Again, the term “imagination” is sometimes confined to intellectual processes in the fine arts: dramas, novels, etc., are works of imagination. Now dramas and novels all proceed upon one method, namely: they begin by stating or insinuating an hypothesis concerning certain persons in a given situation, and then deducing (that is reasoning out) the consequences, occasionally helping the plot by further assumptions: at least that is how it appears, though probably the main incident of the plot is thought of first, and then an hypothesis is framed that conveniently leads up to it. And if the reasoning is feeble, and if the subsidiary assumptions are too numerous or too facile, we say the work is flimsy or improbable—allowing for the _genre_; for a romance is not expected to be as probable as a modern novel. _Gulliver’s Travels_ afford the most perfect example of this method; for each voyage begins with a frank absurdity—men six inches or sixty feet high, a flying island, rational horses; but this being granted, the sequel makes tolerable logic. Well, many scientific investigations seem to follow exactly the same method—begin with an hypothesis, deduce the consequences, and occasionally help out the argument with further hypotheses (though that is not all): and here again the conclusion is usually thought of first, and the hypothesis invented to explain it. If it be said that the scientist believes his hypothesis to be true, whilst the romancer does not, it may be replied that the scientist sometimes expressly warns us that his assumption is only a “working hypothesis,” which may not be true (though he thinks it may be), whereas early epic poets and minstrels often regarded their work as by no means without a foundation in fact.