Personal Pronouns

[Jens Elmegaard Rasmussen]

On Fri, 23 Apr 1999, Patrick C. Ryan wrote:

> [...] My strong suspicion is that *mwe, *twe, *swe
> designated topicality, and their use in other cases is transference of
> function. I believe that *me, *te, and *se are original
> absolutive/accusative forms, and there is no need to derive them from any
> other form.

Even so, the non-occurrence of the very FORM *mwe (in any function) does
look as strong evidence for a rule *mw- > *m-. The rule, however, has a
broader foundation, cf., e.g., Gk. mo:^mar 'blemished' vs. amu:'mo:n
'unblemished' where /mo:-/ : /mu:-/ can only be seen as a regular
alternation if derived from *mwoH-/*muH-.

> Secondly, for your idea to be even vaguely plausible, you should be able to
> show Cwe   -> Ce in non-pronominal words, and, as far as I know, this is not
> recognized in IE studies are as normal phonological developement.

We do have *swe'so:r and *swe'k^s without the *-w- in some languages.

[On the "nominative lengthening":]

> I do not believe that I have ever seen this kind of lengthening
> mechanism asserted. [...]

You have now. It goes back a long way in the literature, over a century
actually. I have given it a twist of my own, but only a little.

[...]

> I am also not aware of any IE who has successfully asserted that apophony is
> governed by specific consonants.

This part of it is, whether assertion to this effect proves successful or
not.

[...]

> Frankly, I wish you could identify a <z> for IE; it would simplify my AA-IE
> comparisons since, from all I can see, AA <z> and <s> correspond to IE <s>
> with no indication of apophonic influence.

I'd like to, of course, and I have let you in on the indications I know.
But all indications are to the effect that the two - if ever opposed -
have merged in PIE. It may be like *o and *a in Germanic which used to be
different, but have merged. Nostratic, though definitely on a productive
track, is in its infancy, and the time may not have come for giving
Nostr. evidence priority over the testimony of IE itself. I admit,
however, that I would probably have welcomed external evidence, had it
been postitive.

[... Further on the length of nominatives:]

> I think the more likely explanation of these lengthenings is the presence of
> a resonant; also, I am rather sceptical of IE <*z> exercising its
> lengthening across an intervening consonant (if only a resonant).

The nominative lengthening also works on stops: *ne'po:t-s, *wo:'{kw}-s.
It never works on an immediately preceding vowel: *-os. I do not see how
its lengthening effect can be used to decide against its being earlier
voiced. Strictly, we can only say - and that's all I claim - that the
nominative marker appears to have been voiced in word-final position when
preceded by a vowel. That could be an effect exerted upon it by the
environment, but then it is an effect to which the *-s marking the 2sg was
immune, so even then we would need two different s's. What is so bad about
that - Hindi /s/ has _three_ known sources.

> > Lengthening is also seen in the s-aorist which then apparently had the
> > lengthening s.  [...]

> I agree that this is an interesting congruity.

That's a pleasant reaction.

[...]

> I do not doubt that different results indicate different conditioning
> factors but I see no major objection to supposing that the conditioning
> factor is probably stress-accent if not simply a desire to differentiate
> phonologically identical forms fulfilling gradually differentiated
> functions.

But what is the probability that morphological differentiation not based
in phonetic change gets to LOOK so much like the result of phonetic change
that its variation can be stated in terms of consistent phonetic rules?

[...]

> You must know that I am merely accepting Beekes on this point. I also once
> toyed with two different s'es but I am convinced that the explanation
> offered by Beekes is the stronger and more economical explanation.

It would be if it explained the facts as they are. The trouble with
Leiden-style IE morphology is that so much of the material has to be
explained as secondary. I want to respect the material - or at least the
parts of it that I cannot explain away.

[...]

> You certainly seem to like "variants".

Yes, that's all there is to work on in internal reconstruction. Without
morphophonemic alternation, there is no clue to the prehistory of the
forms.

