Perhaps not “save” the 21st century in the political or technological sense. But surrealism may still save something more basic: our capacity not to become completely formatted by reality as administered to us.
The 21st century is increasingly a world of optimization, prediction, dashboards, recommendation systems, metrics, identity categories, risk scores, algorithmic sorting, and AI-generated plausibility. Everything tends toward legibility. Surrealism is useful precisely because it reintroduces the illegible, discontinuous, dreamlike, excessive, useless, and unassimilable.
Breton’s surrealism tried to break the monopoly of rational, bourgeois reality by giving dream, chance, automatic writing, eroticism, coincidence, and the unconscious equal standing. Today the enemy is somewhat different. It is no longer merely bourgeois rationality. It is computational rationality: the assumption that whatever matters can eventually be represented, predicted, ranked, and optimized.
That makes surrealism unexpectedly contemporary.
Its most important political gesture may be this: it insists that a human being is not exhausted by his profile. There is always something that does not fit the database. A dream that makes no sense. An attraction with no reason. A ridiculous association. A sentence that arrives from nowhere. A useless walk. A door on the 7½ floor.
In that sense, Being John Malkovich is already a late-surrealist film. The portal is not merely fantasy. It is an ontological insult to ordinary reality. And crucially, nobody spends much time asking how the portal scientifically works. It simply exists. Surrealism says: reality contains trapdoors.
I would distinguish, though, between two kinds of surrealism.
One is now very easy to manufacture: bizarre images, impossible architecture, hybrid creatures, psychedelic video, AI dreamscapes. We are drowning in this kind of visual surrealism. Generative AI can produce it endlessly. Paradoxically, therefore, strangeness itself has become cheap.
The more valuable surrealism today would be existential surrealism: living in such a way that one does not allow the world’s dominant classifications to determine in advance what can happen.
That means preserving coincidence, wandering, uselessness, free association, encounters without purpose, and objects whose meaning has not yet been assigned. The surrealist practice of objet trouvé becomes especially interesting here: you encounter something first and understand it later—or perhaps never.
This comes quite close to your idea of the fragment. A fragment arrives before its meaning. The person who produces it may not know what it means. Its connections become visible only retrospectively. That is almost the opposite of algorithmic production, where the goal, category, and evaluation criterion are specified in advance.
So I might formulate the conflict this way:
Algorithm: “Tell me what you want, and I will predict the next appropriate thing.”
Surrealism: “I don’t yet know what I want. Show me what should not have been there.”
And this is where surrealism intersects surprisingly well with psychoanalysis. Freud’s unconscious is surrealist not because it is fantastical, but because it violates orderly causality: displacement, condensation, jokes, slips, symptoms. Something irrelevant suddenly becomes decisive. The analyst’s free-floating attention is therefore a kind of disciplined surrealism: do not decide too quickly what matters.
The same applies to wandering. The surrealists wandered Paris; Benjamin made the flâneur philosophical; the Situationists later developed dérive. One walks without allowing destination completely to govern perception.
That matters in an age when Google Maps already knows the efficient route, Spotify knows the next song, Amazon knows the next purchase, TikTok knows the next stimulus, and increasingly AI knows the next sentence.
The radical act might therefore be astonishingly modest:
take the wrong turn.
Not because wrongness is inherently virtuous, but because a world from which every detour has been removed becomes a world without discovery.
There is, however, one danger. Surrealism cannot simply celebrate irrationality. The 21st century already contains enormous quantities of conspiracy, hallucination, propaganda, manufactured unreality, and epistemic chaos. We do not need less ability to distinguish truth from falsehood.
So the surrealism worth preserving must have a double discipline:
be rigorous about facts, but anarchic about meanings.
A bridge either collapsed or it did not. A government either enacted a law or it did not. A drug either showed benefit in a trial or it did not. Surrealism should not corrupt that level.
But once the facts are established, what they might mean, what they might connect to, what images they awaken, what unexpected philosophical neighborhood they belong to—there we should resist premature closure.
Thus my answer is: surrealism cannot save civilization from climate change, dictatorship, war, or badly aligned AI by itself.
But it may help save the part of civilization worth saving: the capacity for unexpected association, unprogrammed encounter, useless beauty, dream, humor, eroticism, wandering, and the sudden appearance of something for which no menu option yet exists.
Or, more simply:
The 20th-century surrealists wanted to liberate dream from reality.
The 21st-century task may be to liberate reality from the algorithm.
