Navigating the Future of Interdisciplinarity Through Heisenberg and Generative Artificial Intelligence

Abstract

This ISR 50th anniversary reflection examines the future of interdisciplinarity from the standpoint of generative artificial intelligence's potential challenge to Heisenberg's open/closed theory distinction, which has influenced modern thinking about interdisciplinarity. The challenge is epitomised by Don Swanson's idea of ‘undiscovered public knowledge’.

Interdisciplinarity on the Verge of Openness in Knowledge Production

This essay on the future of interdisciplinarity has been written without citations or notes to illustrate an important aspect of that future. As generative artificial intelligence (‘AI’) absorbs the entire human textual corpus, it will soon be possible for inquirers to discover connections between various figures, findings and ideas in terms suitable for their own purposes, without having to go through the prescribed epistemic channels of academic referencing conventions. Indeed, it will be easier than ever for readers to make connections that go against the grain of an academic author's intended meaning, rendering the academic text closer to a prose poem than an instruction manual. I advise that this article be read in that spirit. The implications of this shift in literary conventions for the organisation of society are considerable, but they will not be discussed here. In any case, Jorge Luis Borges was a harbinger of things to come.

Academic referencing conventions have served several sociologically significant functions, not least the display of disciplinary competence and the crediting of relevant authorities, which unfortunately remain more closely correlated than one might wish. From the choices that academic authors make about what to write about (and how to do so) to the peer review process whereby articles are selected for publication in academic journals, there is an enormous amount of what the social psychologist Kurt Lewin first called ‘gatekeeping’. At one level, ‘gatekeeping’ is simply a euphemism for collective self-censorship in the name of ‘quality control’. But at another level, it is an exercise in boundary maintenance, what I have called ‘academic rentiership’, with an eye to the material basis of the ‘gatekeeping’ metaphor. It entails treating knowledge as property whose value is increased by artificially maintained scarcity (through expert enclosure), which ends up inhibiting its natural productivity (through public use). Interdisciplinarity should abandon this entire way of thinking now, before generative AI makes its abandonment inevitable.

Whatever aura continues to surround academic knowledge production rests on academics concentrating their attention and efforts on a relatively small fraction of the texts produced by their colleagues. That elite gaze establishes a ‘research frontier’, which implies a ‘research trajectory’ that continues to push back the frontier, which in turn creates an illusion of ‘progress’ to spectators. Those ‘spectators’ include academically trained democratic policymakers who regard ‘science’ (in the broad sense of Wissenschaft) as their natural ally in trying to assert a sense of direction on the populace. In any case, it means that most of what academics write probably goes unread – and certainly goes uncited. Consider it as a genteel version of the Orwellian slogan, ‘Ignorance is strength’. It is ‘genteel’ insofar as many if not most academics whose works are never properly read can nevertheless leverage them to keep their current academic posts and perhaps even get promoted to higher posts. This social fact mitigates the brutality – not to mention the absurdity – of not being read after having been trained to write and encouraged to publish in the academic style.

All models of interdisciplinarity presume the existence of well-bounded disciplines that should be combined to produce an epistemic whole greater than the sum of its parts. In that sense, it is not surprising that this journal was founded by a chemist. But at the meta-level, it is equally not surprising because chemistry epitomises what Werner Heisenberg called a ‘closed theory’, that is, a form of knowledge whose stable scope of application renders it sufficiently mature to be widely used in society. Such theories are themselves like atoms that can be combined with others of their kind to produce new forms of knowledge, typically of a more practical kind.

Indeed, when Heisenberg protégé Carl von Weizsäcker co-directed (with Jürgen Habermas) the Max Planck Institute for the Technoscientific Lifeworld in the 1970s, a group of left-leaning philosophers and sociologists dubbed the ‘Finalizationists’ proposed that state science policy should focus on identifying such closed theories and redirecting their efforts from pure to applied research. The proposal was received as controversial at the time because it seemed to amount to some Soviet-style arrest of free inquiry. However, two decades later, once the Cold War ended, the US political scientist Donald Stokes brought the strategy full circle by demonstrating the fecundity of applied research for pure research, in what he dubbed ‘Pasteur's Quadrant’. In effect, Heisenberg had planted a certain epistemic prejudice into the coinage of ‘closed theory’. While Newtonian mechanics can serve as the foundations for nearly all engineering, in the end it remains contained within the trans-Newtonian dimensions of physical reality exposed by Heisenberg, Einstein and others in the early twentieth century. Here, an ‘open theory’ attitude persists, as the scientific frontier continues to be pushed back – though Heisenberg himself believed that the predictive reliability of quantum mechanics also rendered it a closed theory.

