s suggestion is tentatively adopted here.
[D.S.]
1.2.5Derivatives
1.3 Lycian A
ne
‘not’1.3.1Transmission
This lexeme is found across the Lycian corpus.
[D.S.]
1.3.2Forms
The instance in TL 114.2 is taken instead as an enclitic pronoun, i.e. se=i=(e)ne (scil. Esedeplẽme’s name) epñ puñte (cf. DLL:43). The sequence na lau is taken as *ne=lau with vowel harmony on the proclitic element (Hajnal 1995a:163f. n. 189, DLL:43).
[D.S.]
1.3.3Meaning
Using a combinatorial approach, Torp 1898a:30f. recognizes rightly two negative particles in Lycian ne and ni (followed by Bugge 1901a:39, Meriggi 1936b:276, DLL:43, Neumann 2007a:236f.). Thomsen (apud Pedersen 1899a:1101) points out that Lyc. ne must be the objective negation, since the predicate is in the indicative, whereas ni must be the prohibitive negation, since the predicate is regularly in the imperative mood. The negation regularly precedes the verb and is often part of the sentence initial particle chain.
[D.S.]
1.3.4Compounds
1.4 Lycian B
ne
‘not’1.4.1Transmission
This lexeme is found on the inscribed pillar of Xanthos dated contextually to shortly after the ascension of Artaxerxes II in 405-404 BCE and stylistically around 400 BCE (Borchhardt & Eichner 1997-1999a:19f.) and in the poetical inscription engraved on the lion sarcophagus of Antiphellos dated on stylistic grounds to the beginning of the 4th century BCE (Zahle 1979a:318, Mühlbauer 2007a:91).
[D.S.]
1.4.2Forms
| indecl. | ne | TL 44d.12 (Xanthos) | |
| TL 44d.24 (Xanthos) | |||
| TL 44d.35 (Xanthos) | |||
| TL 44d.56 (Xanthos) | |||
| TL 55.2 (Antiphellos) | |||
| TL 55.6 (Antiphellos) |
The instance of TL 44d.56 where neburẽni is divided as ne=(a)burẽni follows Shevoroshkin (apud Neumann 2007a:237) and Shevoroshkin 2014a:255. Schürr 2001e:131f. and Melchert (DLL:122) adduce a potential additional example in alcãnalax, in which they separate alcã na lax. If correct, na would have undergone vowel harmony as a proclitic element to the potential verb lax. Unfortunately, there are different ways of interpreting the sequence; i.e. alcãna (n. pl.) lax or even alcãnalax (denominative verb to agent noun in -ala-); see under lemma.
[D.S.]
1.4.3Meaning
Lycian B ne equates to Lyc. A ne ‘not’ (Neumann 2007a:236f., DLL:122). Just as in Lycian A, the Lycian B negation ne precedes regularly the verb, e.g. ne lelixa (TL 55.2), ne mrssxati (TL 44d.24). In TL 55.6, ne does not appear immediately before the verb, but precedes the indefinite pronoun kike ‘some-/anyone’, which it modifies, i.e. ne kike ‘not any, none’.
[D.S.]
1.4.4Compounds
2. Reconstruction
2.1Reconstruction ‘not’
Comparing Lyc. ne with HLuw. /na/ and probably Ištanuwa Luw. nā, a Proto-Anatolian pre-form *ne can be reconstructed (Morpurgo Davies 1975a:158, Melchert 1994a:290). The very same negation is found in Hittite natta ‘not’, with *e > a due to erosion as a proclitic (see HEG N:288 with further literature; contra Melchert 1994a:139 and EDHIL:597 “o-grade”), nāwi ‘not yet’ (cf. Morpurgo Davies 1975a:157 n. 4, HED N:295f. with further literature, EDHIL:598) as well as in nikku < *ne=ku̯e; see Lyc. B neke and Lyd. nik.
One should also compare Pal. niwa, which contains PAnat. *ne as well.
[D.S.]
3. Proto-Indo-European
3.1Morphological Reconstruction
Phonologically, Cuneiform Luwian nāwa (2nd millennium BCE), Hieroglyphic Luwian /nawa/ and /na/ (only 1st millennium BCE?) might either be derived from PIE *ne or its o-grade form PIE *no, plus additional word-formational material. Phonologically, it is possible to reconstruct Luwian /na(wa)/ as PIE *ne-u̯e, and equations like the following support this reconstruction:
- HLuw /na/ and /nawa/ are used in the same meaning: HLuw. /na kwatti=ha/ (dat. sg.) ‘nobody’ (Aksaray §8), and HLuw. /nawa kwis=ha/ (nom. sg.) ‘nobody’ (Assur 2, §31).
- /na kwis/ is akin to OLat. nequis, Ved. na-kih (Wackernagel 1924a:250, 268, EWAia II:1) from PIE *ne ku̯is.
- Hittite natta kuwapi (UL kuwapi) ‘nowhere, never’ (HHw:88, Hoffner & Melchert 2008a:149) matches Lat. necubi from PIE *ne ku̯u-bhi.
