Evidentials in the GM for proximity



Can/should the evidentials library be used to model levels of proximity of the event?

Here’s a hierarchy from Abui (Kratochvil, 2007):



We used COG-ST:

But the evidentials library wasn’t available yet.


I think in order to decide, we’d need more info on the actual semantics of these (what does “proximate event” actually mean?) but evidentials sounds more plausible than COG-ST to me.

Syntactically, are they adverbs? Auxiliaries?


He is calling them “anaphoric demonstratives”…

They “can occur as adnominal modifier or combine with a VP”.

“The anaphoric demonstrative indicates the discourse location of the referred event when it is followed by another phrase or clause. It indicates the location of the event with respect to the moment of speech (temporal location) when it is the final constituent of a sentence. Clauses with anaphoric demonstratives as their final constituent may express complements.”

They have “nominalising effect and clearly delimit the domain of a clause”.

In (141), hu is said to introduce the information as new to the speaker.

Adverbs, then?..


Those look like clausal mods (i.e. @kphowell’s library)!


Further info on evidentials can be found in the Evidentials Matrix Doc page

Further info on clausal mods can be found in the Clausal Mods Matrix Doc page


Semantically as well as syntactically, you think?


In Ling 567, I came across them first in their use as adnominals, and COG-ST seemed appropriate. The clausal use came up in the corpus sentences – but with a single clause, it doesn’t make sense to analyse them as clausal modifiers:

e-sora e-sapada do lakang masena do
2SG.AL-sword 2SG.AL-machete PRX really be.nice PRX
‘your swords and machetes are really nice!’


The COG-ST analysis is beyond the questionnaire, right? To what degree were you able to use the Q for these, if at all? Also, do you remember if they are obligatory? (That is not immediately obvious from the grammar.) I will ask FK of course, but just in case you know.

What I need to do right now is filter the corpus for sentences which I can hope to model using just the questionnaire or, potentially, to adapt so as to be able to model them using just the questionnaire…


Yeah, within a single clause they don’t look like clausal modifiers, indeed. That doesn’t preclude them being used that way to connect two clauses, though.


Keep in mind that if what you’re really interested in is the syntax & semantics of other phenomena in the sentence, you don’t have to get the semantics of these exactly right so long as their syntax is plausible.


Yes, this was beyond the questionnaire.

I believe they’re optional.