(1) If you know that you see something, you this solely on the basis of your evidence about your environment.
(2) This evidence does not favor the hypothesis that you see over the hypothesis that you do not have a mind at all, and so does not allow you to know that you see something.
(C) You do not know that you see something.
This basic template can be extended to believing something, liking something, feeling something, and so forth.
Berkeley correctly diagnosed the primary flaw with most arguments for external world skepticism as the positing of two external worlds -- what everyone in fact calls the external world and an external world arbitrarily designated as the real one, if it exists. Obviously, if there is any real danger of skepticism about the external world, it has to be skepticism about what people take to be the external world, and not some other 'external world' concocted to make skeptical arguments work. If we start with the former, it then becomes much more difficult to run any kind of argument for external world skepticism at all -- a fact that Berkeley uses to his advantage.
A similar sort of problem can easily be seen to arise for Byrne's argument for internal world skepticism. This becomes quite clear if we look at how he argues that the template above can be extended to liking:
Now take the liking hypotheses. Why do you think you like chocolate? Isn't the answer something about the chocolate? You like chocolate because it tastes good. This is a fact about the chocolate, not about you. When you savor a piece of chocolate on your tongue, your sensory systems are detecting features of the chocolate, in particular its agreeable sweet taste. On the basis of this evidence about the chocolate you conclude that you like it.
It's quite clear here that the argument is working by positing some internal world other than what anyone actually means by 'internal world' and then, unsurprisingly, finding no evidence that this made-up substitute for an internal world exists. Nobody would in fact take 'tasting good' or 'having an agreeable sweet taste' to be a fact about chocolate and not about the liking. It is not even really correct to say that 'tasting good' is 'evidence' for liking chocolate; this is precisely a positing of a distinction between what everyone treats as the internal world and some made-up thing. In the tasting of the chocolate, liking chocolate you taste and the good taste of chocolate are the same thing. Sweet taste is not something that people take to attach to chocolate simply in itself; they take it to be what chocolate has when being tasted. The good taste of chocolate is a fact about chocolate and about you; which is, indeed, how everyone in fact takes it. This point extends to all sensation, and all mental activities closely associated with sensation. As Lady Mary Shepherd might put it, sensing involves mixing ourselves with the world, and the sensation is the mix of the world and ourselves both. And, indeed, there is a problem with talking about this matter in terms of 'evidence', as well, since calling something in the external world 'evidence' locates it in the internal world, as well.
There are other things that could be said. (For instance, Byrne gives an objection to the external world counterpart of (1) that he calls the 'animals objection', and says that (1) is immune to it. This is one of the reasons for his claim that the argument is not as weak as the external world version. But if we formulate the argument in terms of liking, I think it is much less clear that (1) is immune to the animals objection.) But it's a salutary exercise, since it shows the importance of recognizing that minds are not hermetically sealed from the world around them.
*****
Quotation from Alex Byrne, "Skepticism about the Internal World", The Norton Introduction to Philosophy, Rosen, Byrne, Cohen, and Shiffrin, eds., Norton (New York: 2015) pp. 288-289.
