double inflection

[Koontz John E]

On Sun, 3 Aug 2003, ROOD DAVID S wrote:
> For most of the doubly inflected Lakota verbs, there is a clear compound
> etymology and inflection of both parts, but there are lots of wrinkles.
> Iblable is from iyaya, the compound of i and ya, with the unexpected
> reduplication of ya 'go', but the paradigm is not in accord with the
> compound analysis.  The 'start' verbs are all compounds of 'arrive +
> move', so this i must be 'arrive going', and historically we would expect
> *wa'ible for this form -- but it doesn't happen that way.

As I said, I'm pretty embarrassed to have forgotten that this
reduplication is unique to this stem in Dakotan!  The parallel form in OP
is idhe, generally rendered 'had gone' in Dorsey's texts.  It does have a
reduplicated form:  i=dhadha.  It appears in the third person as ai=adha=i
(ai=adhadha=i).  Inflected:  A12 aNgai=adha=i.  I don't have any other
personal forms, but the inflection of the vertitive khi=gdhe is
dhakhi=dhagdha=i 'you had gone back'.  There is a vertitive gi=gdhe
(agi=agdha=i) rendered 'arrive'.

I suppose it's possible that Dakotan simply has an inherited iterative
reduplication in lieu of the basic form in this slot.  That would imply
that it formerly had both forms in all relevant slots.  The problem with
that analysis is that then we'd expect *waiblaye (*waibleye?).  Another
alternative that has occurred to me was that something like *ai=aya might
be reanalyzed as (a)iyaya.

> A similar i- initial verb is iyanka 'run' (wa'imnake 'I run'), but the
> inflection is in a different place there.  The only candidate I know
> of for that i- is the same 'to arrive going', and that doesn't make
> semantic sense there.

However, the inflectional pattern is more what I'd expect, based on hiyu
below, and on Dhegiha.

> Older records for hiyu 'start coming' conjugate it wahibu, but no one
> does that any more as far as I know (today it's wahiyu).  That of
> course is a compound of hi and u.  The suus forms of the compound
> verbs have a further quirk that's unique to them as far as I know:
> glicu 'start coming home' adds an extra -ya- syllable between the
> parts when there's an inflectional prefix, so you say both wagliyaku
> and yagliyaku.  Clearly this is NOT double inflection, but I have no
> idea what it is. Similarly khigla (note that this one is NOT
> reduplicated, though it's the compound of the suus forms of i and ya),
> is wakhiyagle.  The same -ya- shows up with the a- prefix that marks
> collective subject for motion verbs: 'they started for home here' is
> agliyaku.  This could be double inflection with epenthetic /y/, of
> course, and may be the source of the analogy for the other forms, but
> I'm not sure I want to advocate that analysis.

This looks like a fossil remnant of the a-prefix on the third person,
maybe wagliwaku/yagliyaku/agli(y)aku resulted in the first person changing
to match the second and third person.

> 	On the other hand, there are lots of compounds that do not inflect
> both parts -- those with the -ya causative and the -shi 'command' come
> immedidately to mind,

In these cases I'd argue that the causative (and the -shi 'command' form
in Dakotan) are among the rare cases where the higher predicate preempts
the inflection of the lower verb.

> as well as all the nonce (syntactic?) constructions with motion verbs
> second and things like eya-lowan '(s)he said, singing', which I think
> would be eya-walowan in the first person, though that should be
> verfied before being cited.

This is something that as far as I can recall has no parallel in Dhegiha.