Optimal design in language & Reference set computation and focus
Reinhart has argued that reference set computation must be operative at the level of the interface between syntax, semantics and context (Reinhart to appear). Reference set computation is found when a particular linguistic form Form1 is potentially ambiguous; the rules of the interpretative system can assign to it both Interpretation1 and Interpretation2. However, there exists an alternative linguistic form in the language, Form2, which also has Interpretation2, and is a more optimal way to express it. In many such cases, the association of Form1 with Interpretation2 is blocked, because of the existence of the alternative Form2.
Reference set computation has been argued to occur in the following domains: Quantifier raising; stress shift for focus construal; pronominal coreference, and scalar implicatures. In the last two areas, acquisition studies already established that children have difficulties, which, it has been argued, can be interpreted as reflecting difficulties in processing the required reference set computation The present project concerns one of these: focus.
Reinhart assumes that main stress, which is assigned independently during phonetic realization, also determines directly the set of possible foci of a given derivation. The focus set associated with each derivation includes all the constituents that contain the main stress. E.g. (1a) is associated with the focus set (1b). (bold=main stress)
| (1) | a. | [IP My neighbor [VP is building a desk ]]. |
| b. | Focus set: {IP, VP, Object} | |
| (2) | My neighbor is building a desk. |
However, the independent mechanism of main stress assignment is not always sufficient for the needs of the context interface. E.g. the subject is not in the focus set of (1a) For this interpretation, a special mechanism, stress shift, needs to apply, displacing main stress, and deriving (2). Such adjustment operations are permitted only when there is no other way to obtain the intended interpretation. Deciding this requires the construction of a reference set, which, for (2) includes (1). So, (2) cannot have IP-focus, because this interpretation is available also in (1a), without stress-shift.
The present project will involve adult processing experiments in order to establish whether there is evidence indicating the presence of the processing load in cases involving reference set computation in the area of focus with stress shift.