Thomas Kuhn, who interviewed Heisenberg for the US National Science Foundation's oral history of quantum mechanics project in the early 1960s, adapted Heisenberg's closed/open theory distinction to restrict the scope of the history of science to closed theories. It amounted to a division of epistemic labour between science and history. And while a history cannot be written of a science with an open frontier (because its full potential has not yet been realised), that science is bound to contain parts with closed frontiers, whose histories can be written. Thus, the latest scientific episode of which Kuhn himself wrote a history was Max Planck's 1900 explanation of the light emitted as a body heats up (i.e. ‘black-body radiation’), the basis for what is now called ‘Planck's Law’. This is because (so Kuhn argued) Planck did it entirely within the paradigmatic ‘Neo-Newtonian’ understanding of light, which had moved from Newton's own particle-based conception to a wave-based one, but otherwise remained committed to Newton's theoretical and methodological horizons. Kuhn's punchline was that Planck originally postulated quantum discontinuity merely as a mathematical fiction to account for the empirical findings. It took Einstein and the quantum revolutionaries a few years later to convert Planck to the outright reality of quantum discontinuity.

The proposition suggested by Planck's conversion – that mathematics can serve as an ontological generator to expand the scope of reality – will be explored below. Kuhn himself never accepted any overarching sense of progress in science, be it defined in epistemic or ontic terms. On the contrary, he was attracted to Darwinian evolution as a model for the history of science not merely for its clearly defined species lineages but especially for its clarity in defining which species are extant and extinct, based on their reproductive capacity. The former belong to science (biology) and the latter to history (palaeontology). For Kuhn, extant and extinct species exist in incommensurable ‘separate but equal’ realms, each entailing its own kind of ‘relativism’, the one represented by a dynamic ecological niche the other by a static fossil deposit, often in the same physical location, the one on top of the other. That is certainly one very vivid way to imagine the distinction between ‘open’ and ‘closed’ theories, and it helps to explain Kuhn's (again Planck-inspired) belief that paradigm change corresponds to generational change, insofar as a paradigm dies once the successor generation fails to reproduce it in their own scientific practice. But more importantly for our purposes, Kuhn's view also justifies the vision of disciplinary compartmentalisation as endemic to what might be called a ‘natural history of scientific knowledge production’.

Kuhn's philosophical nemesis, Karl Popper, approached matters quite differently, embracing Heisenberg's ‘open theory’ horizon as a general epistemological and ontological perspective applicable to both science and its history, which he regarded as inseparable – as did his followers Imre Lakatos and Paul Feyerabend. In any case, generative AI promises to break down the Kuhnian barrier between science and its history, as all academic publications are made available without the usual biases from disciplinary training and citation counts that end up channelling users’ views about which works are on the cutting edge, the mainstream or irrelevant if not obsolete. (To be sure, programmer- and user-based biases may be introduced, but these are likely to have a non-academic character.) This prospect for textual renewal is comparable to the resurrection of the dead in American Mormonism and Russian Cosmism. Ideas and perspectives that had been previously consigned to the dustbin of history may suddenly acquire a new life, fortified by latter-day empirical insights that now render them viable. Such a prospect would certainly provide a normalising context for the re-emergence of seemingly defunct scientific ideas – including creationism, holistic medicine and parapsychology – in today's ‘post-truth condition’. A specifically interdisciplinary version of this development, ‘undiscovered public knowledge’, will be raised at the end of this article.

