71696Re: [tied] dive (was Re: Sos-)
- Feb 19, 2014
---In email@example.com, <gpiotr@...> wrote:On 2014-02-15 04:59, dgkilday57@... wrote:
> As for Gmc. *xWi:taz 'white', Skt. _s'vindate_ 'shines' shows that theOnly if you practise an extreme form of cherry-picking. In fact,
> root in question can be *k^weid-.
Sanskrit very strongly supports *k^weit-. Suffice it to mention the
Vedic adjectives _s'vítna-_, _s'vitrá-_, _s'vetá-_, _s'vetyá_ 'white,
light', verbs like the aorist _as'vait_, compounds like _su:rya-s'vít-_
'as bright as the sun', etc. Nearly all of them occur already in the
Rigveda. They are supported by plenty of cognates in Iranian and
Balto-Slavic. By contrast, _s'vindate_ is a late (Dhatupatha) dictionary
root. Its only tangible attestation is the isolated aberrant perfect
_s'is'vinde_ (in a text composed in the 7th c. CE!). The _-nd-_ may well
be a Middle Indic development.
Ouch! In the face of this evidence, *k^weid- must be discarded as a phantom root. Since only *k^weit- is securely attested in Indo-Iranian and Balto-Slavic (including loans from some centum language into Slavic), it should also be assumed for Germanic.
I am not so quick to posit an /e/-grade vr.ddhi *k^weit-nó- as the source of Gmc. *xWi:t(t)a-. The zero-grade *k^wit-nó- was evidently inherited as *xWitta- 'white' in OEFris _hwitt_, MLG/MD _wit(t)_, etc., and /o/-grade *k^woit-nó- as *xWaitta- underlying *xWait(t)ja- 'wheat'. But with *xWitt- and *xWaitt- in circulation in PGmc, *xWi:tt- may have been created by analogy with inherited ablaut-sets, such as those provided by Class I strong verbs. Speakers of most dialects may have felt that */i:/ was more appropriate than */i/ for an adjective now meaning 'inherently or intrinsically albescent' rather than the passive sense 'blanched, bleached, whitened' or whatever *k^wit-nó- originally denoted. In this hypothesis, the considerable number of passive participles and deverbatives of Class I verbs containing */i/ would have given a phonesthemic quality of passivity to adjectives with this root-vowel at the PGmc stage under consideration.
Another possibility is interference with *k^weit- by the semantically close *sweid-, found in OE _switol_ 'clear, evident' (not cited by Pokorny under *sweid-(1), IEW 1042, but given by Fraenkel, LitEW s.v. _svidé.ti_). This root could have yielded Gmc. *switt- and *swaitt- by Kluge's Law beside its /e/-grade *swi:t-, providing an analogical template to form *xWi:t- from *xWitt- and *xWaitt-. Reciprocal interference of *k^weit- upon *sweid- would explain Pokorny's isolated *sweit-, listed under *sweid-(1) but having causative usage ('sengen, brennen' against 'glänzen, schimmern'), and confined to Germanic.
Incidentally, it seems quite plausible that Pokorny's *sweid-(2) 'schwitzen' (IEW 1043) developed from *sweid-(1) as a euphemism, independently in several branches of IE. "Women glisten, men perspire, poor people sweat."
Returning now to 'deep', a very good reason to accept *dHeub- as the root is the inclusion of the 'dimple' group: OHG _tumphilo_, MHG _tümpfel_, etc. 'deep place in flowing or standing water, abyss', ME _dympull_ 'depression on a smooth surface, dimple', Du. _dompelen_ 'to plunge, dive, dip'. If *dHeub- is the root, these are simply derivatives of the inherited nasalized zero-grade *dHumb-. But if *dHeubH- is the root, not only must Gmc. *deup(p)a- be produced by an /e/-grade vr.ddhi with *-nó- or some unspecified analogical process(es), but then this result must become fodder for another analogical process which treats it as a primary full grade and inserts a nasal into the zero-grade. Not that such things could never happen, but the scenario of original *dHeub- and *dHumb- is more straightforward.
Of course, if there are any PIE */b/-denialists in this group, they will simply say there is no PIE */b/. The only basis I know of for completely ousting */b/ from PIE is glottalic theory, and that theory is shot down by pre-Grimm's Law loans from Gaulish into Germanic. I cannot explain why */b/ should be the rarest PIE consonant, but some consonant had to be, and it happens to be */b/.
- << Previous post in topic