Re: elementum

1) asrk: of course r/n alternation; in any case */gw/ *in Balles'
etymology* DOESN'T have anything to do with Old Indic -k. Stop.
2) a connection between /ogn/ and /ong/ is irregular from an IE
perspective; maybe they are cognate before PIE, but a PIE metathesis
rule has counterexamples in every preserved sequence without
metathesis. If You discover pre-PIE sound laws it's extremely good,
but incompatibility with known sound-laws requires to operate with
different chronologies, so everyone is happy
3) "it seems unlikely from other ev." is no objection
4) IE alternations: if You mean morphological alternations, they have
to be used until they can't explain anything more and then one has to
reconstruct different forms for one word. If You, on the contrary,
mean phonological alternations, You are operating in a different frame
and so our discussion is useless, because we are dealing with
different scientific objects
5) asser: I've written that YOU have given an explanation, so why do
You critique that very explanation?
Morphological and phonological alternations are competing explanations
(Your sentence that morphological alternation explains less and is
more complicated is a far-reaching utterance that should be based on
much more evidence, otherwise it's just a respectable personal
opinion); if phonological explanations imply optional sound-laws
(therefore not PIE, but at best pre-PIE) they are by no way superior
to accepted morphological ones
6) -i:s: to have a possible different explanation can never mean that
any other regular explanation is false;
chronology of attestation cannot imply necessarily that newer forms
are innovations
7) word order: not 'red blood', but 'red as blood', therefore regular
like any other substantive + adjective compound

2012/10/16, stlatos <sean@...>:

>
>
> --- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
> <bhrihstlobhrouzghdhroy@...> wrote:
>>
>> */gw/ isn't required to match Skt. -k (because this alternates
>> with -n-,
>
>
> It's r/n alt., as normal. The -k appears to come from nothing, or be a
> nom. add-on only (in the mainstream); it is the result of X > q/G/h/?>0 ,
> etc., as I've covered before.
>
>
>> while Lat. -gu- doesn't); it's meant as part of the root
>> *h1egw-, found in:
>> Lat. ignis - Skt. agnih - Slav. ogn' 'fire' < *h1e/ogw-ni-
>> Toch. AB yok 'colour, hair' < *h1e:gw-ni-
>> Gk. erysibe: 'erysipela' (med.) < *h1rudh-si-h1gw-ah2
>> Gk. akri:b'e:s 'precise, exact', which implies a neuter *akri:bos <
>> *h2ak'ri-h1gw-os-.
>>
>
>
> I don't accept that rec., or all cognates above. It ignores, among
> others, * ogni- > * ongy- > unj^ = soot/rust Ar; with a K(Y) not KW since gW
>> gw > g > k (as gWixW-wó+ > keankH Ar; not * c^yankH > etc.)
> . They must also be related since:
>
> ognI OCS;
>
> is related to:
>
> oNglI OCS; : anglìs Lith; aingeal = light/fire Ir; áNgaara-s = charcoal S;
> aNAAr = fire Koh;
>
> and the two 'fire' words must also be related to each other since, just as
> above, apparent K(-Y) corr. to KY or Ky in Armenian:
>
> * anxcuL- > acuL = coal Ar; anjoG EAr;
>
> which suggest a complex cluster (w at least 2 K, gW-xY or sim.), with
> several opt. outcomes. These words all show fire > fire product (charcoal,
> etc.) and/or red/light for any semantic drift.
>
>
>> */nh/ doesn't vocalize, because */h2/ does, so */n/ is
>> consonantal, while */h1/ is onset of following /*h1gwV-/.
>>
>
>
> I don't know of any sure counterexample, but it seems unlikely from other
> ev.
>
>
>> Asins and ariwn reflect *h1osh2-(i-)ni-s and *h1esh2-r-i1/3o:n-
>> respectively and in any case are scarcely relevant for the details of
>> the etymology of sanguis.
>>
>
>
> I disagree. We can't rec. 6 (at least) dif. forms for one word,
> considering the huge amount of IE alt.
>
>
>> I completely fail to understand Your objection w/ref. to asser,
>> since an explanation of this latter from that very root is given by
>> Yourself a few line below.
>>
>
>
> I object to rec. of o/e/0 in one syl. of one word. Morph. alt. in place
> of phon. alt. explains less and is more complicated.
>
>
>> Once the alternance between *-en- and *-i: stem is accepted,
>
>
> Sometimes they're related, others not. The -i:s in L has a dif. expl.;
> thankfully the OL forms are known so there shouldn't be much doubt about its
> origin.
>
>
> an
>> alternation between *en-, *in-, and *-i- stems (still apparent in
>> Latin itself) must be accepted as well.
>>
>> In sum, none of the alleged irregularities is such, therefore
>> *h1sh2n-*h1gw-n/i(n)- is a perfect reconstruction for sanguen /
>> sangui(:)s and not a "ridiculous" one, as You politely write.
>> Is there anything else?
>>
>
>
> If the meaning was 'red blood', the order seems inconsistent with standard
> rec. Also, everything else I said.
>
>
>
>