What semantic notions underlie PIE *meh₂d- ('wet') and Proto-Germanic *matōną, *matjaną (“to feed, eat”)?

Question

I was reading the etymology of amadouer when I lighted on these attested morphemes:

Etymology

From Middle French amadouer (“to coax, lure”), from a- + *madouer (“to lure, give food to”), from Old Norse mata (“to lure, feed”), from Proto-Germanic *matōną, *matjaną (“to feed, eat”), from Proto-Indo-European *meh₂d- (“wet”). Cognate with Icelandic mata (“to bait, allure”), Danish made (“to feed”), Gothic 𐌼𐌰𐍄𐌾𐌰𐌽 (matjan, “to eat, have a meal”), Old English mete (“food”). More at meat.

1. Please see the titled question. How did *meh₂d- semantically shift to mean *matōną? What does eating or feeding have to do with wetness? I know that people drink while eating, but I'd reckon that PIE and Proto-Germanic had verbs for drinking?

2. Which type of semantic shift is this?

Answer 1

Kroonen's dictionary says, following [Lubotsky (1981)]

*matjan- w.v. 'to lap' -- [...] In spite of Torp 1909: 305 and Pokorny IEW: 694-5, unrelated to *mati- 'food' (q.v.)."

The only Germanic reflex given is an old Norse term "to suckle, nurse", agreeing with Albanian evidence and agreeing not so much with the semantics in Sanskrit "to rejoice" in relation to intoxication," and Greek "to drip, melt, ...".

They reconstruct respectively PIE *mh2-d- and tentatively *med-; The later is found in wiktionary under "measure".

*mati- has more explanation. I am not going to quote the whole thing, if I only understand half of it. Follow the sources (Lubotsky 2012 also has some on water):

• Lubotsky, A.M. 1981 [Lubotsky 1981] Gr. πήγνῡμι: Skt. pajra- and loss of laryngeals before mediae in Indo-Iranian. Münchener Studien zur Sprachwissenschaft 40, 133-138

• Pokorny, J. 1959 [Pokorny IEW] Indogermanisches etymologisches Worterbuch. Bern.

• Torp, A. 1909 [Torp 1909) Wortschatz der germanischen Spracheinheit. Vierte Auflage. Göttingen.