Focus, polarity and framing effects.
PhD thesis, University of Glasgow.
Full text available as:
The experiments reported in this thesis examine a reader's ability to draw inferences about a situation from desires, emotion words and logically equivalent frames. Previous research has provided evidence that listeners (or readers) are able to make inferences about current or presupposed states from the speaker's choice of frame. That is, in experiments involving pronominal reference readers can infer that an expectation has not been reached from the use of a negative natural language quantifier (NLQ) (Moxey, 2006). In experiments on framing effects a listener is able to infer a previous volume of liquid in a glass from the way the glass is now described (Sher & McKenzie, 2006). The experiments reported here aim to take a closer look at how these inferences are made during reading. In the first two experimental chapters character desire and emotion words were used to promote references to the complement set without use of a negative NLQ. Where sentences such as 'The waitress was annoyed by the number of customers who left her a tip' were presented to participants a preference for the complement set was found in sentence continuation and eye-tracking experiments, but not with event related potentials. These results suggest that with such complex inferences to be made participants are able to hold a number of referents in mind during reading. The remainder of the experiments examine reader's interpretation of the logically equivalent frames half full and half empty. In eye-tracking and written experiments it was found that when a character desires a vessel to be full a statement of half full does not fulfil this desire. In contrast when a character desires a vessel to be empty a statement of half empty is satisfactory. It is possible that these results are due to markedness of the terms, where empty is considered marked and full unmarked (Greenberg, 1966). The final two experiments look at two sets of logically equivalent frames in neutral contexts and find that participants find it easier to read about situations which are more familiar or likely to occur. The likelihood of a situation is in turn related to polarity and frequency which is found to be connected to markedness. The thesis as a whole suggests that reader's inferences are affected by their pragmatic knowledge of situations. This is discussed in more detailed with respect to pronominal reference and logically equivalent frames.
Actions (login required)