Peggy Speas was a keynote speaker at "The Nature of Evidentiality," a conference held in Leiden on June 14-16. An abstract of her talk, "Evidential Situations," follows.
What is “evidence”? In many formal analyses of evidentials, the denotation of a modal or illocutionary operator simply includes a stipulation that the speaker has “direct evidence,” “indirect evidence,” etc. This is a problem, since “evidence” has no denotation independent of what it is evidence for. I argue that an explanatory theory of evidentials must be a relational one. I will draw on the insights of Nikolaeva (1999), Chung (2007) or Lee(2008), who have analyzed evidentials in terms of temporal or spatial relations, as well as others such as deHaan (1999) who have drawn attention to the fact that evidentials involve overlap or non-overlap with the speaker’s perceptual field. However, I will argue that evidentials denote relations between situations, not times or locations.
Tibetan distinguishes three different types of evidence, ego, direct and indirect. However, within each category there are multiple sub-types. For example, there are three different direct evidential morphemes, (‘dug, shag and song), which are often, but not always, interchangeable.
| (1) | a. |
kha sang khong ‘khrom la slebs ‘dug yesterday he market (LOC) arrived ‘DUG ‘Yesterday he arrived at the market’ (and the speaker witnessed the event’) |
| b. |
kha sang khong ‘khrom la slebs shag yesterday he market (LOC) arrived SHAG ‘Yesterday he arrived at the market’ (and the speaker witnessed the event’) |
|
| c. |
kha sang khong ‘khrom la slebs song yesterday he market (LOC) arrived SONG ‘Yesterday he arrived at the market’ (and the speaker witnessed the event’)
|
In this talk I will focus on the distinction between ’dug and shag . It is fairly well established that song incorporates past tense, while ‘dug and shag are unmarked for tense. (Agfa 1993, Tournadre and Dorje 2003 and Garrett 2001) Although shag is the most frequent evidential in spoken Tibetan, it has not been extensively studied. Documented differences between ’dug and shag do not reveal any obvious pattern. However, a pattern emerges within a theory that treats direct evidentials as denoting a relation of inclusion between the situation being reported and the situation in which the speaker came to know this information. I will argue that ‘dug and shag differ only in the direction of this inclusion relation. This analysis receives surprising support from the behavior of direct evidentials in the antecedent of conditionals, questions and negation. I further argue that apparently idiosyncratic uses of ’dug and shag are extensions of the relevant inclusion relations to epistemic domains.
It would be premature to draw conclusions from Tibetan about the general typology of evidentials, but I conclude by sketching out a typology based on the assumption that evidentials are necessarily relational.