> A look at AA may convince you (e.g.
> Egyptian -wj, dual) that, most likely, the dual -w has a separate origin.
> After all, although a <m/w>  variation is attested is some IE languages like
> Hittite, it is not, so far as I know, established for IE.

I would of course change my mind if I got to know external evidence well
enough to be able to control it and to see its relevance, and then found
it to be in conflict with my present views. Such events would consitute
quite big surprises, for the IE amount of co-variation is not exactly
negligeable, but some of it could of course still be illusory. - That an
m/w alternation is not established for IE is precisely the reason why I
write about it, for it seems to be there. - BTW, I find it difficult to
see that -w- expresses the dual in Egyptian, if the plural ends in -w and
the dual in -wy - and the dual personal suffixes simply add -y to the
plural forms where (outside of the 3rd person) there is no -w-.

[..]

> Not trying to be obtuse but I do not really understand your point here.

I was saying that, in the personal pronouns, the oblique cases are all
built on the acc.: Skt. dat. asma-bhyam, abl. asma-t, loc. asme /asma-y/
parallel to ma-hyam, ma-t tva-t, tve /tva-y/, even instr. tva: and
yuSma:-datta- 'given by you', the underlying acc. being found in Skt. in
the extended form ma:m /ma + -am/, tva:m /tva + -am/ and the normalized
acc.pls. asma:n, yuSma:n - unextended forms seen in Avestan ma, thwa,
ahma. That IS a system like that of Modern Indic where case-forming
postposition are added to the old accusative.

[...]

> All I was saying is that most IEists do not believe (and I agree) that the
> dual is as old as the singular and plural in IE. A dual is certainly not
> necessary in any language to express "two of" something.

That is no valid reason for considering the IE dual forms as we find them
younger than the sg. and pl. forms we find. The real weight of the actual
evidence rather tips the balance the other way: If a dual category is
superfluous, and the IE dual forms are not characterized by productive
elements, the dual looks like a cumbersome luxury present only because the
older generations had it, and not for any important purpose of its own.
That spells archaism if anything does.

[...]

> Obviously, after the dual *was* developed, its rather specialized (and less
> frequent) use would tend to preserve its forms in an archaic state.

If its forms could be in an archaic state, it must have been old enough to
allow me to take it seriously in my analysis of the IE personal pronouns.
That was all it was mentioned for.

[...On 'me' and 'my':]

> This is, IMHO, a completely erroneous analysis. Germanic *mi:na-z is
> composed of IE *me + -y, adjectival/genitive + -nV, nominalizer + -s (case
> marker). *e-mo-s is an emphatic expansion of *me by *e-, in keeping with a
> general tendency to expand monosyllables to disyllables + -s, case marking.
> It has only the last two elements in common with *meynos.

If that is so, you cannot use the same system to account for the
occurrence of /w/ in *tewe 'of thee' (Skt. tava, Lith. poss.adj. tava-)
and /n/ in *mene 'of me' (Av. mana, OCS mene, Lith. poss.adj. mana-).

> > THEN you can link Hitt. to the rest of IE in a smooth and unforced way
> > if you assume that the /u/ has propagated from 'thou' to 'I' and from 'I'
> > to 'me', and finally from orthotone 'me' to enclit. 'me'.

> I do not assume that the basal form is *tu(:), and so cannot justify
> migrating <u>'s.

Then why not change your assumption about 'thou' and get the benefits?

[...]

> The basis for suspecting "topicality" for -w is not present in IE nouns. It
> is necessary to reach beyond IE to "justify" it.

Fine with me, that makes it even older and more deeply rooted when we
find it in the pronouns, where it is then definitely a component to be
accounted for in the underlying forms.

> > BTW, what is the basis for taking the *-g^ to express topicality?

> I do not take it so. I believe it expresses 'maleness', based on analysis of
> IE roots containing <g{^}> and extra-IE morphemes.

Sorry, I misread you. But why "maleness"? Did women not say *eg^?

[...]

Thank you for your trouble with rethinking this intricate question. We
cannot expect full agreement to be right around the corner.

Jens