Quiddities is a more light hearted book than is typical of Philosophy, and is regarded by Quine as being only partly philosophical.
The work is subtitiled:
For this reason these notes will have the same structure, consisting of notes on some of the topics on which Quine writes, ordered alphabetically by Quine's topics. It is likely that these notes will also be cross coupled with entries in my own glossary.
I'm afraid I have to disagree with Quine's casual conflation of logical with physical necessity. In fact (just in case you havn't noticed) the distinction is so fundamental and important in my opinion that it forms a major theme of my Web pages.
We are talking here about the distinction between necessary and contingent truths. This distinction does not lie in our infallible knowledge of the one and our fallible knowledge of the other. To hold that a statement is necessary is not to deny the possibility of error.
Speaking more specifically of the holistic argument, the distinction between a necessary and a contingent claim is not whether we may, in the face of empirical evidence, change our mind about its truth, but about how we justify or explain our new view.
Whatever our grounds for belief in a proposition, whether they be logical or empirical, if we are persuaded that we have mistaken the meaning of the proposition then we are liable to change our mind about its truth without benefit of any further evidence or demonstration. Let us consider then only the cases in which our view of meaning is stable.
If we change our mind about the truth of a logical or mathematical statement because of some empirical evidence, we will do so only because the empirical evidence has helped us in some way to see a flaw in the original proof, or to doubt the existence of a proof we had previously taken on trust. If we thus revise our view of the truth of a mathematical proposition, then it will be possible subsequently to make the case without benefit of the empirical evidence. Even if empirical evidence plays a vital role in changing the mind of a mathematician about a mathematical proposition, he will not admit the evidence into a demonstration or refutation of the proposition.
By contrast, with a contingent proposition there will be no demonstration or refutation which can render the new insight independent of the empirical observation. 1
Elsewhere Quine disputes the notion of analyticity. In this argument he relies uipon indeterminacy of translation. I don't buy this one either.