For these reasons I think that any errors involved in such an anticipation may, without much forcing, be brought under our definition of illusion.  When due altogether to the immediate force of suggestion in a present object or event, and not involving any conscious transition from past to future, or from general truth to particular instance, these errors appear to me to have more of the character of illusions than of that of fallacies.

Much the same thing may be said about the vivid anticipations of a familiar kind of experience called up by a clear and consecutive verbal suggestion.  When a man, even with an apparent air of playfulness, tells me that something is going to happen, and gives a consistent consecutive account of this, I have an anticipation which is not consciously grounded on any past experience of the value of human testimony in general, or of this person’s testimony in particular, but which is instantaneous and quasi-immediate.  Consequently, any error connected with the mental act approximates to an illusion.

So far I have supposed that the anticipated event is a recurring one, that is to say, a kind of experience which has already become familiar to us.  This, however, holds good only of a very few of our experiences.  Our life changes as it progresses, both outwardly and inwardly.  Many of our anticipations, when first formed, involve much more than a reproduction of a past experience, namely, a complex act of constructive imagination.  Our representations of these untried experiences, as, for example, those connected with a new set of circumstances, a new social condition, a new mode of occupation, and so on, are clearly at the first far from simple processes of inference from the past.  They are put together by the aid of many fragmentary images, restored by distinct threads of association, yet by a process so rapid as to appear like an intuition.  Indeed, the anticipation of such new experiences more often resembles an instantaneous imaginative intuition than a process of conscious transition from old experiences.  In the case of these expectations, then, there would clearly seem to be room for illusion from the first.

But even supposing that the errors connected with the first formation of an expectation cannot strictly be called illusory, we may see that such simple expectation will, in certain cases, tend to grow into something quite indistinguishable from illusion.  I refer to expectations of remote events which allow of frequent renewal.  Even supposing the expectation to have originated from some rational source, as from a conscious inference from past experience, or from the acceptance of somebody’s statement, the very habit of cherishing the anticipation tends to invest it with an automatic self-sufficient character.  To all intents and purposes the prevision becomes intuitive, by which I mean that the mind is at the time immediately certain that something is going to happen, without needing to fall back on memory or reflection.  This being so, whenever the initial process of inference or quasi-inference happens to have been bad, an illusory expectation may arise.  In other words, the force of repetition and habit tends to harden what may, in its initial form, have resembled a kind of fallacy into an illusion.