An Investigation of the Lumps of Thought1
Angelika Kratzer · Oxford University Press eBooks · 2012
Chapter 5 investigates premise splitting and lumping in counterfactual reasoning. Understanding what holds a lump of premises together calls for a semantics based on partial possible worlds or situations. In situation semantics, there are many possible denotations for even the most familiar logical words. We need to find the right empirical constraints for possible denotations, and the current chapter makes a start on this project. Before moving on to counterfactual reasoning, it puts a semantic framework in place that includes constraints on possible denotations and draws attention to subtle shades of meaning that can now be expressed. Central in this chapter is a new proposal for the logical forms of lawlike generalizations and their role in counterfactual reasoning: the logical forms of lawlike generalizations represent confirming propositions that correspond to Nicod’s confirmation sets and play a key role in the construction of premise sets for counterfactuals. The analysis leads to a new solution of Goodman’s Puzzle and helps with other conundrums on counterfactual reasoning: Tichý’s hat, Lewis’ barometer, and Veltman’s three sisters. The most surprising consequence of the analysis is the prediction that it should matter how a lawlike generalization is represented. It should be possible for different, but logically equivalent, lawlike generalizations to behave differently in counterfactual reasoning. The results support a premise semantics approach to counterfactuals over those based on unanalyzed notions of overall similarity between worlds.