In “Kripke, the Necessary Aposteriori, and the Two-Dimensionalist Heresy,” Scott Soames argues that the two-dimensionalist (a la Chalmers and Jackson) response to Kripke’s arguments for necessary a posteriori knowledge in Naming and Necessity do not work. I’m not going to go into the details of the article in this post, just point out something that kind of puzzles me. In the article Soames says various things about how speakers might understand a sentence or a name without knowing or believing certain things about the sentence or the name. For example, he says that “Peter Hempel lived on Lake Lane” and “Carl Hempel lived on Lake Lane” mean the same thing “even though speakers who understand them may not realize they do”. I’m not sure if I’ve seen “understand” in the philosophical literature on semantics much. (Maybe I have and it has slipped my memory…) I’m not sure what exactly Soames means by “understanding”. In the articles on propositional attitude reports I remember, everything is framed in terms of belief and propositions. In Lexical Competence by Diego Marconi, there is a delightful discussion of understanding and its degrees, but I am assuming that that is not what Soames has in mind. When one understands a sentence with a name in it, on the direct reference picture, what does one understand about the name? There is nothing to a name apart from its referent. But, if I don’t realize that Carl and Peter Hempel are the same person, then it seems like I do not understand the name. I suppose I understand that they are both names, but that isn’t understanding the sentences and the names; that’s just understanding how constituents function in syntax. I imagine if one has access to Perry’s theory of mental folders and roles that understanding has a natural home. One just starts playing information games with the names. But, that route isn’t available. Direct reference about names is by itself rather austere, so understanding doesn’t seem to happily fit into the picture there. It is important in setting up Soames’s arguments that the agents understand the sentences and names, but it is difficult to see what that comes to.