Roger Penrose is the most senior thinker on the reading list, the most technically demanding to write about, and the most uncomfortable for the surrounding scientific consensus. He is in his nineties, has been working at the highest level of mathematical physics for sixty years, and continues to publish positions that most of his colleagues consider either wrong, over-reaching, or both. The 2020 Nobel Prize for his proof that black hole formation is a robust prediction of general relativity established his place in the canon. The work on consciousness that has occupied much of his later career sits outside the canon and is treated by most working scientists as the eccentric speculation of a man whose technical contributions are otherwise indisputable.

I want to write this piece without pretending to expertise I do not have on the underlying physics. The argument I do think is worth making is more philosophical than technical. Penrose's stubborn insistence that consciousness is a real phenomenon that requires real explanation, rather than an illusion to be dissolved or a software-level emergent property of any sufficiently sophisticated computation, is a useful correction to a dominant cultural posture that treats the question as already settled. Whether his specific proposal about how consciousness arises is correct is a separate question, and probably not. The insistence is what matters.

Who he is

Roger Penrose, born 1931 in Colchester, England, doctorate in mathematics from Cambridge, long career across Birkbeck, Cornell, Texas, and Oxford. The mathematical and physics output has been substantial and varied. The Penrose-Hawking singularity theorems of 1965-1970 are foundational results in general relativity, showing that under reasonable physical conditions, gravitational collapse must produce a singularity. Twistor theory, developed from the 1960s onwards, is a mathematical framework that recasts spacetime in terms of complex projective spaces and continues to be a live area of theoretical research. Penrose tilings, the non-periodic plane tilings he developed in the 1970s, turned out to model real quasi-crystalline atomic arrangements that physicists subsequently discovered. The 2020 Nobel Prize was for the singularity work and capped a career most physicists would consider distinguished by any standard.

The other half of his work is more contested. The Emperor's New Mind (1989) and its sequel Shadows of the Mind (1994) developed an argument that human consciousness, particularly human mathematical insight, cannot be reproduced by classical computation, and that something quantum-mechanical must be involved at the level of brain function. With the anaesthesiologist Stuart Hameroff he developed Orchestrated Objective Reduction (Orch-OR), the proposed mechanism: that consciousness arises from quantum processes in microtubules within neurons, with quantum-state collapse producing moments of conscious experience. The theory has been the subject of more than thirty years of intense criticism and limited but ongoing experimental investigation.

The Road to Reality (2004), at over 1100 pages, is his attempt to lay out the entire mathematical and physical framework underlying modern theoretical physics, intended for the serious general reader who is willing to do the mathematical work. It is one of the more remarkable single-author works of scientific exposition produced in the last fifty years. Cycles of Time (2010) presents conformal cyclic cosmology, his proposed cosmological model in which the universe is one of an infinite series of "aeons" succeeding each other across mathematical transformations of spacetime.

He is, in summary, a working mathematician at the highest level who has refused to treat the largest unanswered questions in science (the singularities at the start and end of the universe, the nature of consciousness, the relationship between mathematical truth and physical reality) as off-limits or as adequately handled by the consensus.

The Three Worlds picture

The simplest entry into Penrose's overall position is what he calls the Three Worlds picture. There is the physical world: matter, energy, spacetime, the things that can be measured and manipulated. There is the mental world: consciousness, perception, understanding, the things that are experienced from the inside. And there is the Platonic mathematical world: the truths of mathematics that hold whether or not anyone is around to think about them.

Each world participates in the others. The physical world apparently produces minds, somehow. Minds apparently grasp mathematical truths, somehow. Mathematical structures apparently underwrite the physical world, somehow. Each "somehow" is, on Penrose's account, a genuinely unsolved scientific problem, and pretending otherwise is intellectually dishonest. The dominant scientific posture is to treat one or two of these problems as solved and to focus on the others. Penrose's posture is to treat all three as genuinely open and to try to think across them.

This is unfashionable. Mainstream physicalism wants to deny that the mental world has any reality independent of the physical, or that the mathematical world has any reality independent of the brains that think about it. Mainstream computationalism wants to treat consciousness as a software-level phenomenon emerging from sufficient computational complexity, which Penrose argues fails to engage with the actual problem. Mainstream string-theoretic and multiverse cosmology wants to treat the apparent fine-tuning of physical constants as a selection effect among an infinite ensemble of possible universes, which Penrose argues is a non-explanation.

The Penrose alternative does not provide settled answers. It insists that the questions are real and that the dominant attempts to dissolve them are intellectually inadequate. A reader does not have to accept any of his specific positive proposals to find the negative argument useful.

The consciousness argument

The argument runs roughly as follows.

Gödel's incompleteness theorems show that any sufficiently powerful formal system contains true statements that cannot be proved within the system. Penrose argues that human mathematicians can recognise the truth of certain Gödelian statements (specifically, statements that say "I am not provable in this formal system") while being unable to derive them by formal proof. This recognition, on his account, requires a kind of insight that is structurally non-computable, in the sense that no formal-algorithmic system can perform it.

