Re: Evidentiality in gjâ-zym-byn

On 8/15/05, Jeffrey Jones <jsjonesmiami@...> wrote:
> On Sun, 14 Aug 2005 22:06:32 -0400, Jim Henry <jimhenry1973@...>
> wrote:
> > It's less concise than the languages that use a single-phoneme affix or
> > fusional inflection or monosyllabic particles for evidentiality
> > (always at least two syllables, sometimes three or more), but still
> > more concise than the roundabout means for indicating evidentiality
> > previously available in the language.
> That's a nice method, since it makes evidentials an open class. I'm trying
> out something vaguely like that, but with certain limitations, in a new
> project. If a word can be used as an evidential, a form with the auxiliary
> affix is used. This takes person marking, unlike yours, where I guess 1st
> person is implied by -EVD.
So, how does the auxiliary affix work?  Does the
evidentiality marker take the form of an auxiliary verb,
with that getting the person marking, while the previous
main verb becomes infinitive...?

> Possibly my system is like your previous method?
I should have been clearer about what the
"roundabout means for indicating evidentiality
previously available in the language" were.
Basically, it would have required recasting the sentence
with the evidentiality as the main clause
(I've seen/heard/read that....) and the previous
form of the sentence becoming a subordinate clause;
or else using an evidential "because" or "therefore"
conjunction to connect the main assertion with the
statement of how one learned this.


On 15 Aug 2005 04:48:07 +0200, Henrik Theiling <theiling@...> wrote:

> I'm a bit destructive now instead of con~: I can only give the reasons
> for not having derived words of evidentiality in Qthyn|gai:
>
>   - technical: evidence/mood is one of the few mandatory categories
>     in Qthyn|gai (there are three: evidence/mood, valence, and case).
>     It felt inappropriate to not have special marker since there
>     are so few mandatory categories.
Yes, if I were making a language with mandatory
evidentiality marking I would probably have the evidentiality
morphemes be a single phoneme, or maybe have fusional
morphs that mark both evidentiality and something else
(aspect or mood, maybe).

>   - semantical: IMHO, it is not important for the evidentiality,
>     i.e., the clarification of *how* you come to a given conclusion,
>     of whether it is heard, seen, smelt, or physically felt.  The
>     point is, it is perception.  I think any deeper clarification
>     would be better done paraphrastically instead of grammatically.
Maybe so.  But given the affix I'm trying out for now
with gzb, it seems to make sense to usually use {riqm-poxm},
"I saw for myself", rather than a vaguer {zqaxw-poxm},
"I've perceived it for myself".

>     A generic approach as yours would feel too paraphrasing to me.
>     From the mere evidence point of view, the essence is that the
>     truth value is derived from perception.
>
>   - pragmatic: I did not want people to be able to chose their own
>     evidentiality markers. :-)))) There is fixed set to chose from
>     and no generic means to form new ones is allowed.  Especially,
>     there is no 'from reliable sources' evidentiality, in order to
>     prevent politicians from using it. :-)
Whereas I would kind of like to be able to solve
the evidentiality problem once by adding a single
additional suffix, which I can apply however it seems
suitable when needed, instead of reanalyzing it
from time to time as I would probably feel a need
to do if I came up with a set of several adverbial
particles for various kinds of evidentiality --
direct perception, hearsay considered reliable,
hearsay considered unreliable, etc.  Also,
gzb is fairly minimalist lexically, and it would feel
weird to have another particle meaning "evidenced
by direct perception" when I already have a word
for "perceive" and more specific ones for "see", "hear",
etc.  I feel like I ought to at least try deriving evidentiality
markers from root words for a while before
deciding they are really too verbose and I need some
monosyllabic particles or affixes for specific kinds
of evidentiality.

> Anyway, all of this is gut feeling.  For instance, the most generic
> case markers, namely verbs, are the basis of my new S11 (Tesäfköm)
> where Qthyn|gai had a fixed set of grammatical cases.  Views and
> solutions change.  Maybe by good examples you can convince me that
> generic evidence markers are better than grammaticalised ones.
Absolutely better, maybe not; but I think the
method I'm experimenting with now is better to retro-fit
into an already well-developed language like gzb.
If I were starting from scratch I would probably do it differently.

pirahanq-lam  i-m        henx gun-nxix'biq miq-i
Pirahã-NAME.L in-part.of not  word-number  TOP-at

sqe   siqnx-flu-kox-poxm.
maybe information-flow-PLACE-EVD

Pirahã supposedly has no words for numbers (I read it on the Internet).


--
Jim Henry
http://www.pobox.com/~jimhenry/conlang.htm
...Mind the gmail Reply-to: field

Reply