Returning the Interdisciplinary Imagination from Chemistry to Mathematics

Interdisciplinary Science Reviews was founded by a cosmopolitan chemist, Anthony Michaelis, who in his 1976 inaugural editorial explicitly drew on the chemical distinction between compound and mixture to contrast inter- and multi-disciplinary research, the former producing a new object of knowledge that is not reducible to its disciplinary parts. This way of thinking about interdisciplinarity would have resonated at the time with certain left-leaning scientists who fancied Friedrich Engels’ recasting of Hegel's logic in Dialectics of Nature as chemistry performed on the world-historic stage, which certainly captured Hegel's fascination with the ‘catalytic’ moment. But for what follows, it is worth pondering the largely behind-the-scenes role that chemists in the past hundred years have played in shaping both the professional and popular understanding of science as whole, which enabled Michaelis to reach so easily for the compound metaphor to characterise interdisciplinarity. Wilhelm Ostwald, George Sarton and Eugene Garfield come to mind as relevant chemists. Ostwald introduced the scientific abstract and keywords to ease access to the content of journal articles, Sarton developed the archival orientation towards the history of science discipline, and Garfield designed the Science Citation Index, the precursor of today's Web of Science.

Taken together, the efforts by these three trained chemists contributed to the creation of an organisational structure for navigating the ever-expanding space of scientific (i.e. wissensschaftlich) knowledge. That chemists were so integral to this achievement may have reflected chemistry's disciplinary status by the early twentieth century as a Heisenberg-style ‘closed theory’, due to its fundamental principles having been discovered and subsumed under the principles of another science – physics – whose own epistemic frontiers were kept ‘open’ just as long as its fundamental principles remained unresolved. Here it is worth recalling that for the 150 years prior to Einstein's explanation of Brownian motion in 1905, the difference between ‘physicists’ and ‘chemists’ mainly turned on the existence of atoms, which the former accepted despite their unobservability and the latter denied in favour of concepts such as energy that were closer to the experience of the scientist.

The battleground for this debate in the nineteenth century was thermodynamics, in which the direction of travel turned decisively in the atomists’ favour with Ludwig Boltzmann's statistical account of Newtonian mechanics, including thermodynamics. He provided the mathematical horizon in which Einstein's explanation for Brownian motion could seem plausible. As it turns out, in the twentieth century, thermodynamics was turned towards the interdisciplinary study of life and mind based on local energy flow reversals, which Norbert Wiener christened ‘cybernetics’. I shall explore this connection in the final section, but for now what matters is that once the chemists lost the battle over atoms, they embraced a ‘closed theory’ understanding of not only their own science but all science – perhaps in the spirit of Freud's ‘return of the repressed’. This is most evident in Sarton's archival focus, which suggests that it is never too early to organise and generate the materials needed to do proper autopsies and obituaries of sciences currently operating as open theories because they too will become closed theories; hence Kuhn's involvement in the oral history of quantum mechanics project. In any case, the three previously mentioned chemist-based metascientific innovations have contributed to contemporary science's distinct character as a form of knowledge.

Thomas Kuhn captured the distinctness well in his idea of ‘normal science’, which occurs within a ‘paradigm’, whose main activity is ‘puzzle-solving’, a term meant to convey a highly constrained problem-space. This is the mentality that regards academic disciplines as ‘fields’ and ‘domains’, whose regions can be ‘mapped’ and then divided and conquered through the methodical treatment of well-defined problems. This way of thinking about knowledge began to acquire a sociological form with the reconstitution of the university as a bureaucracy, starting in nineteenth-century Germany. It is a history that casts a somewhat different light on what Robert Frodeman rightly describes as the mathesis universalis approach to knowledge, which projects an indefinitely expansive knowledge space for exploration. It became part of the modern Western intellectual firmament, courtesy of the philosophers whom historians of philosophy call ‘Rationalists’: Descartes, Spinoza and Leibniz. Whereas Robert Frodeman in his contribution to this issue focuses his fire on mathematics as a method supposedly applicable to everything, wherever and forever, I find this very consequential feature of mathesis universalis much less problematic than the legacy of the specific version that captivated early modern thinkers, namely, that knowledge could be plotted on an indefinitely expanded version of Euclidean geometry.

