UBC Faculty Research and Publications

Evidential-temporal interactions do not (always) come for free Hirayama, Yuto; Matthewson, Lisa, 1968-


Evidentials are usually assumed to encode the speaker's evidence source. However, some authors propose that evidence source can be derived from temporal information elsewhere in the sentence (e.g., Chung, 2007, Lee, 2013, Koev, 2017, Bowler, 2018, Speas, 2021). We argue that evidence source cannot always be derived from temporal information. Using data from five languages, we propose that evidentials can lexically restrict the time the evidence was acquired. Evidentials can do this independently of other temporal marking, and they sometimes encode both temporal and evidential information. We show that English ‘apparently’ and ‘seem’, Japanese indirect ‘yooda’ and reportative ‘sooda’, and St'at'imcets (Lillooet; Salish) perceived-evidence ‘an’ all require that the earliest time their prejacent becomes true, EARLIEST(p) (Beaver and Condoravdi, 2003) precedes or coincides with the Evidence Acquisition Time. Conversely, English epistemic 'should' and German epistemic ‘sollte’ require EARLIEST(p) to follow EAT. A third group encodes no temporal restrictions: English epistemic ‘must’, St'at'imcets inferential ‘k'a’ and reportative ‘ku7’, and Gitksan (Tsimshianic) inferential ‘ima’ and reportative ‘gat’. Comparing temporal with nontemporal evidentials supports the view that a temporal component is hardwired for the former set. The fact that temporal contributions cross-cut evidential ones shows that one cannot be reduced to the other.

Item Citations and Data


Attribution-NonCommercial-NoDerivatives 4.0 International