As clear as rain

[muckefuck]

Novice language freaks generally seem pretty exicited when they discover evidentiality. Imagine, a mandatory inflection indicating the speaker's basis for knowledge! How many disputes would that solve before they begin? Inevitably, some recommend that evidentials should be adopted in their native languages.

It only takes a moment's reflection to see the flaws in this argument. It's not like we don't already have ways of expressing evidentiality in English and experience shows how likely people are to use them to misrepresent (intentionally or not) the basis for their claims. There's simply no way to mandate objectivity in speech, no matter how you go about it. Even a category like "established fact" is so inherently subjective that it alone is fodder for endless metadebates.

And even if you could somehow enforce use of the categories, they wouldn't necessarily mean the same thing. In her chapter on evidentiality in The languages of Native America (Cambridge, 1999), Marianne Mithun points out that, even though both Mohawk and Inuit distinguish "hearsay" information from "direct experience", they aren't necessarily valued the same: The Yup'ik speakers tend to assign more credibility to "hearsay"-marked statements than the Mohawk, since they represent the experience of more than just a single member of the community.

She presents an interesting example for the elabourate system of Central Pomo:

1. čʰé mul=ʔma "It rained" (established fact)
2. čʰé mul=ya "It rained" (eyewitness)
3. čʰé mul=ʔdo "It rained" (hearsay)
4. čʰé mul=nme "It rained" (auditory evidence)
5. čʰé mul=ʔka "It rained" (inference)

This encouraged me to attempt to construct a similar set of contrasts for Osage. (Note that I'm not sure how correct these are; in particular, I wonder if I should've used the continuative apa "moving; absent; out of sight" for all sentences except the first.):

1. nížu akxai "It is/was raining" (simple declarative)
2. nížu akxa che "It is/was raining" (evidential)
3. nížu akxa aape "(They said) it is/was raining" (reportative)
4. nížu akxa ska "I suppose it is/was raining" (suppositive)

"Evidential" for sentence-final che (in the second sentence) is Quintero's gloss. Technically speaking, I guess she considers the other distinctions epistemic modalities rather than "evidentials" in the strict sense. This is reasonable since its usage appears to cover that of categories (4) and (5) from the Central Pomo example and may overlap with (1) as well depending on what the criteria are for considering something an "established fact". That is--as near as I can tell from her examples--seeing the effects of rain (such as puddles) or hearing it fall without seeing it or being in it are both situations which would demand the Osage evidential. That is, if any situations demand it at all. She never says outright that these inflections are compulsory; given the stage of language obsolescence at the time of her fieldwork, it may well have been impossible to tell.

I'm not sure what kind of inflectional epistemic modality is found in other Siouan languages (Dhegiha or otherwise). FWIW, the etymological origin of the Osage suffixes seems pretty transparent. Che is identical to the inanimate "standing" positional article (e.g. mãhkásai che "that coffee standing there [in a pot]"), so the origin might be something like "[the fact] stands [that]" preceded by a nominalisation of the clause. Aape is a third-person plural form of the verb ee "say, tell" (i.e. ee-api-ðe "say-PL-DECL" > eeapie [ð-deletion] > aape [vowel assimilation]). Ska is identical to the adjective meaning "white; clear", so I conjecture an origin like "It is clear [that]" with a nominalised clause. Given all this, the whole system could be of relatively recent vintage.

Comments

[aadroma.livejournal.com]

Such aspects also exist (though not on such as wide of a basis) in Cherokee -- the verb is suffixed with ᎡᎢ, ei, to give verbs like ᎧᏁᎨᎢ, kanegei, "He spoke (though I wasn't actually there, so someone else TOLD me about him speaking).

(I assume you're sticking to Native American languages for this? ^_-)

[muckefuck.livejournal.com]

For this...what? Most of the non-Amerind languages I know with evidentials (e.g. Bulgarian, Turkish, etc.) have a simple two-term system that's less interesting to talk about. Besides, my main mission was to present the Osage system and the discussion of Central Pomo et al. was only to provide some context.

BTW, did you know that there is a Cherokee band in Missouri? I had no idea until I stumbled upon their site yesterday. It doesn't look like they have a significant number of speakers, however.

[joliecanard.livejournal.com]

Macedonian has evidentiality!
The aorist is used for witnessed events and established fact, and the l-past is used for unwitnessed events (and also functions as a perfect tense, though Macedonian has another synthetic perfect tense, има + ppp).