The following is an essay written in 2024; adapted for this blog
I. The Loneliest Philosophy
Solipsism is the philosophical position that only one’s own mind is certain to exist. Everything else, other people, the physical world, the stars, the deep past, the apparent future, may be nothing more than constructions of that single mind. It is, in its strongest form, the claim that you are the only conscious entity in existence and that the entirety of experienced reality is your mental content.
It is also, by most accounts, the most difficult philosophical position to refute. Not because it is likely to be true, but because it is structured in a way that makes counterevidence logically impossible. Any evidence you could present against solipsism, the testimony of other people, the results of scientific experiments, the apparent independence of the physical world, can be absorbed into the solipsist framework as further content of the one existing mind. The theory is not falsifiable, and that is precisely the problem with it.
This essay examines solipsism and its intellectual relatives with a critical lens. The goal is not to prove that other minds exist, since such proof may be impossible in the strict philosophical sense, but to demonstrate that solipsism fails as an explanatory framework, that it is internally incoherent in ways its proponents rarely acknowledge, and that physics, neuroscience, information theory, and evolutionary biology all provide strong reasons to reject it as an account of reality, even if they cannot deliver the kind of deductive certainty that the solipsist demands.
II. Origins and Articulations
Solipsism does not have a single origin. It is better understood as a recurring consequence of certain epistemological starting points, a destination that various thinkers have arrived at by following different paths.
The earliest Western formulation is sometimes attributed to the pre-Socratic philosopher Gorgias of Leontini, who around 440 BCE advanced three claims: nothing exists; if something did exist, it could not be known; and if it could be known, it could not be communicated. Gorgias was likely being deliberately provocative, deploying these arguments as rhetorical exercises rather than sincere metaphysical commitments. But the structure of his reasoning, the progressive retreat from claims about the external world to claims about the isolation of the knowing subject, anticipates the logic of solipsism with remarkable precision.
Descartes arrived at a solipsism-adjacent position through methodological doubt. His project in the Meditations on First Philosophy (1641) was to strip away every belief that could conceivably be false and see what remained. What remained was the Cogito: the certainty that a thinking thing exists. Everything else, the body, the external world, other minds, had to be rebuilt through argument, and Descartes’ reconstruction relied on proving God’s existence and God’s benevolence, a move that most subsequent philosophers have found unconvincing. Without that theological bridge, the Cartesian method terminates in something very close to solipsism. You are certain that you think. You are certain of nothing else.
George Berkeley, the 18th-century Irish philosopher, developed a form of idealism that stopped just short of solipsism. Berkeley argued that material objects do not exist independently of perception. His famous formula, “esse est percipi” (to be is to be perceived), held that the table in front of you exists only insofar as some mind perceives it. Berkeley avoided solipsism by positing that God continuously perceives all things, thereby granting them existence independent of any particular human mind. Remove God from Berkeley’s system and you are left with a single perceiving mind and its contents, which is solipsism.
The German idealist Johann Gottlieb Fichte came closer to making solipsism explicit. In his Wissenschaftslehre (1794), Fichte argued that the self posits both itself and the not-self, meaning that the external world is, in some sense, a product of the ego’s activity. Fichte did not consider himself a solipsist, and his system is considerably more complex than a summary can convey, but the structural similarity is clear. The self generates reality. Everything encountered is a product of the self’s own positing activity.
In the 20th century, solipsism surfaced less as a position anyone endorsed and more as a problem that epistemology needed to solve or dissolve. Ludwig Wittgenstein engaged with it extensively in the Tractatus Logico-Philosophicus (1921) and later in Philosophical Investigations (1953), ultimately arguing that solipsism, when worked out consistently, collapses into realism, a claim we will return to later. The logical positivists attempted to dismiss it as meaningless on the grounds that it could not be empirically verified or falsified. And the phenomenological tradition, particularly Edmund Husserl and Maurice Merleau-Ponty, tried to show that consciousness is always already directed toward a world and toward others, making solipsism a distortion of the structure of experience itself.
III. The Steel-Man Case
Before criticizing solipsism, it is worth understanding why it is difficult to dismiss. The strongest version of the argument proceeds as follows.
All knowledge is mediated by experience. You do not have direct access to the external world. You have access to sensory impressions, which your brain processes into a coherent model of reality. That model feels immediate and transparent, but it is a construction. You do not see the world as it is. You see the world as your perceptual apparatus represents it. This is not speculative philosophy. It is established neuroscience. The colors you perceive do not exist in the physical world; they are your visual system’s interpretation of electromagnetic wavelengths. The solidity of objects is your haptic system’s rendering of electromagnetic repulsion between atomic shells. The sounds you hear are your auditory system’s translation of pressure waves in air. At every level, what you experience is a model, not reality itself.