Alternatively, Luwian /nawa/ might be reconstructed as PIE *no-u̯e (LIPP II:530, 539, 845) and it is likely to be understood as ‘not anyhow’ with indefinite similative *-u̯e or *-u̯a ‘like’ (ibid. 763). The o-grade could be attested in Hittite natta (if from *nó-th2o ‘nothing, not’); cf. Hom. νόϑος ‘nonce-offspring, bastard’ (Hackstein 2016b), with semantic shift from ‘nothing’ to ‘not’, as in OLat. noenum, nōn and NHG nicht (see Wackernagel 1924a:252–253). Also note PIE *ne-u̯e (Lat. nēve, neu ‘or not’) alongside PIE *nei̯-u̯e (OLat. neive, nīve).
3.2Syntactic Reconstruction and Phraseology
The declarative negative can also be employed for a constituent negation, in which case it is placed immediately before the negated nominal constituent; cf., e.g., Hieroglyphic Luwian /na kwin=ha/ ‘no one, none, no’ before /hatturan/ ‘letter (?)’ in wa/i-za NEG2-a kwa/i-i-ha ha-tu+ra/i-na ha-tu-ra+a ‘we are to write no letter (?)’ (Assur letters e, §5). This type of syntactic constituent negation can also be reconstructed for PIE; cf. PIE *n’eĝóh2 ‘not me’ (Hackstein 2021a).
Morphosyntactically, the negation system of the minor Anatolian corpus languages agrees with that of Hittite and reconstructed Indo-European. The major division is between syntactic negation (free morpheme PIE *ne) and morphological (word-formational) negation (*n̥-). The morphosyntax of syntactic negation varies across the three basic sentence types of declarative, interrogative and imperative clauses. The negative syntax of the syntactic negation ramifies into verbal and nominal constituent negation.
The Indo-European morphological distinction between the syntactic negation PIE *ne and the morphological negation PIE *n̥- (e.g. Ved. ájnāta- ‘unknown’, Gk. ἀγνώς ‘unknown, ignorant’, ἄγνωστος ‘unknown’, Arm. ancanawtc ‘not knowing, unknown’ (Olsen 1999a:696), Lat. ignārus ‘ignorant’, OIr. ingnad ‘strange’, Goth. unkunþs ‘ignorant’) in composition was inherited by Proto-Anatolian, as documented by Hittite ammiyant- ‘having no growth; small‚ immature’ built on the participial form of the verb māi-/-i(ya)-(ḫi) ´to grow´. However, this privative prefix seems to have become unproductive in the Hittite language.
In Luwian, the reflexes of the syntactic negation PIE *ne might have replaced the inherited morphological negation PIE *n̥-, if Luw. ni- continues PIE *ne; but cf. Starke 1990a:451f. and now Melchert 2019b:365 against pretonic *e > Luw. i and pleading for *nē > Luw. prefix /ni-/. In any case, these examples include CLuw. ni‑walla/i- ‘innocent’, and Luw. /ni-waralla/i-/ ‘alien, hostile’ and Luw. /ni-warann(i)-/ ‘the non-speaking one, child’ (see also Lyd. niqašl(i)- and niwisw(i)-). With respect to the replacement of the compositional negation *n̥- by *ne, Anatolian largely sides with Balto-Slavic (e.g. Slavic neznanije ‘lack of knowledge’), Albanian, and Tocharian (e.g. TochB. aknātsa ‘ignorant’, with analogical full grade). However, for a new potential example of *ń̥- surviving in Luwian, see ūmmaya- ‘ungrowth’ and its derivative.
The default position of the declarative negation is immediately before the inflected verb (cf. Delbrück 1897a:52, Wackernagel 1924a:260, Hoffner & Melchert 2008a:341, §26.3). A variety of subsetting rules may apply. For instance, negative existential clauses allow fronting of the negation (Hackstein 2012d).
Interrogatives distinguish between information-soliciting interrogatives and rhetorical bias interrogatives. Information interrogatives leave the negation in situ before the tensed verb. Rhetorical interrogatives typically show negation fronting (Hackstein 2016a, Hackstein 2018a).
The negation of commands requires the inherited Indo-European sentence type “prohibitive clause”. PIE forbids the negation of imperatives, demanding instead non-imperative morphology for the prohibitive clause.
[O.H.]
4. Transmission
ni … noh ‘not yet’ < *ne h2u ku̯(e) (Hackstein 2020a:19, EWAhd VI:1017)
n-í ‘(it is) not’ (Thurneysen 1980a:110, 153)
Russ. ne, Slow. ne
ne quidem, ne-quis, necesse(st) < *ne cessis est ‘there is no escape’ (Hackstein 2012d, Hackstein 2018c with literature), nē-ve, neu, OLat. neive, nīve ‘or not’ (Pokorny IEW:757), OIr. no, nu and proclitic no-, nu- hosting infixed pronouns and relative -n-, MWelsh neu- originally ‘or not?’
as, s’ < PIE *ne h2oi̯u-ku̯id (Cowgill 1960a, Schumacher & Matzinger 2013a:241 §6.5.3, 252–253 §8.2.2); nuk < *ne h2u ku̯e (Hackstein 2021a:20)
οὐκ, Hom. οὐκί (Cowgill 1960a, Schumacher & Matzinger loc.cit.)
očc < *(ne …) (ku̯)ote =ku̯id (Fortson 2018-2019a)
Toch. A and B mā from *mē (Hackstein 2017a:1313) by contamination of *mē and *ne, Toch. A māwk < *ne h2u ku̯e (Hackstein 2020a:19)
Vedic ná vā ‘or not’ = Lat. nē-ve, neu (EWAia II:1)
niwalla/i-