Critique of Premise 1
It is hard to deny that my idea of green is different from my idea of red.

Critique of Premise 2
This premise, while not as obvious as premise 1 seems obviously true -- at least to those of us who have seen such rainbow-like displays. One might think, however, that green and red are so different that nothing can be very much like red and very much like green. And that much is true, but that just shows that there must be more than one shade between red and green.

But the premise does seem to assume that there is a finite number of distinct shades of a color and one might think shades of a color literally form a continuum, not just that there are many (but a finite number) of them.

Critique of Premise 3
We can imagine someone being presented with almost every shade of blue. Imagine a big chart in a paint store that has had a swatch cut out of it (for another) and has since been spliced back together with that swatch missing. I am not sure whether Blake would perceive a "gap" without it being pointed out, but the claim that he would is not as important for the argument as the claim made in been premise 4.

Critique of Premise 4
I think most of us are inclined to agree with Hume that Blake could in the circumstances imagine the missing shade of blue. But how could this really be tested? If Blake later would see the shade (on the walls of the previous customer) would he say "Ah, there it is -- I've been wanting to meet it for years?" Suppose we ask Blake to mix the paint to that shade. But of course if he gets it right he will then have the needed impression, so perhaps this does not prove anything. And in any case, why couldn't we suppose that Blake could do this by saying something to the effect of "I want something darker than blue-226 (which he seems on the chart) but lighter than blue-228" and mix the paints until he gets this result--without actually having an idea of blue-227 before the paint is finally mixed right? In short, it is rather difficult to prove (even general skepticism aside) that Blake (or anyone else) has an idea of that particular shade of blue.

Critique of Premise 5
This key premise seems faulty. Hume's distinction between simple and complex ideas is not as clear as it might at first seem. One might think (as Hume did) that an idea of a color is always simple. After all, typically these ideas are copied from impressions. But we also think (with Hume) that we could generate some of these ideas by, say, imagining a darker shade of blue and lightening it in our mind, or by imagining green and imagining blue and combining these in our mind (splitting the difference) and arriving at an idea of green-blue, even if we had never seen such a color.

Critique of Inference
The argument for subconclusion 1 is troubling. It seems to be a fallacious slippery slope. My ideas of colors are not as precise as this argument requires. Just because I distinguish red from green, it does not follow that I can distinguish every shade from every other shade. So it is far from clear that I have a different idea for each shade of a color I have come upon. Indeed, I rather doubt that I do. Hume himself offers a response to this counterargument against his thesis. He simply says it is a "singular" objection that should not trouble us: ". . . this instance is so singular, that it is scarcely worth our observing, and does not merit, that for it alone we should alter our general maxim." [text]

Defense of the first con argument against the critique
In critique of premise 5 I suggested two ways that one might arrive at an idea of a particular shade of a color without having had an impression of it. These two ways are not equally allowed by Hume's system. The suggestion that one might "lighten" a darker shade of blue might be allowed by Hume (if he didn't presume each idea of a shade of blue must be a simple idea), since he allows that one of the (few) things our mind can do to the material of the senses is to diminish it. One might think of a lighter blue as a diminished version of a darker impression. But one cannot make the same move with the suggestion that one interpolate between blue and green and arrive at an idea green-blue that way, since Hume doesn't seem to allow that our mind has such powers of interpolation between two different qualities [footnote] (although between two quantities it would simply be a matter of diminution or augmentation, which are allowed) and this is not merely a matter of "compounding" the ideas of blue and green, is it?

Even though this overall argument about the missing shade of blue may not convince us, it still is an effective argument against Hume. After all, his scheme seems to require a distinction between simple and complex ideas and to require simple ideas to be copies of impressions. Surely most of our ideas of colors are ideas that have been directly copied from impressions, if any of our ideas are. But then the missing shade of blue case does seem to provide a counterexample.

Evaluation of first con argument in light of critique
Hume's nonchalant attitude toward this counterexample cannot easily be condoned. If Hume really holds that all our ideas are derived from impressions and more particularly that all our simple ideas are copied from corresponding impressions, then he cannot simply back off and say "well almost all." If he really is content with "almost all," then the use to which he puts this thesis is gutted. (See the "Significance" section.)

On the other hand the example does seem contrived and unimportant. Why is this? It is, I suggest, because it seems clear that Blake's idea of blue-227 is derived from his impressions. It is not, admittedly, derived from an impression of blue-227, but surely unless Blake had impressions of other shades of blue (or at least of other colors), he would not have had this idea of blue-227. So this example seems to support, not to undermine, Hume's thesis! What it does undermine is Hume's assumption that all our ideas of shades of colors are simple ideas. Surely some of these ideas are derived from impressions, but others might be somehow built from other impressions of colors. Indeed, even so basic an idea as our idea of red is probably not copied from some single impression, but is somehow a conglomerate, or abstraction perhaps, of several impressions of different shades of red. If so, then our ideas of colors are not simple after all. And if that premise is dropped, the argument here (against Hume's thesis) simply does not go through.

And indeed if the distinction between simple ideas and complex ideas is based on the genesis of those ideas, then there is no need to assume (and in fact it seems false) that one can distinguish between complex and simple ideas by inspection, so to speak, as one can distinguish between visual ideas and auditory ideas.