Given that all experience is mediated, you cannot rule out the possibility that the model is generated entirely internally, with no external world giving rise to the sensory inputs. You might be a brain in a vat receiving electrical stimulation. You might be a Boltzmann brain, a momentary fluctuation in entropy that spontaneously generates a conscious experience complete with false memories of a past and false expectations of a future. You might be dreaming. You might be the only mind in existence, generating the appearance of a world and other minds as part of your own mental activity.
The critical point is that there is no observation you could make that would distinguish between these scenarios and the conventional assumption that an external world exists and other minds inhabit it. Every piece of evidence you encounter, every scientific experiment, every conversation with another person, is itself part of the model. It cannot serve as evidence for the existence of something outside the model, because you have no way to step outside the model to check.
This is a genuine epistemological problem. It cannot be solved by hand-waving, by appealing to common sense, or by noting that solipsism feels absurd. Feeling absurd is not the same as being false. And the fact that virtually no one actually believes solipsism does not constitute an argument against it. Virtually no one believed heliocentrism for most of human history either.
IV. Intellectual Relatives
Solipsism exists within a family of related positions, each of which shares some of its structure while differing in scope or emphasis.
Metaphysical idealism holds that reality is fundamentally mental in nature. The physical world either does not exist independently of mind or is itself a manifestation of mind. Berkeley’s idealism is one version. Hegel’s absolute idealism, in which all of reality is the self-development of Spirit (Geist), is another. Idealism does not necessarily entail solipsism, since it can posit multiple minds or a universal mind, but it shares with solipsism the denial that matter exists independently of perception or thought.
Phenomenalism, associated with John Stuart Mill and later with A.J. Ayer, holds that physical objects are best understood as “permanent possibilities of sensation.” A table is not a material object that exists when no one is looking at it; it is a stable pattern of sensory experiences that would occur under specified conditions. Phenomenalism does not deny the existence of other minds, but it reduces the physical world to actual and possible experiences, which makes it difficult to explain what grounds the “permanence” of those possibilities when no one is around to have them.
Skepticism about other minds, sometimes called the “problem of other minds,” is a more targeted version of solipsism. It accepts the existence of the physical world but questions whether other humans (or animals, or machines) have inner experiences. You can observe other people’s behavior, including their verbal reports of subjective experience, but you cannot observe their experience itself. The inference from behavior to consciousness is an inference by analogy from your own case, and analogical arguments are not deductively valid. This is the problem that makes the question of AI consciousness so difficult, as discussed in the companion essay.
The simulation hypothesis, popularized by Nick Bostrom in 2003, holds that we may be living in a computer simulation run by a more advanced civilization. In its structure, this is remarkably similar to Cartesian doubt: an external agent (the simulators, rather than Descartes’ evil demon) generates the entirety of our experienced reality, and we have no way to determine from inside the simulation whether it is “real.” The simulation hypothesis differs from solipsism in that it posits the existence of the simulators and their world, but from the perspective of the simulated beings, the epistemic situation is identical. You cannot trust your senses to tell you about the nature of the system that generates your experience.
The Boltzmann brain hypothesis, drawn from statistical mechanics, posits that in an infinite or sufficiently long-lived universe, random thermal fluctuations will occasionally produce a brain-sized region of organized matter complete with false memories and coherent perceptual experiences. If such fluctuations are more probable than the emergence of an entire ordered universe (which, under certain cosmological assumptions, they are), then it is more likely that you are a Boltzmann brain than a biological organism in a real universe with a real 13.8-billion-year history. This is solipsism derived not from philosophy but from physics, and it is taken seriously enough that cosmologists use the absence of Boltzmann brain domination as a constraint on viable models of the universe.
V. The Thermodynamic Objection
The sciences do not address solipsism directly, because solipsism is not a scientific hypothesis. It makes no testable predictions. It is compatible with any observation. But several branches of science produce results that are deeply incongruent with solipsism, and this incongruence, while not constituting logical refutation, provides strong pragmatic grounds for rejection.
The most straightforward scientific objection to solipsism comes from thermodynamics. The second law of thermodynamics states that the entropy of a closed system tends to increase over time. This law is not merely an empirical generalization. It is grounded in the statistical behavior of large numbers of particles: there are overwhelmingly more disordered configurations of a system than ordered ones, so random evolution will almost certainly move toward disorder.