For more than a half-century now, bibliometrics and scientometrics have provided useful network analysis graphs that depict the distribution, concentration and overall relationship of various fields of inquiry, based on researcher activity. Here, conceptual relations are represented as spatial ones on a Euclidean plane – or sometimes a cube. Eugene Garfield even aspired to an ‘Atlas of Science’ that could instruct science policymakers on which fields of research were over- and under-subscribed, as well as ‘gaps’ of the sort that Kuhn identified with puzzle-solving but now projected on a cross-disciplinary canvas. To be sure, Garfield never specified whether this ‘Atlas’ would envisage knowledge as covering a self-enclosed globe or spreading over the sort of infinite Euclidean space that Newton envisaged. Nevertheless, philosophy has not been immune to the charms of the Euclidean vision of mathesis universalis. Quine's ‘web of belief’ metaphor suggested something similar, as if propositions exist at various logical distances from each other in a conceptual world modelled on physical space. Indeed, he envisaged this ‘web’ as having a core and a periphery, in which empirical challenges first strike the periphery and are normally contained there in what Quine himself admitted as a ‘conservative’ epistemic strategy of belief revision.

With greater modal sophistication but in largely the same spirit, Wilfrid Sellars invoked ‘the space of reasons’, in which we ‘make moves’, chess-style, to justify our actions. Robert Brandom has been the most diligent disciple of this approach, which seems to have impressed Jürgen Habermas, even though it amounts to our living double lives, whereby the ‘moves’ that we make in this conceptual world are orthogonal to those we make in the physical world. And while philosophers typically regard Cartesian dualism as a containable problem within the philosophy of mind, it is really a ticking timebomb beneath all contemporary knowledge production. To be sure, much of its potential explosiveness has been mitigated by the internet, whereby physical space collapses into conceptual space. Thus, as research collaboration no longer requires personal visitation, the very reality of research shifts from the physical location of researchers to the intensity of their online interaction, which scientometric graphs can capture very well. In that sense, the map is indeed becoming the territory.

Jean Piaget's account of cognitive development offers an interesting aetiology for this spatialisation of the conceptual. It marks the transition from what Piaget called ‘preoperational’ to ‘concrete operational’ thinking (starting at roughly age 7). In the former, children treat neighbouring objects as needing to be together, whereas in the latter children can regroup the objects according to characteristics they find common in them, resulting in a reorganisation of physical space to match conceptual space. In the final stage of Piaget's account, ‘formal operations’, this process happens in the head, such that one's vision of physical space comes to be strongly mediated by the conceptual lens, resulting in, for example, a distrust in the causal significance of perceived physical contiguity. A verbal indicator of this treatment of physical space as conceptually rearrangeable is the word ‘extension’, which, on the one hand, Descartes used to describe a bounded physical space – the functional equivalent of a material object – and, on the other hand, modern logic uses to refer to a collection of actual or possible objects sharing common properties – the functional equivalent of a concept.

Piaget had some provocative things to say about this conversion of physical to conceptual space, based on his acceptance of the ‘ontogeny recapitulates phylogeny’ thesis advanced by the late nineteenth-century evolutionist Ernst Haeckel. For Piaget it meant that the cognitive development of children re-enacts the history of science, with the Newtonian worldview – slightly modified by Einstein – corresponding to the final stage of formal operations. In the 1970s, Lawrence Kohlberg extended this approach to moral development, with Kant as modified by Rawls serving as the developmental endpoint. In this context, agape, a sense of love that treats strangers the same as friends, represents formal operations. One might go further and project this mentality on the geopolitical stage in the manner of David Ricardo, who held that in a universal free trade regime, each nation would play to its ‘comparative advantage’ vis-à-vis productive efficiency, which entails, in the case of comparable goods, the removal of any protection that native producers have over foreign ones. This would constitute formal operations at the level of political economy. From this standpoint, Aristotle's worldview appears to be stranded between the preoperational and concrete operational stages because he presumed what nowadays we would call an ‘ecological’ perspective, whereby the sustained spatial proximity between entities is ipso facto taken to be indicative of a deeper conceptual relationship, which can be characterised as one of functional interdependency. Thus, Aristotle's sense of ‘economy’ was centred on sustaining the family household, not facilitating international exchange. Put bluntly, Aristotle never thought about people with whom he was not in regular physical contact.