If human mathematical insight is non-computable in this sense, and if the brain is a physical system, then the brain must be doing something that exceeds what classical computation can do. The candidate Penrose proposes is quantum-mechanical processes that involve genuine quantum-state collapse, which on his account is not itself a computational process and is in fact one of the major unsolved problems of fundamental physics. The specific mechanism, developed with Hameroff, is that quantum-coherent states in microtubules within neurons undergo orchestrated objective reduction (objective in the sense of being a real physical process rather than an observer-relative one) and that these reductions are the moments of conscious experience.

The argument has been extensively criticised. The Gödel argument is rejected by most working mathematicians and philosophers as not licensing the strong conclusion Penrose draws. The microtubule mechanism is considered speculative by most neuroscientists, partly because the brain is too warm and wet to sustain the kind of quantum coherence required for the timescales involved, partly because no specific experimental signature uniquely supporting Orch-OR has been observed in three decades. There has been some recent work (Bandyopadhyay's group, the Posner-molecule speculation) that has revived interest, but the theory remains unverified and is widely considered unlikely.

My own view, for what it is worth, is that the specific Orch-OR proposal is probably wrong. The underlying argument that consciousness presents a real explanatory problem that conventional computational accounts fail to address is, however, much stronger than its critics typically acknowledge. The hard problem of consciousness remains genuinely hard and the dominant attempts to dissolve it (functionalism, eliminativism, illusionism) all involve moves that look more like changing the subject than answering the question.

Where he is right

Three places where the work earns its standing regardless of the specific consciousness proposal.

The mathematical physics is real and important. The singularity theorems, twistor theory, and Penrose tilings are settled contributions to the foundations of theoretical physics and mathematics. Anyone trying to assess the consciousness work has to do so against a background in which the technical contributions are not in serious dispute. Penrose is not a fringe figure who happened to develop unusual views on consciousness. He is one of the most technically capable mathematical physicists of the late 20th century who happened also to develop unusual views on consciousness.

The insistence that consciousness is a real explanandum is correct and unfashionable. The dominant cultural and scientific posture in the last several decades has been to treat consciousness either as an illusion that careful philosophical analysis will dissolve, or as an emergent property of any sufficiently sophisticated computational process, in either case requiring no separate explanation. Both moves involve substantial intellectual sleight-of-hand. Penrose's refusal to accept either is not eccentricity. It is rigour. The hard problem of consciousness, as articulated by David Chalmers in 1995, remains the actual philosophical problem, and any account of mind that does not engage with it directly is in some sense dodging the question.

The Platonist position about mathematical truth has more going for it than its critics often acknowledge. The reasonable effectiveness of mathematics in the natural sciences, the apparent objectivity of mathematical truth across cultures and across mathematical practitioners, the experience of mathematical discovery as discovery rather than invention reported by working mathematicians, are not nothing. The conventional alternative (mathematical truths as social conventions, or as features of human cognition projected onto the world) has its own difficulties and is not obviously more parsimonious than the Platonist account. Penrose's willingness to defend Platonism in a generally anti-Platonist scientific culture is intellectually substantive.

Where he is vulnerable

Worth being honest about.

Orch-OR has not produced the experimental confirmation it would need to count as a working scientific theory. Three decades of investigation have not yielded the unambiguous experimental signature that would shift the consensus. The theory remains, at this point, a live possibility rather than a confirmed mechanism. Penrose has continued to defend it against criticism in ways that some readers find admirable and others find stubborn. My reading is that the specific proposal is probably wrong even if the underlying intuition (that consciousness involves something beyond classical computation) turns out to be defensible.

The Gödelian argument is contested by serious philosophers and mathematicians and the strong form Penrose defends in Shadows of the Mind has not been generally accepted. The objection that human mathematicians make mistakes, that the recognition of mathematical truth is fallible, and that the asymmetry between human insight and algorithmic derivation is therefore weaker than Penrose's argument requires, has not been adequately answered.

The Orch-OR mechanism in particular has the appearance, to some critics, of being motivated by what is needed to support the philosophical conclusion rather than by independent biological evidence. The microtubule choice, the temperature requirements, the specific quantum coherence claim, all read to many neuroscientists as having been selected for their fit with the desired conclusion rather than discovered as independent biological facts. This is not necessarily wrong, but it is an inferential pattern that responsible readers should notice.

The current AI moment is putting Penrose's anti-computationalist position under increasing pressure. The performance of large language models, while not consciousness in the philosophically interesting sense, has demonstrated that computational systems can do substantially more of the work that we previously thought required something beyond algorithm. The extent to which this undermines Penrose's specific argument is contested. The argument that human consciousness is non-computable depends on a specific technical claim about Gödelian recognition that the LLM evidence does not necessarily speak to. Whether the broader intuition holds in light of recent AI progress is, however, an open question that the work needs to engage with more directly than it has.

