The Meanings of Middle, or mana kartam

From: tgpedersen
Message: 47230
Date: 2007-02-03

The semantics of the middle has always escaped me. He did it for
himself vs. he did it? Why would a language find that distinction so
important that it needed a separate category? And why would that
category at times turn into a passive?

In Old Persian we find the phrase '(ima tya) mana kartam' "(this is
the) by-me done", ie "this is what I did". In late Sanskrit similarly
with the instrumental: sa maya: dr.s.tah "he (was) seen by me" for
"I saw him". "This type of construction becomes the basis of the
preterite of all modern Iranian (and Indic) languages" and "This type
of later Sanskrit is largely Prakrit in disguise. By such devices the
wealth of the Pa:nin.ean verbal morphology can be mostly ignored, and
this simplified Sanskrit was understandably popular".(Burrow).

Now in view of its later success, it is tempting to see this
construction as 'Low Indo-Iranian'. It could probably be used in at
least two shapes, her recast as bad Latin:

'illui factum' "for him (is) done" which corresponds to the
self-benefactive sense traditionally ascribed to the middle.

and

'(ab) illo: factum' "by him (is) done" which corresponds to a passive.

Now in 'High Indo-Iranian' this would have been frowned upon, so it
was hyper-corrected to

'ego factum'

which is actually horrible, but cf the passage from 'me thinks' to 'I
think'.

Now, by this mistake, we have created a new preterite. Also, its
semantics will oscillate between "it is done for me" and "it is done
by me", same as we observe in the middle. Note that in Skt. the
aor.mid.3sg ending is -ta, the Greek and IndoIranian forms together
suggest that a middle could have been formed by analogy: pret. *-mo,
*-so, *-to; *-nto, pres. with *-i *-moi, *-soi, *-toi; *-ntoi.

Unlike in Indo-Iranian, there is no trace of an earlier phase with
pret. "subject" in dative or instrumental in Germanic, but nonetheless
it has the same common ppp./pret. stem construction as modern IIr
does. One wonders. Cf the Ossetic preterite, which also has its
subject in the nominative.


Torsten