Re: Analyzing Ayeri's syntactic and voice alignment (long)

Hallo!

On Tue, 18 Mar 2008 21:51:16 +0100, Carsten Becker wrote:

> Hi,
>
> I have started to rewrite the grammar of my conlang from
> scratch some time ago, this time sticking to the table of
> contents that Describing Morphosyntax suggests. So far it
> has been going nice, but now that I have come to the point
> of describing the language's syntactic alignment I have run
> into problems. I want to be sure that what I have thought
> about makes sense.
>
> So far I have been assuming that Ayeri was a "trigger" lang
> with some things misunderstood. Scrap that idea for the
> moment.
>
> Ayeri has three core cases: agent, patient and benefactive
> (dative). It also frequently uses the genitive and the
> locative. Marginally there are the instrumentative (by
> means/help of X, can be both a person or a thing) and even
> more marginally, the causative (due to/because of X, I'm not
> sure whether the term "causative" is right here). I am
> reluctant to call the active and patient nominative and
> accusative or ergative and absolutive respectively, because
> both systems don't seem to fit particularly well:
>
>    (1a) Maliyāng.
>         Mali-yāng
>         sing-3s:m.AGT
>         'He sings.'
>
>    (1b) i.  Haruyās.
>             haru-yās
>             beat-3s:m.PAT
>             'He is beaten.'
>
>    (1b) ii. Haruyāng.
>             haru-yāng
>             beat-3s:m.AGT
>             'He beats.'
So far, just as in my conlang Old Albic:

(1a')     Dairasa.
          sing-PRS-3SG:A
          'He/She sings.'

(1b') i.  Taphá.
          beat-PRS-3SG:P
          'He/She/It is beaten.'

      ii. Taphasa.
          beat-PRS-3SG:A
          'He/She beats.'

> The examples (1a) and (1b) work all just fine. So my
> assumption is that it might be actually split-S in
> intransitive clauses depending on whether S is acting
> themselves or acted upon, rather than depending on
> volitionality because of the following:
>
>    (2a) Rua məhasuvāng.
>         rua mə-hasu-vāng
>         must PST-sneeze-2s.AGT
>         'You had to sneeze.'
>
>    (2b) Tenyayong.
>         tenya-(i)yong
>         die-3s:n.AGT
>         'It dies.'
>
>    (2c) Məvantayāng.
>         mə-vanta-(i)yāng
>         PST-stumble-3s:m.AGT
>         'He stumbled.'
>
> All of (2a) to (2c) are more or less accidental happenings,
> not caused by the experiencer, so the split is probably not
> dependant on volition, as E is still marked as the agent.
This is somewhat different in Old Albic.  The verb mar-
'to die' uses patient marking, as death is something that
happens to a being, not something it performs:

(2b') Mará.
      die-3SG:P
      'He/She/It dies.'

In the other two sentences, dative subjects come into play:

(2a') Ineththa         than.
      AOR-sneeze-2SG:A 2SG-DAT
      'You sneezed.'

(2c') Athambtha         than.
      AOR-stumble-2SG:A 2SG-DAT
      'You stumbled.'

The subject is in the dative case because, while performing
the action, it happens accidentally.  However, with these
verbs, that would be understood that way even if no dative
pronoun was specified.

Ayeri behaves more or less like an accusative language here,
it seems to me.  To this point, the example (1b i.) could
be considered a synthetic passive, compare Latin:

(1b'') i.  Caeditur.
           beat-3SG.PRS.PAS
           'He/She/It is beaten.'

       ii. Caedit.
           beat-3SG.PRS.ACT
           'He/She beats.'

> Let's now move on to transitive sentences; note that the
> distinction between a definite and an indefinite noun is not
> mandatory, so I have been writing 'a' in the example
> sentences:
>
>    (3a) Ang maliya malinoas.
>         ang mali-(i)ya-Ø malino-as
>         AGTFOC sing-3s:m.FOC song-PAT
>         'He sings a song'
>
>    (3b) Le maliyāng malino.
>         le mali-yāng malino-Ø
>         PATFOC sing-3s:m.AGT song-FOC
>         'A song he sings', 'A song is sung by him'
>
>    (3c) Maliyo malinoas (yari).
>         mali-yo malino-as (yari)
>         sing-3s:n song-PAT (3s:m.INS)
>         'A song is sung (by him).'
>
> The sentences (3a) and (3b) illustrate what I thought
> "triggers" would work like (see the other current thread
> about triggerlangs). Sentence (3a) can be analyzed as an
> active sentence with S=A, thus '-(i)ya' would be in the
> nominative and also it would be the subject, since its is in
> focus.
Indeed, "subject" is perhaps the better term here than "focus".

>       On the other hand, (3b) has S=P, which would make
> '-(i)ya' be in the ergative, which would be the subject in
> this case instead. This is why I said that neither the
> definition nom/acc nor erg/abs fits. Sentence (3b) however
> could be analyzed as a passive sentence, since the focus
> moves on from A to P.
Yes, it looks somewhat like a passive.  The notional patient
becomes the subject (zero-marked case), and that is indicated
by the preverbal particle _le_.  That would be a passive if
not ...

>       However, the role marking stays the
> same,
Yes, I realize that, the verb carries the same suffix as
in sentence (1a), and that suffix cross-references the
agent.  This is indeed something else than a "normal"
passive.