If solipsism is true and reality is the content of a single mind, then the regularities described by thermodynamics require explanation. Why does the imagined world obey the second law? The solipsist’s mind would need to be generating, at every moment, a reality that is consistent with the statistical mechanics of approximately 10^80 particles interacting according to consistent physical laws over 13.8 billion years of apparent history. The information content of this simulation is staggering. It requires the solipsist’s mind to be vastly more complex than the physical universe it is supposedly imagining, which raises the question of what supports the solipsist’s mind and why that substrate is not itself a form of external reality.
This is a version of the argument from complexity. The world as experienced contains regularities, structures, and information that would require an extraordinary substrate to generate. It is more parsimonious, by a factor that is difficult to overstate, to posit that the regularities reflect an actual external world governed by physical laws than to posit that a single mind is fabricating those regularities from scratch while maintaining perfect internal consistency across every domain of experience.
VI. Quantum Mechanics and the Refusal of Privacy
Quantum mechanics poses a subtler problem for solipsism, one that touches on the relationship between observation, measurement, and the nature of physical reality.
In the standard Copenhagen interpretation of quantum mechanics, a quantum system exists in a superposition of states until it is measured, at which point it collapses into a definite state. This has sometimes been interpreted (incorrectly, in the view of most physicists) as implying that consciousness causes wave function collapse, which would seem to support idealism. If mind is required for physical reality to become definite, then perhaps mind is primary.
But this interpretation runs into serious problems. Bell’s theorem, proven by John Stewart Bell in 1964 and experimentally confirmed in increasingly rigorous tests culminating in the 2015 loophole-free experiments by Hensen et al. and the 2022 Nobel Prize in Physics awarded to Alain Aspect, John Clauser, and Anton Zeilinger, demonstrates that quantum mechanics violates local realism. Specifically, measurements on entangled particles show correlations that cannot be explained by any theory in which (a) particles have definite properties prior to measurement, and (b) influences cannot travel faster than light.
The relevance to solipsism is this: Bell’s theorem shows that the correlations between entangled particles are not pre-determined. They are established at the moment of measurement and depend on the measurement settings chosen by the experimenters. If solipsism were true, the solipsist’s mind would need to generate these correlations in real time, maintaining consistency with the predictions of quantum mechanics across all possible combinations of measurement settings, including settings chosen by random number generators or by the light from distant quasars (as in the “cosmic Bell test” experiments). The mind would need to be, in effect, running a full quantum mechanical simulation of the universe at the subatomic level.
This is not impossible in the strict logical sense. A sufficiently powerful mind could, in principle, generate any pattern of experience. But the explanatory cost is enormous. The solipsist’s mind must be at least as complex as the physical system it is simulating, which eliminates any parsimony advantage that solipsism might have had over realism.
Furthermore, the many-worlds interpretation of quantum mechanics, which avoids wave function collapse entirely by positing that all possible measurement outcomes are realized in branching parallel universes, is explicitly incompatible with any form of idealism that privileges a single observer. In the many-worlds framework, there is no special role for consciousness. Branching occurs whether or not anyone is watching. The universe does not require a mind to actualize it.
VII. The Evolutionary Objection
If solipsism is true, then the elaborate cognitive machinery that humans possess for understanding other minds is a colossal waste of resources. And evolution does not waste resources, at least not on this scale.
Humans are equipped with a sophisticated “theory of mind,” the ability to attribute mental states to other beings and to use those attributions to predict behavior. This capacity develops in early childhood (typically between ages 3 and 5), is supported by dedicated neural architecture (including the temporoparietal junction, the medial prefrontal cortex, and the mirror neuron system), and is impaired in specific and well-characterized ways by conditions such as autism spectrum disorder and certain forms of brain damage.
The mirror neuron system, discovered in macaque monkeys in the 1990s by Giacomo Rizzolatti and colleagues and subsequently studied in humans, consists of neurons that fire both when an organism performs an action and when it observes the same action performed by another. This system appears to provide a neural basis for understanding others’ actions and, by extension, others’ intentions and mental states. It is deeply integrated into the primate brain and has been shaped by millions of years of natural selection.
If other minds do not exist, there was no selection pressure to develop this machinery. The solipsist must explain why evolution, a process that operates through differential reproductive success in a population, would produce an organism equipped with elaborate tools for modeling the mental states of entities that do not exist. The evolutionary explanation is straightforward: other minds do exist, they are a critical feature of the environment in which social primates evolved, and the ability to model them confers significant survival and reproductive advantages. The solipsist explanation requires either that evolution itself is part of the illusion (which raises the question of what is generating the illusion and why it is so internally consistent) or that theory of mind serves some other, unknown function.
