As Segal has it in The Function of Dreams, Freud didn't have the concept of 'working through' at the time of writing The Interpretation of Dreams. But once we have that concept available to us we can start to see dreams - 'good dreams' at least - as not just a defence against, but as the constructive processing of, difficult emotional experiences. For Freud, dreams are compromises of repressing and repressed forces; for him only the repressed was symbolised. In Segal's hands, however, 'symbolism' comes to mean something much more like what Langer means by it: that is, the expressive bodying forth of meaningful action tendencies whereby difficult emotional experience is given shape and representation, and so can be renewingly woven into the flow of an ongoing subjective life, regardless of whether the symbolising subject can now avow an intentional object for his emotion. (This isn't surprising: Segal read Langer in the 1940s.) What I want to stress here is that this isn't a different theory of symbolism: it is a different concept of symbolism. In fact it's a rather radically different concept: for Freud, symbolism has to do just with what is unthinkable; for the later Kleinians it becomes the process whereby difficult experience becomes thinkable. The difference is rather underplayed by Segal. If I had a psychoanalytic fiefdom I think I should insist there on 'symbolism' for the Freudian, and 'symbolisation' for the Langerian, phenomenon.
As is well known, Segal enriched psychoanalytic theory by distinguishing between symbolisation proper and symbolic equation. The latter involves a collapse of symbol and symbolised, something we may also call 'concrete' symbolising. Segal often talks of the psychotic patient not being able to tell apart the symbol and the symbolised, but a more helpful way of putting this would be in terms of an actual indistinction - and not just 'for them' - between thoughts that treat of metaphoric and those that treat of literal truth, between self and object. This occurs, Segal tells us, when projective identification is in the ascendant: i.e. when aspects of the self are truly lost from it and instead become lodged in the self's representations of its objects. The distinction becomes important when we turn to psychotic dreaming. Symbolisation proper, Segal tells us, by contrast, can only obtain when the subject is capable of mourning. After all, one might think, the ability to tolerate the absence of the object is essential before one can represent it as absent - and this capacity is essential to a symbol functioning as such.
In true psychosis there is no difference between wake and sleep, hallucination and dream. (In the ambit of the complexes, that is; it's not as if the psychotic subject is always inexorably psychotic, nor the sane subject always inexorably sane.) Segal suggests that the patient may treat their ‘dream’ as a real happening. By contrast we may suggest (with e.g. Sass) that the schizophrenic patient experiences reality as having the subjective quality of a dream. In fact, rather than say either of those, I propose that we stick instead to an ontological, and avoid the epistemological, claim: the schizophrenic's mind does not, when he is in his autism, instantiate in its very form the distinction between reality and fantasy. Dream and reality are of a piece. This is because they are in that radically dissociated - 'autistic' - state of mind. They've lost 'reality contact'.
Borderline psychotic subjects, by contrast, ‘can use dreams for getting rid of, rather than working through, unwanted parts of the self and objects’. Hence the curious Kleinian talk of ‘projection into a dream’.
It is the use, rather than the content, of dreams which comes more to the fore in Kleinian analysis, and this affects the kinds of interpretations given when working with borderline subjects.
In chapter 7 of Learning from Experience, Bion proposes that the function of dreaming is to create a distinction of conscious emotional experience from unconscious emotional experience. He describes this in concrete terms: as a semi-permeable 'barrier' composed of alpha elements. I cannot find a way to treat this as more than a mythology useful for keeping our eye on the relevant phenomena whilst yet remaining explanatorily impotent. A way to think about the creation of this conscious/unconscious distinction is in terms of the narrative structure of dreaming, which binds salient tolerable emotional experience into a thinkable/dreamable whole and thereby excludes otherwise overwhelming unmanageable proto-thoughts and proto-feelings (which now can be said to be unconscious). Psychosis, then, involves the failure of dreaming to create the distinction between the conscious and unconscious.
Whilst a mentally healthy person can be both awake and asleep, the psychotic person cannot be either when in her psychosis.
Here is a helpful passage from Learning from Experience. Unlike in Freudian theory,
in alpha-function theory the powers of censorship and resistance are essential to differentiation of conscious and unconscious and help to maintain the discrimination between the two. This discrimination derives from the operation of the “dream”, which is a combination in narrative form of dream thoughts, which thoughts in turn derive from combinations of alpha-elements. In this theory the ability to “dream” preserves the personality from what is virtually a psychotic state. It therefore helps to explain the tenacity with which the dream, as represented in classical theory, defends itself against the attempt to make the unconscious conscious. Such an attempt must appear indistinguishable from destruction of the capacity to dream in so far as that capacity is related to differentiating conscious from unconsciousand maintaining the difference so established.To end I should like to share my own Bionian impression of the psychotic dream. Which is that it is a dream which is violent, disturbing, a fusing of love and hate, containing objects broken into pieces, not rich in imagery, with barely any development, etc. Perhaps it is just of some dangerous teeth or threatening faeces. It's a dream which hardly functions as such since the narrative containing function barely gets going.