In chapters 13 and 14 we worked through some simple examples showing how the truth value of a sentence uttered at a particular time and situation can be calculated based on the denotations of the constituent parts of the sentence at that same time and situation. In this chapter we discussed a variety of linguistic features which make this calculation more complex. For many of these opaque (or intensional) contexts, we can only calculate the truth value of a sentence in a given situation if we know what the denotation of a constituent would be in some other situation.9 For example, statements in the past or future tense, like examples (24–25), require knowledge about denotations at some time other than the time of speaking. Statements of possibility (23) and counterfactuals (27–28) require judgments about ways that the world might have been, i.e., other possible situations or “possible worlds”. Some of the non-intersective adjectives, such as former and potential, have similar effects.
As we stated in Chapter 2, it is knowing the sense of an expression that allows speakers to identify the denotation of that expression in various situations. What all the phenomena discussed in this chapter have in common is that the denotation of some complex expression (e.g., the truth value of a sentence) cannot be compositionally determined from the denotations of its parts alone; we have to refer to senses as well.
Kearns (2011: ch. 7) presents a good overview of referential opacity, and Zimmermann & Sternefeld (2013: ch. 8) provide a good introduction to the analysis of intensions as functions on possible worlds. Van Benthem (1988) and Gamut (1991b) provide more detailed discussions of intensional logic and its applications. Partee (1995) discusses non-intersective adjectives (among other issues) in relation to compositionality. For an introduction to lambda abstraction, see Coppock (2016: 93ff.); Kearns (2011: 62–75); Heim & Kratzer (1998: 34ff.).