>       unlike in sentence (3c), which would be a *true*
> passive in that what has originally been the A now became
> demoted to an oblique object which is not necessarily
> required anymore. Accordingly, the person agreement on the
> verb changes.
Yes; (3c) looks more like a passive than (3b).

>       I may be thinking too strongly along the lines
> of Indo-European languages here I fear. In case you wonder,
> the case marking on the verb is always overridden by an
> according non-clitic NP existing in the sentence. One could
> say that person marking is a cliticized pronoun minus the
> case marker, or in other words, that a pronoun replaces the
> person marking on the verb if no other non-clitic NP exists.
> Furthermore, in order to put arguments in focus, the case
> marker of the respective NP is drawn in front of the verb,
> while the focussed NP itself is left without any case
> marking. In so far I assume that the case marker morpheme
> may be reinterpreted as an anaphora to the focussed NP which
> more or less accidentally shares the case marker's
> superficial form and which is fronted instead of the whole
> original NP. I guess that this kind of marking might be
> unstable in real life, though.
I don't know.  I see no reason why this shouldn't work.

> Now let us have a look at ditransitive sentences:
>
>    (4a) Ang ilya migorayas yeyam.
>         ang il-(i)ya-Ø migoray-as yeyam
>         AGTFOC give-3s:m-FOC 3s:f.BEN
>         'He gives her a flower.'
>
>    (4b) Le ilyāng migoray yeyam.
>         le il-yāng migoray-Ø yeyam
>         PATFOC give-3s:m.AGT flower-FOC 3s:f.BEN
>         'A flower he gives to her.'
>
>    (4c) Yam ilyāng migorayas ye.
>         yam il-yāng migoray-as ye-Ø
>         BENFOC give flower-PAT 3s:f-FOC
>         'To her he gives a flower.'
>
> The sentence (4b) could be transformed like (3c) as well,
> except that 'yeyam' would still be in the sentence, also in
> benefactive, since 'she' does not stop to be the recepient
> of the flower of course.
Yes.  Passivization usually does not affect indirect objects,
cf. English "A flower is given to her."

>       I have been told that the promoting
> of indirect objects (in this case, of 'yeyam' to 'yam ...
> ye') is called 'applicative voice',
As David J. Peterson already said, it is not, but that
doesn't really matter much.

>       but again, (3c) is not
> strictly a passive or an applicative sentence then -- I
> think. Of course the whole focussing can be extended to
> non-core roles as well:
>
>    (5)  Ya məradanyang sirutay.
>         ya mə-radan-yang sirutay-Ø
>         LOCFOC PST-wake_up-1s.AGT night-FOC
>         'At night I woke up.'
>
> In sentence (5), as you can see, the location, which
> provides additional but not crucial information, is in focus
> here. Whether one can call this a locative applicative I do
> not know, since again the role marking stays the same and
> nothing is demoted or dropped. It would not be sensible even
> to drop something here.
>
> If someone could help me make sense out of this mess (well,
> at least it's a mess in my head) that would be very kind ...
The whole shebang makes sense to me.  I don't know how much
it has to do with a real trigger language like Tagalog, but
I don't see why a language like Ayeri should not work in
practice or could not be the ethnic language of a fictional
country.  Whether Tagalog works that way, too, or not, is
secondary.

On Wed, 19 Mar 2008 11:49:27 +0000, R A Brown wrote:

> David J. Peterson wrote:
> [snip]
> >
> > Emphasizing the parenthetical comment, I'd say you're done!
> > As I said (and Ray quoted), just because the conlang trigger
> > system doesn't exist in the real world doesn't mean it's a bad--or
> > even necessarily unrealistic--system.  It simply means that it's
> > unattested.  If you're system makes sense to you, then I don't
> > see what the problem is.
>
> Amen! Amen!
>
> Surely one of the reasons some of us have conlanging is to try out
> ideas. Simply because a system is apparently unattested in the real
> world is IMO no reason at all why it should not be used in a conlang.
> Indeed, i would argue there is even more reason to try it out.
Yes.  This holds not only for engelangs but also for naturalistic
artlangs.  After all, we are all designing languages that did not
exist before, so why not do something that has not been done
before?  I once observed that if the Afro-Asiatic family had died
out without leaving intelligible written records, triconsonantal
root morphology would probably be considered "unnatural" by many
conlangers.

My own conlang Old Albic has a degree-of-volition marking system
which I haven't yet seen in any natlang (see the examples (2a')
and (2c') above), but is it unnatural?  I don't think so.  I also
haven't heard of a natlang that has active/stative alignment,
suffixaufnahme *and* VSO basic word order - does that mean that
there is a "law" against this combination?  I think there is no
such law, the absence of this combination in natlangs being simply
due to all three features being uncommon enough to make their
non-cooccurence a matter of chance entirely, and I happily
implemented this combination in Old Albic.

> I endorse David's comment entirely: *If your system makes sense to you,
> then I don't see what the problem is.*
>
> Experimenting with different systems, and especially original systems,
> surely helps give more insight into linguistic systems generally.
Concurred.

> Whether the system you are trying out originated from some
> misunderstanding of an existing but unfamiliar natlang system, or from a
> flash of inspiration or from whatever source seems to me immaterial. The
> thing is, if the system makes sense to you and seems to work, then I say
> 'Stick with it!'
>
> Originality is interesting - natlang clones can be so boring  ;)
Very much so.

... brought to you by the Weeping Elf