Austin proposes a linguistic solution to the sceptical argument from dreaming. We use the phrase “dream-like quality” in everyday life, when describing certain types of events, but what does this mean? Presumably it refers to a quality that separates dreams from reality. But if there is such a quality then the argument from dreaming, which rests on the idea that we cannot tell waking and dreaming experiences apart, fails.

According to Austin then, the premise “dream-experiences are indistinguishable from ‘real’ experiences” is false, as is demonstrated by our linguistic practices. He writes “If dreams were not ‘qualitatively’ different from waking experiences, then every waking experience would be like a dream”. Indeed if the premise were true then “the phrase [“dream-like quality”] would be perfectly meaningless”, he claims. [Austin, Sense and Sensibilia, p49]

Austin’s attack need not worry the sceptic. Firstly, we must note that linguistic practices are not the infallible guide to truth that Austin may hope. Ridiculous examples abound here, for instance, we may say that someone has “undergone a change of heart” without implying that major surgery has taken place. Linguistic conventions are related more to folk-knowledge than scientific knowledge. As a source of complex theories they are not to be trusted.

But even if we accept Austin’s point we have yet to dismiss the argument from dreaming. Consider the phrase “summer weather”. Let’s say it means weather which is hot and dry. To describe a hot, dry day in March as “summery” is not to say that all summer weather is like this, it is simply to say that a high temperature and an absence of rain are typical of summer days. Similarly, if we accept that there is a valid sense in which we use the phrase “dream-like quality”, this means nothing more than that dreams are often a particular way. We cannot attribute this quality to all dreams any more than we can deny that it ever rains in the summer. This leads to a revision of the premise of the argument from dreaming to “Some dream-experiences are indistinguishable from ‘real’ experiences”. For any real experience which lacks Austin’s “dream-like quality” we can still wonder if it is one of the minority of dreams which lacks that quality rather than a waking experience. The weakened premise therefore still supports the sceptical conclusion.