*immō

Melchert 1985a (Melchert 2002d:230) equates Hitt. imma, CLuw. imma, HLuw. i(a)-ma with Latin immō (followed by HEG I-K:355f., EDHIL:384). Concerning the Proto-Anatolian reconstruction, we can reconstruct an adverb with geminate -mm-, i.e. immV, and a final vowel that can be either *a, *ā, *o or *ō. The last option is chosen here on Indo-European grounds; see below. Further, Melchert assumes, based on its Proto-Indo-European etymology, a case of assimilation, i.e. *id-mō > *immō, which should have occurred already at the Proto-Anatolian stage, but this is no longer necessary; see below.

[D.S.]

While the CLuw. attestations do not shed much light on the meaning of PAnat. *immō̆, the attestation in HLuw. as a postfix (‘=ever’) of a free choice pronoun meaning ‘any’ implies original status as a focus particle. Similarly, in Hittite the use as a focus particle is attested earlier than the one as a marker of free choice indefinites (‘=ever’), cf., e.g., Melchert 1985a:201 n. 20. Typologically, the development of focus particles to indefiniteness markers is well attested.Cf. Sideltsev 2018b on Hitt. =imma. Also PIE 3.*=k^u̯e and PIE 2.*=k^u̯id (LIPP II:448–451) can be reconstructed as focus particles showing very similar developments, viz. free choice postfix and usage in (affirmative) rhetorical questions which is attested for Hitt. imma, too (HW² I:49f.), cf. Opfermann 2019a:198f.It is recommended to reconstruct PAnat. *im=mō̆ as a focus particle which only secondarily developed the use as an indefinite postfix, most likely independently in Hittite and (Hieroglyphic) Luwian.
Additionally, Hitt. imma may also be used with sentential scope meaning ‘indeed, actually’ vel sim.,The CLuw. attestations may or may not be categorized under this function.cf. Sideltsev 2020a:367–378. It is typologically frequent and semantically plausible (cf. Sideltsev 2020a:362 with ref.) that focus particles with scope over the whole sentence devolop into adversative conjunctions, cf. ModE but ‘only (ptcl.); but (conj.)’ (eOED s.v. but B.2, C.III.11.a) and PAnat. *=ma ‘(emphatic particle)’ on which cf. additionally Sideltsev & Molina 2015a. This fits with the use of Lat. immō “to introduce the correction of a preceding statement, etc., or of an idea contained in it” (OLD²:919). Although both branches that continue PIE *im=mo‑h₁ attest the reading as an adversative marker, it is, due to phonological considerations (cf. below), recommended to reconstruct PIE *im=mo‑h₁ only as a not yet univerbated collocation (developing into a focus particle).

The attested geminate in Proto-Anatolian *immō̆ and Latin *immō suggests that the constituents *im and *moh₁ have not yet been univerbated in the parent language. Otherwise, the geminate would have been simplified and yielded *imoh₁ already in the proto-language, cf. LIPP II:527 n. 10 and LIPP I:92f.The assumption of assimilation of *‑mH- > *‑mm- holds only for Anatolian (cf. Melchert 1994a:79–81) but not for Latin, cf., e.g., the continuants of the root PIE *demh₂- ‘to (op)press’ (cf. PAnat. *dmaH-s- ‘to (op)press’): Lat. domāre ‘to tame’ vs. Hitt. dammisḫa- ‘oppression, forcing’. The phonological option *‑dm- > *‑mm- (as per Melchert 1985a:202–205) works for Hittite (cf. Kimball 1999a:198f.) but the Luwian evidence is difficult (cf. Melchert 1994a:160, Melchert 2002d:229; phonetic erosion, as mentioned above sub PAnat. Reconstruction, cannot be proven) and, furthermore, there is only weak evidence from Latin (cf. Weiss 2020a:181). “Expressive gemination” (as per, e.g., EM:310) must be rejected because it is rather ad hoc and had to be assumed for both branches independently.It is thus recommended to reconstruct PIE *im=mo‑h₁ as a (habitual) collocation of *im and *moh₁.In view of this situation, it is not impossible that PAnat. *immō̆ and Lat. immō go back to different collocations with different second constituents: pre-PAnat. *im=ma (cf. PAnat. *=ma ‘(emphatic particle)’) and pre-PItal. *im=moh₁. The reconstruction pre-PAnat. *im=m(i̯)a is more likely in terms of internal reconstruction because the particle *=m(i̯)a is attested in Anatolian (cf. PAnat. *=ma ‘(emphatic particle)’) which is not the case with *moh₁ ‘somehow’.

Lat. immo is virtually always found at the beginning of a sentenceThe most prominent exception to this is the construction Lat. #quīn immo. However, this collocation does not appear earlier than the 1^st c. CE (Plin., Quint., etc.) which weakens its evidential power. Furthermore, Lat. #quīn immo may be explained as a case of semantic reinforcement since Lat. quīn (in non-pronominal use, cf. OLD²:1712) is nearly synonymous with immō.which shows that it is not enclitic and thus not a particle focussing a specific word in the sentence but a focus particle with sentential scope. According to Sideltsev 2020a, the same sentential scope can be observed in the case of Hitt. imma which may, with due caution, also be reconstructed for PAnat. and, taking into accout the Latin state of affairs, for PIE *im=moh₁, if it functions as a particle already in the proto-language.

[A.O.]

Lat. immō ‘(a particle to correct a previous statement, hence) rather, more correctly’ (Plaut.+)
WH I:632f., EM:310, EDLIL:300, OLD²:919f.

PIE pron. abl./instr. *i‑m ‘thence; thereby’, cf. (?)Ved. emph. ptcl. īm (EWAia I:205, Grassmann 1996a:231–233), Lat. in=de (< *im=de) ‘thence’, etc. (WH I:692, EM:315, EDLIL:162, OLD²:967, Opfermann 2016a:230–237), (?)Osc. =en, =en ‘from’ (Untermann 2000a:223–225, esp. 224, Opfermann 2016a:237 n. 676 with lit.)
On the deictic pronominal stem PIE *í‑, cf. LIPP II:363–374, on the form PIE 2.*im, cf. LIPP II:367

On the interr./rel./indef. pronominal stem PIE *mó‑, cf. LIPP II:518–523, on the (alleged) particle PIE *mṓ, *=ma, cf. LIPP II:526–528 and PAnat. *=ma ‘(emphatic particle)’