Classical views of temporal indexicals like `tomorrow' in English claim that they are rigid indexicals: they refer strictly to the time of utterance. In other languages, however, such indexicals can shift reference when embedded under verbs of speech. In these languages, `John said that he would make cookies tomorrow,' can mean that John committed himself to making cookies the day after he spoke, rather than the day after the entire sentence is spoken. This study seeks to test the conditions in which such shifted readings are possible for 'tomorrow' in American English.
We will test American English speaker's acceptance of shifted tomorrow through a comic captioning task. Participants will be shown a comic consisting of three panels, each labeled with a day of the week. The first panel will have a caption describing the conversation taking place in the first panel. Participants will be shown another sentence and asked to judge how accurate of a caption it is for the third panel. They will rate the caption on a 7-point Likert scale.
There will be four conditions for the main items, defined by which temporal adverbial is used in the critical caption: `tomorrow’, ‘the next day’, Day 1 day-of-week name, Day 2 day-of-week name. The Day 1 items serve as a bad baseline, since captions containing the day-of-week name for the first panel are factually incorrect. The Day 2 items provide a good baseline, since captions containing the day-of-week name for the second panel are both factually correct and do not require any kind of context-shift or anaphor resolution to interpret. `Tomorrow’ is the critical condition, while `the next day’ provides a comparison of `tomorrow’ against a temporal adverbial that requires anaphora resolution, and therefore may have a processing cost, but is predicted to be grammatical.
The critical sentences are designed to control for three known sources of indexical shift: (1) quotation, controlled by making sure 'tomorrow' does not appear in the character's Day 1 speech bubble; (2) Free Indirect Discourse, controlled by embedding 'tomorrow' under an utterance-time oriented attitude report and including utterance-time oriented expressives; and (3) indexical shift as found in languages with shifty indexicals, controlled by embedded 'tomorrow' in a non-finite clause rather than the CP of an event-time oriented attitude report.
We will also gather and analyze demographic features of participants such as age and geographic location.
We will ask a few debriefing questions to understand how participants are interpreting the tasks. We will ask whether they thought of the captions as being read from any particular perspective, and include a few free response questions as a data quality control measure.