The evolutionary argument does not prove that other minds exist. It demonstrates that the best available scientific explanation for a well-documented feature of human cognition presupposes the existence of other minds, and that no solipsist-compatible explanation of comparable quality has been offered.
VIII. Wittgenstein’s Private Language Argument
The most philosophically sophisticated attack on solipsism comes from Ludwig Wittgenstein’s later work, particularly the private language argument developed in Philosophical Investigations (1953). The argument is notoriously difficult to interpret and has generated an enormous secondary literature, but its core structure can be stated relatively concisely.
Wittgenstein argues that a language understandable only by a single individual is impossible. Language requires rules, and rules require the possibility of being followed correctly or incorrectly. But a rule that only one person can apply has no external standard of correctness. The individual cannot distinguish between actually following the rule and merely believing they are following the rule, because there is no independent check. Without the possibility of error, there is no rule. And without rules, there is no language.
The relevance to solipsism is this: if private language is impossible, then the language in which the solipsist formulates their position is necessarily public. It presupposes a community of language users. The very act of stating “only my mind exists” employs concepts (mind, existence, only) that derive their meaning from shared use within a linguistic community. The solipsist’s claim is parasitic on the existence of the other minds it denies.
Wittgenstein’s point is not merely that solipsism is socially inconvenient. It is that solipsism is linguistically incoherent. The conditions required for the solipsist to formulate and understand their own position are conditions that the position itself denies. This is a deeper problem than the thermodynamic or evolutionary objections, because it strikes at the internal consistency of the theory rather than its empirical adequacy.
This is what Wittgenstein meant when he wrote in the Tractatus that “what the solipsist means is quite correct; only it cannot be said, but makes itself manifest.” The solipsist is pointing at something real, the privacy and immediacy of subjective experience, but the attempt to articulate this as a metaphysical claim about the non-existence of other minds generates a self-undermining paradox.
IX. Information Theory and the Compression Problem
Claude Shannon’s information theory, developed in 1948, provides another lens through which to evaluate solipsism, one that connects the thermodynamic objection to a more precise quantitative framework.
The information content of experienced reality is immense. A single second of visual experience contains roughly 10 million bits of information from the retina alone (though the brain compresses this substantially). Over a lifetime, the total sensory input amounts to something on the order of 10^16 to 10^17 bits. But this is the compressed, experiential version. The physical processes that give rise to those sensory inputs, including the behavior of approximately 10^80 particles governed by quantum mechanical laws, contain vastly more information.
If solipsism is true, all of this information is generated by a single mind. The mind must produce sensory inputs that are consistent with an enormously complex physical world, including consistency across time, across sensory modalities, and across the testimony of other apparent minds. This is equivalent to saying that the solipsist’s mind must contain or generate at least as much information as the physical universe it claims to be imagining.
Information theory tells us that a system cannot generate more information than it contains. If the solipsist’s mind is generating the appearance of a universe, the mind must be at least as informationally complex as that universe. This does not refute solipsism logically, but it eliminates any theoretical advantage it might claim over realism. If explaining experience requires positing a system of universe-level complexity, it is simpler to posit the universe itself than to posit a mind of equivalent complexity that is generating the illusion of a universe.
The Kolmogorov complexity of experienced reality, the length of the shortest program that could generate it, is a useful concept here. If the world is real, its regularities are explained by physical laws, which are remarkably compressible. The Standard Model of particle physics, general relativity, and thermodynamics can be expressed in a few pages of equations. The laws are simple; the resulting behavior is complex. If the world is a mental construction, the solipsist’s mind must generate all of that complex behavior without recourse to external laws, which means the generating system itself must encode the complexity directly. The solipsist’s mind, in other words, would need to be incompressibly complex, containing within itself the full specification of every particle interaction, every quantum correlation, and every thermodynamic process that constitutes the experienced world.
This is, to put it mildly, a lot to ask of a theory that was supposed to be simpler than realism.
X. The Pragmatic Dissolution
There is a school of thought, associated with William James, Charles Sanders Peirce, and the broader pragmatist tradition, that holds the solipsism debate to be fundamentally misconceived. The pragmatists argue that the meaning of a proposition is exhausted by its practical consequences. If two propositions have identical practical consequences, they are, for all purposes, the same proposition.
Solipsism and realism have identical practical consequences. Whether or not other minds exist, you will still behave as though they do. You will still feel pain when struck, pleasure when comforted, loneliness when isolated, and connection when understood. You will still find that the world resists your will in consistent and often inconvenient ways. You will still discover regularities that allow you to predict future experiences based on past ones. Nothing about your behavior, your decision-making, or your experience changes depending on which metaphysical framework you adopt.
