*mé-moh₁-/-mh₁-

Considering its original inflection as a stem in -āi-/-i(ya)-^(ḫi) with reduction to -a-/-i(ya)-^(ḫi) (as expected for reduplicated formations; see Oettinger 1979a:463, Kümmel 2012b, Sasseville 2021a:428), the Proto-Anatolian root of Hitt. mēma-/-i(ya)-^(ḫi)‘to speak’, Pal. memiya- ‘to speak (?)’, Išt. Luw. mammiya-⁽ⁱ⁾ ‘to speak (?)’ and Lyc. B mẽmiya- ‘language’ should be reconstructed as *meh₁-. The reduplication found for each cognate in the various Anatolian languages allows for the PAnat. reconstruction of the following derivational chain: *móh₁-/mh₁-´ → *mé-moh₁-/-mh₁- >> mé-moi̯h₁/-mih₁- (for the transfer, see Sasseville 2021a:429-437). Kloekhorst (EDHIL:574) suggests a connection with the root *h₁em- ‘to seize’, which is morphologically and semantically unconvincing. Contra Kloekhorst (EDHIL:581f.), Hitt. mimma-/mimm-^(ḫi) must come from the root men- on account of its geminate -mm- and its conjugation -a-^(ḫi). Further, its lack of transfer to the stem class in -āi-/-i(ya)-^(ḫi) in Proto-Anatolian speaks against a root-final *h₁ (see Sasseville 2021a:429-437). For Indo-European etymologies, see Jasanoff 2003a:118, HEG M:188 and below.

[D.S.]

Hitt. mēma-/-i(ya)-^(ḫi) ‘to speak’ is sometimes seen as onomatopoeic in origin (cf., e.g., Jasanoff 2003a:117f.), which may be supported by the fact that it “is the verb marking foreign language” (HED M:139).The etymological connection with PIE 1.*men- ‘to catch a thought’ (LIV²:435–437, LIV²⁺ s.v. 1.*men- [2019-04-17]) supported by Puhvel (HED M:139f.) is refuted on phonological and morphological grounds by EDHIL:575 and others.However, the claim that Hitt. mēma‑/‑i(ya)‑^(ḫi) and CLuw. mammiya‑⁽ⁱ⁾ ‘to speak’ go back to a onomatopoeic root (or formation) is not compelling, since the allegedly underlying form is not straightforwardly reconstructable and other, more plausible solutions are available.Cf., e.g., the comments on Ved. mā- ‘to moo’ above sub Transmission of PIE *meh₁- ‘to measure’.
According to Sasseville 2021a:429–437, the morphologically most normal derivation would be from a root PIE *mei̯h₁- or secondarily, as is opted for above (sub PAnat. reconstruction), from PIE *meh₁‑. While the former is attested as PAnat. *mei̯h₁- ‘soft’, PAnat. *mói̯h₁‑/*mih₁‑´⁽ⁱ⁾ ‘to grow’ and does not fit semantically at all, the latter is reconstructed with the meaning ‘to measure’ (LIV²:424f.). As is shown by Iranian continuants like Budd. Sogd. pr’m’y, etc. ‘to command, say’, the development of verba dicendi out of verba metiendi can indeed be observed (cf. EDIV:254–257 passim).The semantic path probably is ‘to measure’ > ‘to apportion, allot’ > ‘to command’ > ‘to say’; cf. for the last steps PAnat. *mér-/mr̥-´^(ti) ‘to order, request’.However, a quite remarkable number of intermediary steps is necessary to arrive at ‘to say, state’, the meaning that is attested for the Anatolian lexemes. Furthermore, no traces of the alleged more original root semantics (‘to measure’, ‘to allot’, ‘to command’) can be observed in Anatolian.
Another root is mentioned by Kloekhorst (EDHIL:575) as possibly underlying: PIE 1.*(H)mei̯- ‘to fix, fasten’ (LIV²:426), which may develop via ‘to state’ to ‘to say’. This semantic development is not only paralleled in Anatolian itself in Hitt. tē‑^(mi) ‘to speak, state’ vs. Lyc. A ta-^(di) ‘to lay down’ [lemma link] (: PIE *d^heh₁- ‘to put, place; to do, make’, LIV²:136–138), but also typologically well established, cf., e.g., Buck 1915a:129–137 passim. The only minor semantic caveat is the lack of cognates meaning ‘to state, speak’ outside Anatolian.
Nevertheless, the semantic development ‘to fix, fasten’ (*1.(H)mei̯‑) > ‘to state, speak’ is more straightforward and intuitive than ‘to measure’ (*meh₁‑) > ‘to allot’ > ‘to command’ > ‘to say’.On the morpho(no)logical issues of the root PIE 1.*(H)mei̯- ‘to fix, fasten’, cf. below.