These are not fatal to the broader insistence that consciousness is real and requires explanation. They are reasons to treat the specific proposals as live hypotheses rather than as settled science.

How it lands in Ireland

Ireland has, contrary to the impression most Irish people have, a genuinely substantial mathematical-physics culture. The Dublin Institute for Advanced Studies (DIAS), founded by de Valera in 1940 and modelled on the Princeton Institute, has produced and hosted some of the most consequential theoretical physicists of the 20th century. Erwin Schrödinger spent the 1940s and 1950s at DIAS and wrote What Is Life? (1944) there, the book that motivated Watson and Crick's discovery of DNA structure. John Bell, born in Belfast, did the work at CERN that produced Bell's inequalities and the experimental foundations of modern quantum-information theory. Lochlainn O'Raifeartaigh's work on gauge symmetries was foundational. The current Irish mathematical-physics community remains internationally significant.

The relevance of Penrose's work to this Irish tradition is substantial and underexplored. Penrose's twistor theory is a working area at DIAS. Penrose's Platonism is intellectually compatible with the Irish-Catholic philosophical tradition that took mathematical and metaphysical objects seriously, in ways that the dominant Anglo-American materialist tradition does not. The Irish poet-philosopher tradition discussed in the McGilchrist piece is doing closely related work from a different starting point. There is, in principle, an Irish intellectual position that would integrate mathematical-physics rigour with the Penrosean Three-Worlds picture and the McGilchrist hemispheric account into something philosophically coherent. None of the current Irish public discourse is doing this work.

The Irish public conversation about consciousness, mind, and AI is currently held in vocabularies that cannot reach what is at stake. Religious frameworks treat the question as already answered by theology. Mainstream secular discourse treats it as already answered by neuroscience or as not really a question at all. The Penrosean position, that consciousness is a real phenomenon presenting a genuine open problem in physics and philosophy, is largely absent from Irish public engagement. This matters because the AI moment is putting these questions back at the centre of public concern in ways the existing Irish vocabularies are not equipped to handle.

A specific implication: Irish education, particularly at second-level, has substantially abandoned the kind of mathematical and philosophical formation that would let citizens engage with these questions at adult depth. The current curricular emphasis on procedural mathematical competence and on standardised philosophical material does not produce the kind of formation Penrose's body of work is written for. Whether Irish education recovers this depth in the next generation is a question of cultural and institutional choice rather than of resource availability.

Where to start

If you have an evening: any of the longer Penrose interviews from the last few years, particularly his 2020 and 2022 conversations with Lex Fridman or the 2018 Closer to Truth interviews. He is, surprisingly, an excellent interviewee at his current age, with the technical depth and the willingness to think slowly on camera that the format rewards.

If you have a week: The Emperor's New Mind (1989). Despite its age, this is still the most accessible introduction to the consciousness argument and to Penrose's broader scientific worldview. The first half (general physics and cosmology) is excellent introductory material. The second half (the consciousness argument specifically) should be read with awareness that the field has moved on and that the specific mechanism has not been confirmed.

If you have a month: the first 200 pages or so of The Road to Reality (2004). The whole book is a magnificent but enormous undertaking. The first 200 pages cover the mathematical foundations and are a genuinely useful introduction to the formalism that underlies modern physics, written for the serious general reader. Anyone willing to do the mathematical work will be substantially repaid.

For the consciousness debate broadly: David Chalmers, The Conscious Mind (1996), is the standard statement of the hard-problem position. Daniel Dennett, Consciousness Explained (1991), is the canonical opposing view. Reading the two against Penrose gives a reader the major positions in the debate.

For the Irish-specific extensions: Schrödinger's What Is Life? (1944), written at DIAS, remains worth reading. John Bell's collected papers (in Speakable and Unspeakable in Quantum Mechanics) are essential for anyone interested in the foundations of quantum mechanics and have been substantially under-read in Ireland relative to their importance.

The thing Penrose offers, that almost no other major contemporary scientist offers, is the willingness to take the largest unanswered questions in science as genuinely open and to bring serious technical capability to thinking about them. Whether his specific answers are right is contested. The standard he sets, that scientific work should engage with the actual problems rather than redefining the problems out of existence, is the part of his legacy most worth carrying forward.

Related in the Political Literacy series

  • Iain McGilchrist — the neuroscientist-philosopher whose hemispheric work pairs with Penrose's consciousness work
  • Andrej Karpathy — the AI/ML position Penrose's anti-computationalist argument is in tension with
  • Daniel Schmachtenberger — the civilisational-systems frame on the same large open questions
  • Yuval Noah Harari — the big-history popularisation of related questions
  • Alex O'Connor — public dialogues engaging with consciousness, mind, and metaphysics

Plus the framing piece, What Do Ireland's Parties Actually Stand For?, and the full Political Literacy archive.