Er, I didn't say that the behaviour was chosen because it signals something. I offered that as one possible explanation.

In this instance, signalling-to-self seems like about as good an explanation as signalling-to-others for the fact that people give to charities at all (though I'm sure these are nothing like the whole explanation). But it doesn't work for explaining why someone would do it in their will rather than earlier, whereas signalling-to-others does.

For what it's worth, I think my #1 is likely more important than my #2. I would hazard a guess that, e.g., Robin Hanson might disagree. But I'm not sure what your objection is: do you think "signalling" explanations are never an interesting part of the story?