The reduplicated formation reconstructable for Proto-Anatolian originally belongs to the ḫi-conjugated stems that end up in the class of verbs in Hitt. ‑āi‑/‑i(ya)‑^(ḫi), etc. (cf. PAnat. reconstruction above). This means that the underlying root prototypically has to be reconstructed as *mei̯h₁‑, which is attested as PAnat. *mei̯h₁- ‘soft’, PAnat. *mói̯h₁‑/*mih₁‑´⁽ⁱ⁾ ‘to grow’, but is semantically unattractive (cf. above). The analogically remodeled type of these verbs consists of roots ending in *‑h₁, cf., e.g., Luw /hishiya‑⁽ⁱ⁾/ ‘to bind’ : PIE *sh₂eh₁- ‘to bind’. Thus, above sub PAnat. reconstruction the root PIE *meh₁- ‘to measure’ is proposed as underlying the current lemma, yet with the semantic issues emphasized above.
EDHIL:575 (followed by LIV²⁺ s.v. 1.*mei̯- → *Hmei̯-) alternatively proposes the connection with PIE 1.*(H)mei̯- ‘to fix, fasten’ (or ‘firm, fixed’ as per above). While this etymology is not supported by parallel cases of the same stem formation (*Ce-Coi̯‑), there are no counter-examples, either.The verb Luw. /dwaya‑⁽ⁱ⁾/ [lemma link], if from PIE *du̯ei̯- ‘to fear’, is the only example for a ḫi-verb formed to a root of the structure *Cei̯‑ but shows no reduplication and no accent retraction as is necessary for Hitt. mema‑/‑i(ya)‑^(ḫi) ‘to speak’, etc. The accent on the reduplication syllable most likely induced the change of stem class, cf. Sasseville 2021a:428f. with refs.Therefore, this semantically more attractive solution should be put up for discussion.

PIE
PIE *méh₁-ti‑/*mh̥₁‑téi̯- ‘measure, measurement’, cf. OInd. māti- ‘id.’ (Lex.) (EWAia II:341–343, KEWA II:638, Grassmann 1996a:154), Gk. μῆτις ‘id., skill, craft’ (Il.+) (GEW II:232f., GEW III:153, DELG:699, EDG:948f., LSJ:1130), Lat. *mēti- ‘measure’, continued in Lat. mētīri ‘to measure’ (WH II:81f., EM:401, EDLIL:377, OLD²:1216), OE mǣd ‘measure’ (EDPG:367)
IEW:703f., EIEC:374

PIE *méh₁-n‑os‑/*mh̥₁‑n‑s‑´ vel sim.The PIE stem of the word for ‘moon’ and its einzelsprachlich developments is a complex matter, cf., among others, Steer 2015a:270f., EDAIL:48, EWAhd VI:127f.‘moon, month’, cf. Ved. mās- ‘moon, month’ (RV+) (EWAia II:352f., KEWA II:631f., Grassmann 1996a:1036f.), Av. māh-, OPers. māh- ‘moon, month’ (Bartholomae 1904:1170f., Schmitt 2014a:208f.), Myc. gen. sg. me-no, Ion., Dor. μήν, Att. μείς ‘month; moon-sickle’ (GEW II:227f., GEW III:153, DELG:695f., EDG:945, LSJ:1093f.), Arm. amis ‘month’ (EDAIL:47f., Olsen 1999a:48f.), Lat. mēnsis ‘month’, Umbr. abl. sg. menzne ‘moon(?), month(?)’, (?)Osc. dat./abl. pl. minnaris ‘(?)lunar’ etc. (WH II:71f., EM:398, EDLIL:373, OLD²:1209, Untermann 2000a:471f., Poccetti 2016a), OIr. m. (→ f.) mí, Gaul miđ, etc. ‘month’ (EDPC:272, LEIA-M:46, de Bernardo Stempel 1999a:148, eDIL s.v. mí [http://www.dil.ie/32120] [2021-03-22], Delamarre 2003a:226f.), Goth. mena, ON máni, OHG māno ‘moon’ (EDPG:365, EWAhd VI:125–129, de Vries 1977a:378, GED:251), Goth. menoþs, ON mánaðr, OHG mānōd ‘month’ (EDPG:365, EWAhd VI:129–131, de Vries 1977a:378, GED:251), Lith. mė́nuo, gen. sg. mė́nesio, etc. ‘moon, month’, Latv. mẽness ‘id.’, OPr. menig ‘moon’ (EDBIL:311f., ALEW 2.0 s.v. mė́nuo [2021-03-22], LEW I:438f., Karulis 1992a:582, PKEŽ² s.v. menig [2021-03-22]), OCS měsęcĭ ‘moon, month’, Russ. mésjac ‘moon’ (EDSIL:312f., Vasmer II:125), Alb. muaj ‘month’ (Schumacher & Matzinger 2013a:222, 261, Demiraj 1997a:279f., AED:276, Huld 1984a:96), TochA mañ, TochB meñe ‘moon, month’ (DTB²:503)
IEW:731f., EIEC:385, Kölligan 2018b:2253