So, what exactly is wrong with the mathesis universalis treatment of conceptual space as bounded like Euclidean physical space? It suggests that conceptual space is closed under its own set of laws, which a research programme or paradigm then ‘disciplines’ with its methods of empirical inference and problem-solving. Piaget's final stage of formal operations is the epitome of this vision. Yet, Heisenberg's invocation of ‘closed theories’ to capture this vision of organised inquiry equally presumes ‘open theories’ whose frontiers are indeterminate – where something different happens. Interestingly, for reasons we shall touch on in the final section, Heisenberg himself did not regard quantum mechanics as ‘open’, notwithstanding the opinion of his fellow physicists, both then and now. Most would judge both quantum and relativity theories as still active research programmes in physics whose open horizons depend on suspending one or more of the fundamental assumptions of the now closed Newtonian mechanics (including its dependency on Euclidean geometry). In the early twentieth century, this openness served to solve long-standing anomalies in the Newtonian paradigm, but increasingly these theories have extended the explanatory range of physics to data domains that Newton himself never imagined. And while physics was the science that most directly benefitted from this open-theoretic horizon, the strategy itself was largely a product of nineteenth-century mathematicians inspired by Carl Friedrich Gauss.

Contrary to such Enlightenment mathematicians as D’Alembert, who laboured under a broadly materialist metaphysics, Gauss treated his discipline not simply as a methodological handmaiden to the empirical sciences but as an independent field of inquiry for discovering new domains of objects. Moreover, the motions of these objects may violate the ‘mechanical’ version of materialism associated with Newtonianism. These ontological innovations started to reconfigure the other sciences in the early twentieth century, starting with physics (non-Euclidean geometry) and logic (set theory). Indeed, the ‘open-theoretic’ character of the history of post-Enlightenment mathematics inspired Karl Popper to propose a ‘World Three’ of ‘objective knowledge’, which contains all such conversions of tools to objects of inquiry. In terms of philosophical anthropology, the emergence of World Three marks the shift from developing a tool simply in terms of its anticipated uses to extending the tool's various functional dimensions and then projecting worlds in which such extended tools would be useful. This systematic means-ends reversal effectively primes tool-users to imagine unexplored conceptual space as much larger and more varied than that which led them to use the tool in the first place. It is this second-order intellectual revolution that I associate with Gauss in mathematics.

Gauss himself is credited with the development and application of two speculative fields of mathematics, complex numbers and differential geometry, to solve standing problems in physics and astronomy. He was not deterred by the counter-intuitiveness of the objects generated by these fields. On the contrary, he used the situation as an opportunity to expand mathematical intuition's purchase on reality, a strategy that Einstein later adopted when treating Bernhard Riemann's version of differential geometry as the platform for addressing long-standing problems about the motion of light, resulting in the relativity revolution in physics. In that sense, modern mathematics follows in an especially rigorous way the path of those who enter politics by way of utopias and those who enter science and technology by way of science fiction. (The lazy Kantian term ‘a priori’ doesn’t do it justice.) In each case, the fictional is leveraged into the factual by expanding the sphere of the realisable. Conversely, it helps to explain the early twentieth-century revanchist move by L.E.J. Brouwer (as well as Husserl and the later Wittgenstein) to disown these intellectual innovations and return mathematical intuition to the sort of human-level visual experiences that originally produced arithmetic and geometry in ancient Greece.

