A really cautious etymology would be something like *X-g^Huh2. As for the the initial X-, we have the following reflexes:
(1) Slavic has *(j)eN(zy-kU), as if from *In- < *n.- or *eN-;
(2) in Baltic X = *lei- (Lith. liez^uvis), cf. Arm. lezu, Class.Lat. lingua, all perhaps influenced by the 'lick' verb;
(3) Indic ji- (jihva:) ~ ju- (RV juhu:), as if from reduplicated (pre-Grassmann) *g^Hi-g^Huh2 or *g^Hu-g^Huh2;
(4) Iranian *hi- (Av. hizva: ~ hizu:, OP hizba:na:), as if from *si-; as in the case of Baltic and Slavic the striking thing is that Indic and Iranian should yield conflicting reconstructions;
(5) Geg gûh, Tosk gluhë suggest Pre-Albanian *gln.- or *dln.-;
(6) Celtic has *tan- (OIr. tenge), as if from *tn.-;
(6) Tocharian A käntu, B kantwo points to *tn.g^Hw- or *dn.g^Hw- metathesised to *g^Hn.tw- or *g^n.dHw-. Of these, *tn.g^Hw- would be compatible with the Celtic form, while *dn.g^Hw- is compatible with Old Latin dingua and Germanic *tungo:n-. The reconstruction *dn.g^Huh2 can therefore be regarded as the "best fit" since it accounts for reflexes in three branches.
(7) Greek and Anatolian have nothing interesting to offer in this case.
Rather than reconstruct a hypothetical monster onset like *sdln.- plus ad hoc simplifications to get the attested forms, I prefer to regard the 'tongue' word as an ancient compound, possibly with *-g^Huh2- from *g^Heuh2- 'call, invoke' (cf. Skt. indra-hu: 'calling Indra' and other similar compounds) and with _various_ first elements. The structure of the compound has been obscured in most cases and one can only speculate about the individual cases (I'm inclined to assume genuine reduplication in Indic and perhaps *hu-zu: (dissimilated to *hizu:, cf. Vedic su-hu:- 'well-invoking'). Folk-etymological association with *leig^H-/*ling^H- ('tongue' = 'licker') accounts for the frequent occurrence of *l.
Piotr
----- Original Message -----
From: "Glen Gordon" <glengordon01@...>
To: <cybalist@yahoogroups.com>
Sent: Sunday, December 22, 2002 2:04 PM
Subject: [tied] Lick me (was: the tongue)
>
> Piotr:
> >The PIE word was *dn.g^Huh2 (with irregular treatment of the initial in
> >several branches). [...] The word was certainly influenced in a
> >folk-etymological manner by (formally unrelated!) <lingo:> 'lick' (PIE
> >*leig^H-).
>
> Right, but aren't there also forms demonstrating *ngHux too?
>
>
> - gLeN