My idea of a particular friend contains, among other things, numbers of vague representations of his habitual modes of feeling and acting, and numbers of still more vague expectations of how he will or might feel and act in certain circumstances.
Now, it is plain that such a composite idea must have been a very slow growth, involving, in certain stages of its formation, numerous processes of inference or quasi-inference from the past to the future.
But in process of time these elements fuse inseparably: the directly known and the inferred no longer stand apart in my mind; my whole conception of the individual as he has been, is, and will be, seems one indivisible cognition; and this cognition is so firmly fixed and presents itself so instantaneously to the mind when I think of the object, that it has all the appearance of an intuitive conviction.
If this is a fairly accurate description of the structure of these compound representations and of their attendant beliefs, it is easy to see how many openings for error they cover.

To begin with, my representation of so complex a thing as a concrete personality must always be exceedingly inadequate and fragmentary.

I see only a few facets of the person's many-sided mind and character.

And yet, in general, I am not aware of this, but habitually identify my representation with the totality of the object.
More than this, a little attention to the process by which these compound beliefs arise will disclose the fact that this apparently adequate representation of another has arisen in part by other than logical processes.

If the blending of memory and expectation were simply a mingling of facts with correct inferences from these, it might not greatly matter; but it is something very different from this.