After Michael Dummett, it has been common to characterise this attempt to re-anchor intuition in the vision of the physical eye instead of the mind's eye as ‘anti-realism’. The difference in optical horizon is important, since the anti-realist stress on ‘verification’ and ‘proof’ turns on the ability to inspect each step of the path from premise to conclusion. Its strategy is to reduce the conceptual to the empirical. It is not ‘intuition’ in the holistic sense of glimpsing the world all at once, as God might do. On the contrary, it is about imposing the physical eye's frame of reference on the mind's eye, which historically amounts to a reversal of fate, since the epistemic fixation on the eye was originally about light's alleged ability to open the mind's eye, back when the modern distinction between experience and imagination had yet to be enforced. However, this reversion to the literal eye, which becomes especially pronounced after Kant, understands our physical vision not as an imperfect participation in – or reflection of – the divine mind but as an expression of the intrinsically limited nature of the human mind. And while it is conceptually easy to see this shift in perspective as simply two different ways of talking about the same thing, it provides an initial clue to why this post-Kantian view of intuition is characterised as ‘anti-realism’, notwithstanding what the relevant ‘anti-realist’ philosophers thought they were doing.

Consider the most important downstream thinker in this ‘anti-realist’ trend, Martin Heidegger. He saw his task as trying to recover ‘the ground of being’, an ultimate sense of reality out of which our linguistic and mathematical constructions emerge, which are themselves only partial if not distorted representations. In that sense, which Heidegger himself coined, philosophy was about ‘deconstructing’ such artifices of representation. Robert Frodeman has similarly promoted what he calls ‘field philosophy’, which would take the idea of a ‘reflective practitioner’ to an ontologically deeper level, whereby the inquirer's activity would become better aligned with that of the object of inquiry. The result would be a kind of ‘ecologically valid’ epistemology, which Frodeman himself calls ‘sustainable’. Most philosophers, not only Dummett, are inclined to call this ‘anti-realism’ because ‘realism’ nowadays refers to the idea that objects – typically material – have a determinate identity independent of any observer. It is based on the historiographical convention that Aristotle reversed Plato's perspective on the origin of objects: Whereas Plato held that an object is just one of many possible (better or worse) realisations of an idea, Aristotle held that an idea is real only to the extent that it is realised in objects.

This complementarity of a materialist Aristotle to an idealist Plato has fostered an understanding of reality as reducible to the materialisation of one possible world, which constitutes what after Aquinas is called ‘actuality’, tipping the balance clearly in favour of Aristotle. From that standpoint, Heidegger's philosophy is radical because it positions the human not as a possessor of Plato's ideational creativity that can conceive if not construct alternatve worlds, but rather as itself no more than one of the ‘thrown’ (geworfen) objects in the actual world. Moreover, Heidegger purports to articulate ‘what it is like to be an object’ without presuming to know the corresponding subject – or even if there is one. Here he flips the import of Kantian epistemology: Heidegger reads Kant's claim that we only know our representations of reality as implying that we only know our alienation from reality, as expressed by those representations. But for our purposes, Heidegger's key message is that any attempts to extricate ourselves from our alien predicament – such as through the artifices of mathematics and technology more generally – need to remain grounded in our thrown nature to be true to our ‘being in the world’.

Plato would certainly not have accepted such a restricted understanding of the human condition – and neither did Dummett (nor do I). Dummett's own qualified support for the position he called ‘anti-realism’ was a perhaps too subtle attempt to recover philosophy's Platonic/theistic starting point, whereby the ‘anti’ in ‘anti-realism’ was meant to signal the creative or voluntary (what positivists self-deprecate as ‘conventional’ or ‘arbitrary’) element in the determination of reality, which humans and God share in the Abrahamic theological traditions. From that standpoint, the commonsense understanding of ‘reality’ inherited from Aristotle and Aquinas is merely a biased part of Plato's larger sense of reality as all that is possible, which in recent times Nick Bostrom has characterised as an ‘observer selection effect’. Once we regain the Platonic vision, what the physical eye sees turns out to be no more than a snapshot of all that the mind's eye sees, albeit more dimly when it is not in focus. The great Cartesian Nicolas Malebranche called it our ‘vision in God’.

