Wolfgang Schwarz

The Most Certain Principle

Cresswell calls this the Most Certain Principle:

MCP: if we have two sentences A and B, and A is true and B is false, then A and B do not mean the same.

Last year, I thought that this principle was most certainly false: if I say something true that is false at another world w, and somebody in w says something with the same content, then our utterances mean the same while they differ in truth value. To quote myself,

Consider a possible world that differs from ours only by containing an extra isolated electron in some remote part of the universe, far outside our galaxy. When I say "the number of electrons is even", my utterance differs in truth value from the corresponding utterance of my twin at this world. Does it follow that we mean different things by "number" or "electron" or "even" (or "is")?

I still believe the Most Certain Principle is false, but I no longer think this is a good reason. Cresswell could respond that our two utterances of "the number of electrons is even" really do mean different things: I make a claim about my world, my twin makes a claim about his world. After all, if (per impossibile) a trusting hearer were to hear both utterances, they would learn different things -- as we do when we simultaneously hear two people in different rooms say "it's cold in here". If my twin wanted to say something that has the same meaning as my sentence, he would have to say something like "in alpha, the number of electrons is even", where "alpha" denotes my world.

Contents, on this view, have their truth value eternally and essentially. The content expressed by "it's raining" in the present context is not just that it's raining on May 08 at 23:25 in Canberra, but that it's raining on May 08 at 23:25 in Canberra in world alpha. This is true everywhere, at every time, at every world. Such contents therefore can't be modeled as divisions of possibilities: what is expressed is either true at all worlds or false at all worlds.

This, I thought, is silly: obviously we can express and believe contingent things. But an advocate of MCP could reply that there are two ways of evaluating thoughts or contents at a world. On the first way, we evaluate a tokening of the very same thought or content located at the other world. For instance, we consider an utterance with the exact same content there and figure out its truth value. On this way, we always get the same truth value, because sameness of content entails sameness of truth value. But on the second way, we abstract from the content a certain condition for the actual world and then determine whether that condition is satisfied at other worlds. The condition abstracted from the content that it's raining on May 08 at 23:25 in Canberra in world alpha is the condition satisfied at a world w iff it's raining on May 08 at 23:25 in Canberra in w. If a content is contingent, then that's because it is false at some worlds according to the second way.

The second way of evaluating contents at worlds effects a division among possibilities, a function from possible worlds to truth values. But this function is not the meaning or content itself, nor is the actual content a Russellian proposition of an ordinary kind. Something at another world can make the very same division of possibilities, and express the same Russellian proposition, but have a different truth value. By MCP, therefore, these candidates are not the meaning.

Thus it is a little puzzling that many outspoken advocates of MCP, including Cresswell and Stalnaker, are also outspoken advocates of modeling content by divisions among possibilities. Maybe they don't really endorse MCP. Most philosophers simply ignore the trans-world application of MCP and pretend that if contents don't change their truth-value within a world, then MCP will be satisfied. This indicates that they might have a weaker version in mind:

MCP2: if we have two sentences A and B, and A is true and B is false within the same world, then A and B do not mean the same.

I would recommend going further:

MCP3: if we have two sentences A and B, and A is true and B is false in the same context, then A and B do not mean the same.

I would even strengthen this to a biconditional: A and B have the same meaning iff they have the same truth value in all contexts. People often object to this on the ground that meanings are the ultimate bearers of truth values. I don't quite see how that argument goes: why can't meanings bear a truth value without essentially and eternally bearing it, like St. Petersburg bears the name "St. Petersburg" today, but didn't bear it in 1937? But anyway, if that argument is any good, it works just as well against a notion of meaning based on MCP2.

If MCP was as obviously false as I thought last year, that would be a good reason to suppose that everyone who claims to endorse it actually only endorses something like MCP2. But now that I no longer think it obviously false, I wonder if maybe some really accept MCP. Here is a curious passage from Stalnaker's Inquiry that looks like he does.

Take Mabel, for example, who has arthritis and also has many beliefs about arthritis. One of them is that she has recently developed arthritis in her thigh [...]. Now compare the actual situation with the counterfactual situation where Mabel's inner states and dispositions are just as they actually are [...]. What is different in the counterfactual situation is the way certain words are used, not by Mabel but by other people. The word "arthritis", in this counterfactual situation, is properly used to refer to a wider class of rheumatoid ailments, including the one that is afflicting Mabel's thigh. [...] [T]he belief that she would express there by saying "I have arthritis in my thigh" is a true belief. It follows that Mabel's belief in the actual world is different from her belief in the counterfactual world [...]. (Inquiry, p.67, my emphasis)

By what principle does this conclusion follow? The only candidate I can see is the MCP: Mabel's belief in the counterfactual situation is true, her belief in the actual situation is false, therefore different content. By this reasoning, just about any contingent belief Mabel has in the counterfactual world will have a different content from her corresponding actual belief (and this has nothing to do with linguistic division of labour).