*nun-

What regards the further etymological connection of the numeral, it is frequently assumed that it is connected with the word for ‘new’, PIE *néu̯‑o- ‘new’.“Unless ‘nine’ can be connected with ‘newness’ […], it remains unanalyzable” (Winter 1992a:17).One of the proposals to connect ‘9’ with ‘new’ is that PIE *néu̯‑n̥ is a locative of an r/n‑stem PIE *néu̯‑r̥/*néu̯‑n̥- ‘new’, thus originally meaning ‘in the new’.LIPP II:582 reconstructs an original accusative *(én) néu̯‑o‑m ‘(to) the new’ with subsequent (irregular) syncope and assimilation to the dental anlaut of PIE *dék̑m̥ ‘10’, thus *(én) néu̯‑n̥. Alternatively, the ending *‑m may be interpreted as the instrumental-ablative adverbial ending *‑m (LIPP I:137–148), which would result in the meaning ‘with/in/from the new’.Such a semantic etymology may be in accordance with the interpretation of the numeral ‘8’ as an original dual formation PIE (?)*h₂ók̑t‑o‑h₁ ‘both hands without the thumbs’ (cf. Lyc. A aitãta ‘eight(een)’), which points towards a kind of quaternary counting system for PIE. However, since this morphological and semantic analysis of the numeral ‘8’ and its consequences on the counting system is only tentative, there is no good external evidence for the analysis of the numeral ‘9’ as ‘new’.On the morphological caveats, cf. below.
Blažek 1999a:285–287 reconstructs as a preform the u‑stem acc. sg. (phrase) PIE *(en) h₁néu̯‑m̥ ‘(one is) in lack’, thus proposing an arithmetical explanation for the numeral ‘9’ (= 10–1). Besides the unexpected locative semantics of the accusative (not †‘to/into the lack’), this solution poses various morphological problems, on which cf. below.
On these and various other proposals for the semantic reconstruction of the PIE numeral ‘9’, cf. Blažek 1999a:283–287.

The exact phonological and thus morphological shape of the numeral ‘9’ is hard to determine for two reasons. Firstly, Gk. ἐννέα and Arm. inn ‘9’ show both a kind of prothetic vowel that may go back to *h₁- or *e‑. In view of the geminate ‑νν- in Greek,The Armenian geminate is not inherited but a result of syncope: *inown (cf. the marked plural inown‑k‘ ‘9’) > inn, cf. Winter 1992b:350 and EDAIL:302 for further proposals that only differ in other details.it may be assumed that the numeral goes back to a prepositional phrase *en=néu̯m̥ ‘to the new’ (cf., e.g., LIPP II:582). If *h₁- is reconstructed to cope the Greco-Armenian anlaut, the connection with PIE *néu̯‑o- ‘new’, which is doubtful anyway (cf. above), can hardly be upheld.The local adverb PIE *én ‘in’ does not contain a laryngeal (not †*h₁én, cf. LIPP II:221). Peters (in EWAia II:25) proposes analogical influence from ‘8’ (if *h₁ok̑t‑) although the anlaut of ‘8’ may rather be *h₂‑, cf. Lyc. A aitãta ‘eight(een)’. Alternatively, it may be assumed that *h₁ is transferred (or, rather, copied) from the dual ending of ‘8’ in allegro counting: PIE *…=h₂ók̑to‑h₁=néu̯‑n̥= “…8, 9…” → *…=h₂ók̑to‑h₁=h₁néu̯‑n̥=. However, this solution, while in principal supported by other kinds of assimilation in numerals (cf., e.g., the anlaut *d- of most of the Balto-Slavic forms of ‘9’), is ad hoc.
On the other hand, also *‑m̥ → *‑n̥ can be explained, viz. by assimilation to the following dental of *dék̑m̥ ‘10’ (cf. LIPP II:582, Blažek 1999a:287 with ref.). Both nasals may have a possible source in morphology: *‑m̥ can be the acc. sg. c. (cf. LIPP II:582, Blažek 1999a:287) or the instr.-abl. adverbial ending (LIPP I:137–148) and *‑n̥ can be seen as a nasal suffix with zero ending, viz. loc. sg. (or even neuter nom.-acc. sg.).
As per Winter 1992a:13f. and others, the connection with an alleged r/n-stem PIE *néu̯-r̥/*néu̯‑n‑ ‘(?)new(ness)’ is rather unlikely for various reasons (on the semantics, cf. above), not least because Gk. νεαρός ‘young’, Arm. nor ‘new’ as the only evidence for the stem *néu̯‑r̥ had to be explained by derivation from the strong stem, which is unexpected.The putative parallel PIE *ud‑r‑ó- ‘living in the water’ (← PIE *u̯ód‑r̥/u̯éd‑n- ‘water’) is an exocentric derivative, a property not to be assumed for Gk. νεαρός ‘young’ and Arm. nor ‘new’.Furthermore, the proposed endingless loc. sg. *néu̯‑n̥-Ø ‘in the new’ shows unexpected ablaut – not †n(u)u̯‑én‑Ø – in comparison with the adduced parallel Ved. ud‑án‑Ø ‘in the water’ (cf. PIE *u̯ód‑r̥/u̯éd‑n- ‘water’).According to NIL:525, the Greek and Armenian lexemes are not derived from an athematic r/n-stem, but directly from the o-stem.While metathesis *nu̯‑én → *néu̯n̥ is always possible, such a solution is ad hoc.LIPP II:582 proposes an allegro form *(en)=néu̯‑n̥ ← (assimilation to the anlaut of *dék̑m̥ ‘10’) *(en)=néu̯‑m̥ ← (syncope due to allegro) acc. *(én)=néu̯‑o‑m ‘(to) the new’. While the preposition *en demands the accusative (or locative), the ending *‑m may also be interpreted as instrumental-ablative (LIPP I:137–148), viz. ‘with/in/from the new’.Lastly, there is, contra NIL:524, no “root” PIE †neu̯- ‘new’, that could serve as a derivational base: PIE ‘new’ is a (thematic) vr̥ddhi-formation derived from the particle PIE *nú(n) ‘now’, cf. NIL:525 n. 1, s.v. PAnat. *néu̯‑o- ‘new’. A metanalyzed, secondary root PIE †neu̯- ‘new’ is not reconstructable. As Winter 1992a:14 concludes, there is not enough clear evidence in favor of the connection of the PIE words for ‘9’ and ‘new’.
Blažek 1999a:286f. proposes a connection with Goth. inu ‘without’, Gk. ἄνευ ‘id.’, etc., which he traces back to a u-stem PIE *h₁énu‑/*h₁néu̯- ‘lack’. This reconstruction, however, is based only on very weak evidence and therefore not followed in this dictionary.On the prepositions mentioned, cf. the lemma *énu ‘along; behind, after’ in LIPP II:241–244.

[A.O.]