Indo-Iranian
Ved. pres. act. mímā‑, midd. mímī- ‘to measure, allot’ (RV+), etc.
EWAia II:341–343, KEWA II:638, Grassmann 1996a:1023–1025Ved. mā- ‘to bellow, roar, bleat’ (RV+) (together with OCS mĭmati ‘to stammer’ and others, cf., e.g., EIEC:394) is usually connected with Hitt. mēma-/-i(ya)-^(ḫi) ‘to speak’, etc. and considered to be onomatopoeic, cf., e.g., Jasanoff 2003a:117f., Kümmel 2000a:377f. Phonologically, this may be possible, but semantically it is problematic, because the Vedic verb is used of cows and therefore rather means ‘to moo’. This suggests that the verb goes back, if onomatopoeic, to a “root” pre-PIIr. */moː/ ‘to moo’ (> Ved. /mā/) which fell together with possible PIE */maː/ and */meː/ that may be onomatopoeic roots for ‘to bleat, maa, baa (of sheep, goat)’ (cf. OED online s.v.). There is a strong cross-linguistic tendency that the sound imitating a cow and thus the verb ‘to moo’ (if onomatopoeic) consists of a labial consonant (mostly /m/) and a long rounded vowel (mostly /uː/, but also /oː/ or /œː/) while the same for sheep and goats also consists of a labial consonant (/m/ or /b/), but followed by an open and unrounded vowel (/aː/ or /εː/), cf. the list in Wikipedia. The Ved. root mā- may thus be traced back to a pre-PIIr. onomatopoeic root *mō- ‘to moo’ (not **moh₁‑ or **meh₃‑!), which may be contaminated with *mā- or *mē- ‘to maa, baa’ and cannot serve as a source of PAnat. *mé-moh₁‑/‑mh₁‑⁽ⁱ⁾ ‘to speak’, cf. above sub PAnat. reconstruction for the need of the laryngeal in this case. For a proposal to explain the various stems of Ved. mā- ‘to moo, (?)maa’, cf. Kümmel 2000a:376–378.
OAv. fra-mīmaθā ‘you command’ (hapax, Y.32.4), OPers. fra-mātar- ‘ruler’, Budd. Sogd. pr(’)m’y, etc. ‘to command, say’
EDIV:254–257, Bartholomae 1904:1165f., Schmitt 2014a:176, Gharib 1995a:279, 288

Greek
Gk. μέτρον ‘measure’ (Il.+), μήτρα f. ‘(pl.) register of house-property; (Gk.) κλῆρος, allotment (Hsch. μ 1291 Latte)’Another meaning given by Hesychius is the following: τῶν ξύλων τὸ ἐντός, ὅ καρδίαν τινές, ἢ ἐντεριώνην καλοῦσι. “The inner part of logs that some call καρδία (‘heart[‑wood]’) or ἐντεριώνη (‘heartwood’).” This notion reminds of dendrochronology, hence time measurement.
GEW II:220f., EDG:939f., DELG:691f., LSJ:1123, 1130

Italic
(?)mēnsa ‘table’, Umbr. acc./abl. sg. (?)mefa, mefa, loc. sg. mefe ‘round flat bread/cake’
WH II:70, EM:397f., EDLIL:372, OLD²:1209, Untermann 2000a:463f.

