At the same time a further conclusion follows from these phenomena, which is of importance for the understanding of all the processes of association. It is impossible to imagine that a combination of letters, such as we have given above, could work as a whole, and then, because it was wrong, be replaced by the right word. It is on the contrary obvious that processes of assimilation and displacement have only occurred at certain places. It is also difficult to take for granted that the observer has ever seen the word printed in exactly the same size and type as employed in the reading experiments. It cannot, therefore, be a single definite word-image that he calls to memory, but there must be an indefinite number of similar word-images, which affect assimilatively the given impression, and cast it into the word-form which we ultimately apprehend. From this it follows that these associations do not by any means consist of a combination of complex ideas, but of a combination of ideational elements, which may possibly belong to very different ideas. With this we see that assimilation is at the same time closely connected with the associations by fusion considered above. In both cases the association is an elementary process. The difference between the two forms consists only in the fact that the elements in a fusion are constituent parts of a complex impression, whereas in an assimilation they already belong to complex ideas, from which they then break away in order to enter into new ideational compounds. Thus fusion and assimilation work together in all sense-perceptions. The moment we see an object, hear a musical chord, &c., not only do the parts of the impression itself fuse together, but the impression also immediately gives rise to reproductive elements, which fill up any gaps in it, and arrange it among the ideas familiar to us. These processes continually overlap each other, and extend over all the regions of sense. What we imagine we perceive directly, really belongs in a great extent to our memory of innumerable previous impressions, and we are not aware of a separation between what is directly given us and what is supplied by assimilation. Only when the reproductive elements attain to such a striking ascendancy, that they come into an irreconcilable contradiction with our usual perceptions, are we accustomed to speak of a deception of the senses or of an illusion. But this is only a limiting case, and it goes over by unnoticeable intermediate gradations into normal associations, which we might just as well call "normal illusions." Many words of a lecture are imperfectly heard; the contours of a drawing or painting are only imperfectly represented in our eye. In spite of this we notice none of the gaps. That does not happen because we perceive the things inaccurately, as this phenomenon is often incorrectly interpreted, but because we have at our disposal the rich stores of memory, which fill out and perfect the perceived image.