Re: Trivalent logic in Aymara?

Pablo Flores wrote:

>I just finished reading Eco's book and I found
>a very interesting passage towards the end [1]
>that I'd like to comment. According to it, the
>Jesuit Ludovico Bertonio described the Aymara
>language (still spoken in parts of Bolivia and Peru)
>as "a language of extraordinary flexibility, with
>an incredible ability to create neologisms, and
>especially adequate for the use of abstractions".
>
>Eco says that "recent studies have shown that
>Aymara is based on a trivalent logic system instead
>of the bivalent logic (true/false) on which the
>Western thought is based, thus being capable of
>expressing modal subtleties that our Western
>languages can only express by resorting to
>unwieldy periphrases".
>
>  [1] "The search for the perfect language", chapter 17,
>   "Conclusions", section "The translation" (Umberto Eco, 1993)
>
>
>Does anybody know anything about this "trivalent
>logic"? I assume that Aymara, same as Quechua and
>most (or all?) American languages, is heavily
>agglutinating and has a lot of attitudinal affixes
>and such (which would explain Bertonio's description),
>but I'd never heard of the trivalent system in this
>context -- and it sounds really intriguing (and cool
>to try). Have you seen anything like this in natlangs
>or experimented it in your conlangs?
>
Since bivalent logic has the true/false opposition, I take it that a
trivalent logic must be something like true/unconfirmed/false. With
that as my understanding of what trivalent logic means, I take it
that what Aymara really has is a logical category of irrealis -
making no assertion as to the validity of a specific event or state
of affairs.

Consider modal categories (realis vs irrealis) together with the
negative below, and we get a trivalent logic, right?:

Realis:   Strongly asserting that a specific event or state of
          affairs has actually happened or holds true.
Irrealis: Making _no_ assertion whatsoever that an actual event
          or state of affairs actually happened or holds true.
Negative: Asserting that events or state of affairs do _not_
          hold.

Is that what is meant by trivalent logic?

One subcategory of the irrealis that many American languages have is
something called evidentiality - the linguistic coding of
epistemology or certainty of truth. For instance, according to
Payne's "Describing Morphosytax", Huallaga Quechua (a language in
the same region as Aymara) has three enclitics that are clearly
evidential. These enclitics are -mi "direct evidence, -shi
"hearsay", and -chi "inference" (view in monospace font like
courier):

     Qam-pis maqa-ma-shka-nki a.-mi
                              b.-shi
                              c.-chi
     you-also hit-1-PERF-2
     "You also hit me"
  a. I saw/felt you hit me and I was conscious.
  b. I was drunk, and someone informed me that you hit me.
  c. A group of people beat me up, and I think you may have been
     one of them.

In the above example, Quechua appears to have this "trivalent"
logic. Example (a) asserts truth, while (b) and (c) asserts
something that is not entirely confirmed (or irrealis). The event
could then be asserted as false by using the negative.

Incidentally, Boreanesian also has evidential clitics. These are
also hearsay and inferential, and can only be used together with the
irrealis modal clitic. So Boreanesian could perhaps in this sense be
a conlang with a trivalent logic as well.

-kristian- 8)