Germanic
(?)PGerm. mēla- n. ‘point in time’, cf. Goth. mel n. ‘hour, time’, ON mál ‘(meal)time’, OHG māl ‘time, moment’; PGerm. mēla- m. ‘measure, amount’, cf. Goth. mela m. ‘bushel’, ON mál ‘measure’,
EDPG:362, EWAhd VI:62–64, de Vries 1977a:376, GED:250

Albanian
Alb. mat ‘to measure; to hesitate’, mot ‘period of time, year, weather’
Demiraj 1997a:258, 278, AED:246f., 274f., Huld 1984a:95, Schumacher & Matzinger 2013a:587f.

(?)Tocharian
TochA me‑ ‘to gage’, TochB mai- ‘to gage, gauge, estimate’, TochB maim ‘thought, thinking’
Malzahn 2010a:776–778, DTB²:506f., 507f.

PIE
PIE *mEi̯‑to- ‘pole’, cf. ON meiðr ‘tree’, OHG wolla-meit ‘distaff’ (8th c. CE) (EDPG:347, EWAhd VI:—, de Vries 1977a:381, Magnússon 1995a:611, ChWdW8:208), (?)Lat. mūtō/muttō, ‑ōnis ‘penis’ (Lucil., Hor.) (WH II:138, EM:426, EDLIL:398, OLD²:1266),The variant Lat. mŭttō is certainly a result of the Iuppiter-rule (V̄C > V̆CC, on which cf. Weiss 2020a:155). EDLIL:398 reconstructs Lat. mūtō as derived from *mHú-to- and connects it with Luw. /muwa‑⁽ⁱ⁾/ ‘to overpower’ [lemma link], etc.(?)Lith. miẽtas ‘pole, fence post’, Latv. mìets ‘post, setting/hop/bean-pole’ (EDBIL:—, LEW I:451, SEJL:398, Karulis 1992a:592f., Ulmann 1872a:155)
IEW:709, EIEC:441, Steer 2008a:150–152, 154f.

PIE *mi-tó- ‘fixed’, cf. Ved. ní-mita- ‘fixed’ (RV+), sú-mita- ‘well fixed (of a pillar)’ (hapax, RV.5.45.2) (EWAia II:314, KEWA II:636f., Grassmann 1996a:1037f.), YAv. bərəzi-mita- ‘highly built’ (Bartholomae 1904:961, Scarlata 1999a:384f.), Gk. μίτος ‘yarn of the warp (Il.+); thread of destiny (Lyc.+)’This may easily be explained as *mí-to- ‘what is fixed’ (substantivized from *mi-tó- ‘fixed’ by accent shift) because the yarns of the warp are fixed in the loom while the weft is loose and to be threaded through the warp.(GEW II:245f., GEW III:153, DELG:706, EDG:958f., LSJ:1138, Blümner 1912a:141f., 149), (?)Lith. mità ‘lucet (for producing nets); fishing net float’ (ALEW 2.0 s.v. mità [2021-03-16], LEW I:461, SEJL:398)
IEW:709, Steer 2008a:152f.

Indo-Iranian
Ved. minóti ‘fixes, fastens’ (RV+), (?)sū-máya- (hapax, RV.8.77.11) ‘well fashioned’ or ‘made of (boar’s) tusk’, mít‑ ‘post, pillar’ (hapax, RV.10.18.12), °mít- ‘id.’ (in compounds), methí- ‘pole’ etc.
EWAia II:314, KEWA II:636f., Grassmann 1996a:1037f., 1038, Steer 2008a, Scarlata 1999a:384f., Jamison & Brereton 2020a comm. ad RV.8.77.11 (2021-03-12)

Armenian
(?)Arm. moyt‘* ‘pillar’ (Bible+)
EDAIL:—, Olsen 1999a:81f.

Celtic
OIr. do-dímen ‘to attach, fix’
Schumacher 2004a:484, EDPC:271, LEIA-D:131f.

MIr. methas (/methus) ‘frontier’
EDPC: —, de Bernardo Stempel 1999a:418, LEIA-M:45

Balto-Slavic
Latv. méju (mìet) ‘to impale, drive in a stake’
Karulis 1992a:592f., Ulmann 1872a:154