So far we have been attempting to ascertain how the dream represents the relations among the dream-thoughts, but we have several times extended our consideration to the further question of what alterations the dream material undergoes for the purposes of dream formation. We now know that the dream material, after being stripped of the greater parts of its relations, is subjected to compression, while at the same time displacements of intensity among its elements force a psychic revaluation of this material. The displacements which we have considered were shown to be substitutions of one idea for another, the substitute being in some way connected with the original by associations, and the displacements were put to the service of condensation by virtue of the fact that in this manner a common mean between two elements took the place of these two elements in the formation of the dream. We have not yet mentioned any other kind of displacement. But we learn from the analyses that another exists, and that it manifests itself in a change of the verbal expression employed for the thought in question. In both cases we have displacement following a chain of associations, but the same process takes place in different psychic spheres, and the result of this displacement in the one case is that one element is substituted for another, while in the other case an element exchanges its verbal expression for another.

This second kind of displacement occurring in dream formation not only possesses great theoretical interest, but is also peculiarly well fitted to explain the semblance of phan-tastic absurdity in which the dream disguises itself. Displacement usually occurs in such a way that a colourless and abstract expression in the dream-thought is exchanged for one that is visual and concrete. The advantage, and consequently the purpose, of this substitution is obvious. Whatever is visual is capable of representation in the dream, and can be wrought into situations where the abstract expression would confront dream representation with difficulties similar to those which would arise if a political editorial were to be represented in an illustrated journal. But not only the possibility of representation, but also the interests of condensation and of the censor, can be furthered by this change. If the abstractly expressed and unwieldy dream-thought is recast into figurative language, this new expression and the rest of the dream material are more easily furnished with those identities and cross references, which are essential to the dream activity and which it creates whenever they are not at hand, for the reason that in every language concrete terms, owing to their evolution, are more abundant in associations than conceptual ones. It may be imagined that in dream formation a good part of the intermediary activity, which tries to reduce the separate dream-thoughts to the tersest and simplest possible expression in the dream, takes place in the manner above described - that is to say, in providing suitable paraphrase for the individual thoughts. One thought whose expression has already been determined on other grounds will thus exert a separating and selective influence upon the means available for expressing the other, and perhaps it will do this constantly throughout, somewhat after the manner of the poet. If a poem in rhyme is to be composed, the second rhyming line is bound by two conditions; it must express the proper meaning, and it must express it in such a way as to secure the rhyme. The best poems are probably those in which the poet's effort to find a rhyme is unconscious, and in which both thoughts have from the beginning exercised a mutual influence in the selection of their verbal expressions, which can then be made to rhyme by a means of slight re modification.

In some cases change of expression serves the purposes of dream condensation more directly, in making possible the invention of a verbal construction which is ambiguous and therefore suited to the expression of more than one dream-thought. The whole range of word-play is thus put at the service of the dream activity. The part played by words, in the formation of dreams ought not to surprise us. A word being a point of junction for a number of conceptions, it possesses, so to speak, a predestined ambiguity, and neuroses (obsessions, phobias) take advantage of the conveniences which words offer for the purposes of condensation and disguise quite as readily as the dream.* That dream conception also profits by this displacement of expression is easily demonstrated. It is naturally confusing if an ambiguous word is put in the place of two ambiguous ones; and the employment of a figurative expression instead of the sober everyday one thwarts our understanding, especially since the dream never tells us whether the elements which it shows are to be interpreted literally or figuratively, or whether they refer to the dream material directly or only through the agency of interpolated forms of speech.* Several examples of representations in the dream which are held together only by ambiguity have already been cited ("her mouth opens without difficulty." in the dream of Irina's injection; "I cannot go yet," in the last dream reported, p. 312), etc. I shall now cite a dream in the analysis of which the figurative expression of abstract thought plays a greater part. The difference between such dream interpretation and interpretation by symbolism may again be sharply distinguished; in the symbolic interpretation of dreams the key to the symbolism is arbitrarily chosen by the interpreter, while in our own cases of verbal disguise all these keys are universally known and are taken from established customs of speech. If the correct notion occurs at the right opportunity, it is possible to solve dreams of this sort completely or in part, independently of any statements made by the dreamer.

* Cf. Der Witz und seine Bezichung zum Unbewussten., 2nd edit. 1912, and "word-bridges," in the solutions of neurotic symptoms.