From: dgkilday57
Message: 68002
Date: 2011-08-27
>In fact, it should be dismissed. The reconstruction which I had seen was WGmc *larw(a)riko:n in the AHD, evidently a misprint for *laiw-. No connection with *larr- 'meadow' can be justified.
> --- In cybalist@yahoogroups.com, Rick McCallister <gabaroo6958@> wrote:
> >
> > ________________________________
> > From: dgkilday57 <dgkilday57@>
> > To: cybalist@yahoogroups.com
> > Sent: Friday, August 5, 2011 5:49 PM
> > Subject: [tied] Re: floor
> >
> > [...]
> >
> > > But before proceeding further, I should check what Du Cange has to say about <larricium>.
> >
> > Enough to posit a West Mediterranean *larr-. I am trying to see whether West Germanic 'lark' can be referred to a NWB compound based on this substratal word.
> >
> > *****R I don't think lark can be linked to *larr because lark is from something like *lawarkaz; see Old Spanish laverca from Gothic or Frankish. There is a Scots word in an old song laverock which I think means "lark" but it's been a while
> >
> Yes. All the attested older forms have *law- but I have seen one suggestion that WGmc earlier had *larwarkjo:n vel sim. with the first /r/ lost to dissimilation. This raises at least the possibility of a NWB compound *larr(a)-warkan- vel sim. 'meadow-defender'(??), borrowed into WGmc and reinterpreted (at least in some OE dialects) as 'treason-worker', hence folk-etymological deformation. Here the loss of the stem-vowel of the first element would be regular in NWB in this position (as I have postulated for *Der(u)-went-) and the syllable-final -rr- would be shortened to -r- before its dissimilative loss. But the devil is in the details, and unless I can come up with a phonologically concrete scenario, this connection must be regarded as highly speculative.
>