The pragmatist concludes that the question of whether solipsism is “true” is not a meaningful question. It is an intellectual artifact of taking certain epistemological problems more seriously than they deserve. The inability to prove the existence of other minds is a limitation of proof, not of reality. Proof is a tool with a specific scope of application, and that scope does not include the foundations of experience. We do not prove that the ground exists before walking on it, and the impossibility of such a proof does not make the ground less real.
This is not a refutation of solipsism in the logician’s sense. It is a refusal to engage with it on its own terms, a refusal grounded in the observation that the terms themselves are unproductive. The pragmatist is saying: you can hold solipsism as a position if you wish, but it will never make a difference to anything you think, feel, or do, and a position that makes no difference to anything is not a position worth holding.
XI. The Phenomenological Counter
Edmund Husserl and Maurice Merleau-Ponty offer a different strategy for addressing solipsism, one rooted in the analysis of experience itself rather than in science or logic.
Husserl’s approach, developed in the fifth of his Cartesian Meditations (1931), argues that the experience of other minds is built into the structure of perception. When you perceive another human body, you do not first perceive a physical object and then infer that it has a mind. You perceive it as an embodied subject, as a being that perceives you in return. This “appresentation” of the other’s subjectivity is not an inference added to perception; it is part of the perceptual act itself. Husserl calls this “pairing,” a process by which the other’s body is constituted in experience as analogous to your own lived body, carrying with it the same horizon of inner life.
Merleau-Ponty, in Phenomenology of Perception (1945), extends this analysis by arguing that the body is not merely an object in the world but the medium through which the world is experienced. Perception is always embodied, always situated, always oriented from a particular position toward a particular horizon. And the perception of other embodied beings is not a problem to be solved by inference but a basic feature of perceptual life. We encounter others as subjects before we encounter them as objects. The “problem of other minds” arises only when we abstract from this lived experience and ask how a disembodied intellect could know that other intellects exist, a question that presupposes precisely the detachment from embodied experience that makes solipsism seem plausible.
The phenomenological objection does not prove that other minds exist in the way that a mathematical proof establishes a theorem. It argues that solipsism is a distortion of the structure of experience, that it takes a derivative, theoretical attitude and treats it as primary. Lived experience does not begin with an isolated mind wondering whether anything else exists. It begins with a being already entangled with a world and with others. Solipsism is what you get when you extract the subject from that entanglement and ask it to reconstruct everything from scratch, a project that was misconceived from the outset.
XII. Why It Cannot Be Killed
Despite all of these objections, solipsism persists. It persists not because anyone seriously believes it, but because it exposes a genuine gap in our epistemological toolkit. We cannot prove the existence of other minds. We cannot prove the existence of the external world. We cannot step outside our own experience to verify that experience corresponds to something beyond itself. These are real limitations, and solipsism is the position you arrive at if you take those limitations with absolute seriousness and refuse to supplement them with pragmatic, evolutionary, phenomenological, or scientific considerations.
The reason solipsism cannot be killed is that it is not, strictly speaking, a theory about the world. It is a theory about the limits of knowledge. And those limits are real. The solution is not to pretend they do not exist but to recognize that knowledge is not the only tool we have. We also have coherence, parsimony, explanatory power, pragmatic adequacy, and the brute fact that we find ourselves embedded in a world that behaves as though it exists independently of our observation of it. None of these constitute proof. All of them constitute reasons.
The physicist Sean Carroll has argued that we should evaluate metaphysical positions the same way we evaluate scientific theories: by their ability to account for the data of experience in a way that is simple, comprehensive, and predictive. By this standard, realism wins and it is not close. Realism explains the regularities of experience by positing a world governed by physical laws. Solipsism explains the same regularities by positing a mind of incomprehensible complexity generating them for no discernible reason. The second explanation is not wrong in the way that a failed prediction is wrong. It is wrong in the way that epicycles were wrong: it can be made to fit the data, but only at the cost of abandoning everything that makes an explanation useful.
We cannot prove that solipsism is false. But we can observe that it explains nothing, predicts nothing, simplifies nothing, and illuminates nothing. It is a philosophical dead end masquerading as a profound insight. The insight it contains, that subjective experience is private, immediate, and epistemologically privileged, is real. But the conclusion it draws from that insight, that nothing else exists, does not follow. It is a leap from a genuine observation about the limits of knowledge to an unfounded claim about the limits of reality. And that leap, for all its logical tidiness, lands nowhere worth standing.