Re: sum

--- In cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@...>
wrote:

> > So now both structures, the commonly reconstructed *H1és-mi and
> the
> > Schmalstieg-inspired *H1só(:)mi posited in honour of Latin sum,
> are
> > possible morphological structures in PIE. Then the question is,
> > (what was the form of "I am" in PIE?
>
> What was the form of "I am" in Victorian English?
>
>
> >Now, "I am" is (1) Gothic im,
> > 2) Old Irish -amm (Gaulish imi or immi), (3) Old Lithuanian

esmi,

> > OCS jesmI, (4) Albanian jam, (5) Armenian em, (6) Greek eimí,

(7)

> > Sanskrit asmí (Av. ahmi, OP amiy), (8) Hittite esmi. Tocharian
> uses
> > a different word, ad Latin and Oscan have sum. So, out of the

nine

> > branches that have diagnostic material, eight point to *H1esmi,
> and
> > one looks like *H1somi.
>
> Two. You left out Serbo-Croat and Slavic Macedonian 'som'. And now
> your proposal are exposed to the same criticism as mine: How did
> those similar forms arise in several languages?

They arose as reduced forms in South Slavic. Serbo-Croatian has full
forms jesam jesi jest jesmo jeste jesu and enclitic sam si je smo
ste su. Slovene has only the reduced forms sem si je smo ste so. The
Freisinger Denkmäler (10th cty. Slovene) have several occurrences of
iesem. The vowel in Slov. sem (which is a schwa), SbCr. sam and
Bulg. sUm is the regular propvowel which appears when clusters move
to final position by loss of a final reduced vowel (as in the name
Peter, Petar from PetrU). The Macedonian -u- in sum (sic) may show
secondary influence from thematic -u (OCS -oN). There is no high
antiquity about sem/sam/sum/sUm. The story of its development is
strikingly parallel to that of Latin sum.

> >Of the eight votes for *H1esmi, some at
> > least (1-6) are from languages in which the very form /esmi/ is
> > itself a gross irregularity.
>
> A gross irregularity in the respective languages, but in PIE it
> would not be rocket science to construct a *h1es-mi from *h1es-si
> and *h1es-ti.

I grant you that.

> >That already makes it compelling to
> > posit a protoform of the structure *H1esmi, provided there is

any

> > conceivable avenue by which this form can turn into Latin sum.

And

> > of course there is, there are probably many, but the shortest

way

> is
> > of course adjustment to the partly de-thematicized allegro-based
> > outcomes of *bhero: bheresi etc. and subsequent internal
> adjustment
> > of its alternants to get the same stem-form in sum as in sumus
> sunt,
> > because these forms go together in fero fermus ferunt and edo
> edimus
> > edunt. When that is done, *edo(:)mi goes on to become edo: after
> > fero: (and lego: etc.), while *som(:)i stays like that and
> develops
> > regularly into sum.
>
> Alright. So you have
> *h1esmi
> *h1essi
> *h1esti
> and
> fero:
> fersi
> ferti
> and you come up with
> somi
> essi
> esti
>
> I don't get it.

I do. First, it was
*esmi *essi *esti *smos *stes *senti
*fero: *feressi *fereti *feromos *feretes *feronti

which became (leaving put * for ease):
esmi essi esti smos stes sonti
fero: ferssi ferti feromos fertes feronti.

"Be" takes over endings from "carry", cf. edo:, volo:, only 'I am'
is so frequent that its *-mi is not given up, only superimposed:

eso:mi essi esti somos stes sonti
fero: ferssi ferti feromos fertes feronti

Rearrangement of vowel grades: On the model of *somos, *sonti forms
with thematic -o- take stem /s-/, ergo *eso:mi > *so:mi. The others
go by *essi *esti and take /es-/, ergo *stes > *estes.

That has now given
*so:mi > sum
*essi > es
*esti > est
*somos > sumus
*estes > estis
*sonti > sunt.

That was not so very hard. It may therefore well be the truth.

Jens