What dreams are for: the psychoanalytic view

Freud's theory in four ideas

Sigmund Freud published The Interpretation of Dreams in 1900. Four ideas from it still organise the psychoanalytic view.

1. A dream has a meaning. It is not noise. It can be understood.
2. Two layers. The dream as you remember it is the manifest content. The thoughts it came from are the latent content. Freud describes the manifest dream as a translation of the dream-thoughts.
3. The dream-work. The process that turns the latent into the manifest has named parts. In condensation, several thoughts or people are merged into one image. In displacement, emotional weight moves from something important to something trivial, so that the dream's centre seems to lie somewhere else.
4. Meaning comes from the dreamer. In his second chapter Freud sets his method against the popular dream book, the key that assigns a fixed meaning to each image. He rejects that "decoding" method because it has no guarantee of being right. His own method asks the dreamer to report whatever comes to mind about each piece of the dream, without censoring it. The same image can mean different things to different people, so no universal dictionary can work.

The last point matters most for what follows.

Where Jung differed

Carl Jung agreed that dreams are meaningful products of the unconscious. He disagreed about how they work.

• No disguise. Jung held that the remembered dream is not a facade. In his view it expresses what it means, in a language that is concrete, emotional and layered.
• Compensation. He saw dreams as restoring balance: they tend to present the viewpoint the conscious attitude overlooks.
• Amplification. Besides the dreamer's associations, he widened an image with parallels from myth, fairy tale and religion, circling back to the dream image rather than moving away from it.
• Symbols are not signs. Jung distinguished a sign, which stands for a known thing, from a symbol, which points to something not yet consciously recognised and can carry many meanings. So he too had no dictionary.

These summaries come from a Jungian centre's account and a Wikipedia article, both secondary. For exact wording, read Jung's own texts.

What sleep science says

The science does not settle the question, and parts of it contradict Freud.

• Brainstem origins. Hobson and McCarley (1977) proposed that dreaming is physiologically determined: a brainstem mechanism periodically activates the forebrain, which synthesises the dream by comparing that activity with stored memory. On this view, a dream's strangeness reflects the brain's improvisation, not a disguise.
• Dreaming is not just REM. Solms (2000) argued that dreaming and REM sleep are controlled by different mechanisms. He pointed to forebrain systems, probably dopaminergic, as the final common pathway to dreaming, and to evidence that dreaming can change without a change in REM measures. This weakens the older equation of REM with dreaming. It does not by itself confirm Freud.
• Emotion shapes the image. Hartmann proposed that a dream's most vivid image, its central image, pictures the dreamer's dominant emotion, and that intensity of image tracks intensity of feeling. We read this through a secondary summary of his 2008 article.
• Continuity with waking life. Schredl and Hofmann (2003) found continuity between waking activities and dreams to be supported, with emotional involvement affecting how much of waking life appears. They also concluded that the general continuity hypothesis "in its present general form is not valid" and needs refining.

Taken together: dreams do seem related to concerns and feelings from waking life. Whether they carry a hidden, disguised wish that interpretation can recover has not been shown, and the main alternative accounts do not require it.

How dream work is done in practice

In a psychoanalytic session, the dream is told, then the dreamer is asked what comes to mind about each part of it. The therapist listens for what the dreamer circles back to, what is said with unusual feeling, and what is skipped. The analyst does not announce a meaning. They offer a guess and watch whether it lands. If it does not, it is dropped.

This is why the same dream can be productive in one hour and empty in the next, and why a reading from someone who does not know the dreamer is weak evidence. The method depends on a person's associations, not on the dream text alone.

A type of situation many people recognise: a dream returns to a childhood house, with a room you never entered. Neither Freud nor Jung would say what the room means. Each would ask what the dreamer thinks of when they picture it. A fixed answer such as "the room is a hidden part of yourself" answers a question the dreamer did not ask.

What stays contested

Freud's central claim, that a dream is a disguised wish whose meaning interpretation can recover, is not established by sleep research. Hobson and McCarley explained the dream's strangeness without a disguise; Solms argued for a motivational forebrain system and so is sometimes read as friendlier to the psychoanalytic view, though the abstract we read does not say that. Continuity research supports a link with waking life, not a hidden layer. Practitioners also differ: Freudians, Jungians and cognitive dream researchers use different methods and measure success differently.

A modest position is available. Dreams are often related to what matters to the dreamer, and talking about them can bring that into view. That does not require a hidden code.

Where AI dream apps go wrong

A common design is the symbol dictionary: the app finds a word in your dream, such as a house, water or a key, and returns a stored meaning. By the logic of both Freud and Jung this is the wrong tool. Freud rejected the fixed key because the meaning of an image depends on the dreamer. Jung treated symbols as open, not as signs with a single referent. A dictionary also gives a confident answer where no confident answer has been shown to exist.

The risk is practical. A fixed meaning offered with authority can send someone down a wrong line, or settle on a reading that feels satisfying and closes the question.

What TwentyThird claims, and does not

TwentyThird lets a member write a dream down on waking. Oneli, an AI agent grounded in psychodynamic thinking, reads it with them and names what the dream keeps returning to, in plain words. It does not look symbols up. It notices what recurs and asks about it the next time you talk. A setting for psychologists describes the aim as exploring a dream's images with the member rather than decoding it.

What it does not claim:

• It does not tell you what a dream truly means.
• It does not recover unconscious wishes. That outcome has not been shown by sleep research and cannot be checked.
• It is not therapy and not a replacement for dream work with a clinician.

The example dream shown on the method page is invented, and is there to show how the page marks recurring phrases. Psychologist supervision is in early access today; no human is alerted when a member writes something alarming. See what TwentyThird is not.

What this does not replace

A dream is most useful when explored with someone who knows your life, who can ask what each image brings to mind for you. If disturbing dreams are frequent, or come with distress you cannot manage, speak to a clinician. If you are in danger, contact local emergency services now. In the EU, call 112.