How does one refute the idea that life is possibly a dream?

Leonard Peikoff wrote in OPAR on page 41:

Once a mind acquires a certain content of sensory material, it can, as in the case of dreams, contemplate its own content rather than external reality. This is not sense perception at all, but a process of turning inward, made possible by the fact that the individual, through perception, first acquired some sensory contents. Nor, as Aristotle observed, is there any difficulty in distinguishing dreams from perception. The concept of "dream" has meaning only because it denotes a contrast to wakeful awareness. If a man were actually unable to recognize the latter state, the word "dream" to him would be meaningless.

Aristotle in On Dreams:

If, then, the exercise of the faculty of sight is actual seeing, that of the auditory faculty, hearing, and, in general that of the faculty of sense-perception, perceiving; and if there are some perceptions common to the senses, such as figure, magnitude, motion, &c., while there are others, as colour, sound, taste, peculiar [each to its own sense]; and further, if all creatures, when the eyes are closed in sleep, are unable to see, and the analogous statement is true of the other senses, so that manifestly we perceive nothing when asleep; we may conclude that it is not by sense-perception we perceive a dream.

Rather than attempt to refute the idea that life is possibly a dream, a simple validation of consciousness as - a state of awareness by means of the senses, as distinct from a validation of the concept of dreaming - a state that the awareness is not by means of the senses, should be sufficient for any rationally honest individual to differentiate that there is a distinguishable difference.

By pinching the person hard and ask him if that was a dream !

I think the best answer would be the one provided by the one making the positive assertion. The onus of proof isn't on you to refute the statement. I'm sure that a person could assert that we are all living in a dream. But, it's their job to prove that we are. Since an argument would have to start with an arbitrary statement, there are really two ways to handle this:

1) Ignore the arbitrary statement (i.e. "suppose life is a dream") from the outset, regardless of the deductive argument used and regardless of how brilliant it might seem. The arbitrary is neither true nor false.

2) Make them prove their assertion. In other words, have them show you how they themselves arrived at such a notion, inductively. They won't be able to do it. When they cannot, you may simply point out the rationalism in their previous argument and dismiss the arbitrary notion as arbitrary. That or simply tell them that they really don't know that what they're asserting is true (though, if they are a skeptic, you might get yourself into a whole 'nuther argument).

Suggest the idea to a child (say, 5-7 years old). The only evidence he has is his senses. He has been using that to learn to distinguish objects, make sounds, walk, talk, use the swing, count and read. Suggest to him that this might all be a dream. Any but the most highly controlled child (one that relies on others to understand the world instead of his own senses), will laugh and tell you, you are so silly! It takes a special kind of mind acrobatics to come up with such an idea and believe in it. To achieve it, one must separate from reality so dramatically, that finding his way back is near impossible.

So in terms of rhetoric, you will probably never win the argument. Your opponent is too far in the clouds to talk about the earth.

But here is a shot: * define reality as that, which we experience, observe, or can deduce based on observations (e.g. the other side of the moon). * anyone who claims something other than that, will need to define what he means by reality. This is hard without using concepts, such as existence, senses, experience (which would then have to be defined in this system, too!) You will find that anyone making the dream reality argument is using stolen concepts that mean something specific, then applying them to some alternate system, in which they would have to mean something different - without defining them!