Re: Etymology of _insula_ (was Re: Thoughts on Word building)

On Thu, 8 Dec 2005 08:10:01 +0000, R A Brown <ray@...>
wrote:

>Rob Haden wrote:
>> On Wed, 7 Dec 2005 21:20:17 +0100, =?iso-8859-1?Q?J=F6rg?= Rhiemeier
>> <joerg_rhiemeier@...> wrote:
>[snip]
>>>
>>>Irish has _inis_, Welsh has _ynys_, and Latin _insula_ looks
>>>like a diminutive of a similar word (perhaps _insula_ < *inis-ula?).
>>>So this looks like an Italo-Celtic etymology, but there may be
>>>problems with this I do not see.  Ray?
>
>As you will have seen, I quite independently wrote in with the some
>etymology   :)
>
>>>=========================================================================
>>
>>
>> AFAIK, *inis-ula would have given Latin *inirula or *inilla (< *inirla).
>
>That assumes that the original was *inisula and, presumably, that the
>second 'i' was long. The comparison with the Celtic forms does not AFAIK
>demand this.
I suppose the second 'i' would indeed have to be long; otherwise, it would
have been lost to syncope in Latin.

>In his "An Etymological Dictionary of the Gaelic Language", Alexander
>McBain gives the etymon of the Celtic forms as *n=ss.
Is that a syllabic /n/?  If so, why is the whole form zero-grade?  I'm
assuming that the IE etymon (if there was one) would mean 'not (something)'.

>This would give Latin *inss- and thus the (originally) diminunitive would
>be ins(s)ula.
Sure, that could work.  Many other words in Latin traditionally spelled
with a single intervocalic <s> originally had <ss>, e.g. <casus> < <cassus>
(from IE *kadtos).

>Now, I am not an expert in Celtic etymology, so I cannot tell how sound
>McBain's etymology is. But if he is correct about *n=ss, then there is
>no problem with the Latin form as far as I can see.
Where did the medial vowel in the Irish and Welsh forms come from?

>What I find less convincing is relating this stem to the Greek forms
>that I cited in my last mail.
>--
>Ray
I'll have to look those up. :)

- Rob

Reply