NATLANG: Re: Hungarian tense, aspect, mood...

On 27 Apr 2004 Frank George Valoczy <valoczy@...> wrote:

> It is the -l- infix which makes the distinction of meaning here -
> indicating 2nd person object.
  The things that prevents me thinking of the -l- as an infix are
the following:

  It's a morphological "hapax": you may insert only in 1st person
singular forms (Sg1). But it's impossible to infix into other
forms, e.g. into 1st person (Pl1) plural _szeretu:nk_ 'we love' + -
l- > _*szeretlu:nk_. Morphological cathegories, like infixes,
should have a bit more systematic characteristics for me.

  The -ek on the end of _szeretlek_ is the marker of the
*indefinite* Sg1. Therefore, -l- influences the next suffix and
transforms it from indifinite to definite one. It seems a bit odd
behaviour when speaking about infixes.

  You wrote: "'szeretek' exists on its own; *szeretl obviously does
not". The string *szeretl (in form *szeretel) can be analysed as a
valid word in a previous state of the language. Note that -l is the
original Sg2 marker in "ik"-type paradigm, and the "ik"-type
paradigm originally was used to transform the Agent into an
Experiencer or into a Patient. Thus _*szeretel_ is a possible word
for 'thou are loved, X loves thee'.
  Rarely, even verbs with personal markers can be a base in
Hunarian to derive other verbs, cf. _la'ttamoz_ 'to countersign,
lit. to [write] "la'ttam" (= 'I have seen, I agree') [on it]'. So,
the variable "X" could be resolved by attaching another (namely
Sg1) personal marker to the verbal form _*szeretel_ 'X loves thee'.
And this marker is systematically comes from the indefinite series,
because the "ik"-type verbs are intransitive by definition.
  (Note: The omission of the vowel from _*szeretelek_ is a regular
process. There's a rule in Hungarian, called "double open syllable
rule" that tends to elide either vowel of two consecutive open
syllables.)

  A very similar situation exist is in Mordvin (this language has a
nearly full bipersonal marker system). E.g. vechk-at 'thou love' +
vechk-an 'I love' :> *vechk-at-an > vechk-t'an 'I love thee'. Most
of the Mordvin bipersonal markers adopt the scheme: personal suffix
of the direct object (as it were subject) + personal suffix of the
subject.

  [Mordvin data came from the Erza dialect: ch = /tS/, ' =
palatalization of previous consonant]