Thus, Dummett's use of ‘anti-realism’ should be understood somewhat ironically. Notwithstanding Heidegger's clever perspectival reversal of the Christian ‘fallen’ to the atheistic ‘thrown’, his concern was ultimately with the ‘so-called reality’ of someone who takes the Blue Pill in the film The Matrix and imagines that their world is the only possible one, perhaps for reasons they will never fully understand. This also applies to the latter-day ‘intuitionists’, with whom Dummett seems nominally aligned. However, Dummett himself was a Red Pill taker, whose interest in intuitionist logic extended to the full range of many-valued logics, including those covering quantum-mechanical settings. He clearly wanted to scope out all the worlds that could become our own – not simply the actual one that Heidegger thought we are stuck with. This openness to realities beyond Aristotle's one to Plato's many (note the reversal of imagery from Raphael's School of Athens) prepares us for a world in which the search for knowledge is not academically path dependent – a world to which generative AI has begun to introduce us.

The Future of Interdisciplinarity as Undiscovered Public Knowledge

Earlier, I referred to the key role that thermodynamics played both in turning chemistry into a Heisenberg-style ‘closed theory’ and providing the basis for arguably the greatest interdisciplinary project of the second half of the twentieth century, cybernetics. The connection was provided by Leon Brillouin, a founder of solid-state physics. It is perhaps unsurprising that this field inspired Brillouin to turn his attention to unifying information and energy under one set of mathematical principles, which could reasonably aspire to constitute a unified science of life and mind. After all, solid state physics studies how microscopic phenomena, including those at the quantum level, interact to produce stable macroscopic structures. Of course, physical reality as we normally perceive it has this ‘solid state’ character, but the field becomes interesting once we start thinking in terms of how microscopic phenomena might be harnessed to produce novel solid-state structures. The electronics revolution in the second half of the twentieth century resulted from this line of thought, which in recent years has opened the door to more adventurous projects such as quantum computing and even biocomputing, which ultimately envisages the integration of neural wetware and silicon-based hardware. An interesting feature of these developments is their reliable use of the fundamental equations of quantum mechanics without taking a stance on the ontology behind the mathematics. It was for this reason that Heisenberg himself regarded quantum mechanics, somewhat surprisingly, as a ‘closed theory’.

Heisenberg's judgement presumed the irreducibly observer-dependent nature of quantum-level phenomena. He concurred with his senior colleague Niels Bohr that fundamental ontology was off the table. But that didn’t prevent quantum mechanics from specifying higher-order microscopic phenomena that escape this radical indeterminacy by maintaining their functional stability in the face of various interventions. What in the twenty-first century has been called ‘nanoscience’ is about identifying the smallest functionally stable material bodies and testing the limits of productively intervening in their operation. The canonical source for this repurposing of quantum mechanics – seeing it from its flipside, as it were – is Erwin Schrödinger's 1943 Dublin public lectures on the topic, ‘What Is Life?’ Schrödinger focused his speculations on the physics underwriting the smallest units of life: the cell and the gene. Famously, he proposed that genetic material can be understood as an ‘aperiodic crystal’, which inspired a young audience member, Francis Crick, to discover what turned out to be the double helix structure of DNA.

Schrödinger believed that behind his speculations lay a missing law of physics, which Brillouin, borrowing a term of Schrödinger's, called ‘negative entropy’. If order tends towards disorder in a closed system through the degradation of the energy used to maintain the system, then the prospect of increasing order requires that the system receives sufficient outside energy to use productively. Here the sun, understood as a very large and nearby hot body, is posited as the stopgap source of the requisite ‘free energy’. Brillouin likened this entropy-reversing environment to memory storage of those structural features that sustain system functionality. In this sense, new energy is equivalent to new information, whereby the reversal of entropy amounts to regular reminders for the system to continue being itself, and maybe even to evolve, if there is enough free energy to permit expansive system-level reorganisation. Norbert Wiener proposed mathematical equations purporting to show that, at least in principle, the behaviour of humans, animals and machines can all be understood this way, namely, as ‘cybernetic systems’. For our purposes, it is worth recalling the subtle but pervasive influence that this entropy-counteracting cybernetic vision has had on our understanding of discipline formation and interdisciplinarity.

One clear example is what I earlier described as Kuhn's ‘natural history of scientific knowledge production’, which begins with the consolidation of a paradigm that establishes an epistemic monopoly over a field of inquiry by parsing problems in the field into its own theoretical terms and pursuing them by its own authorised methods. However, that modus operandi over time brings diminishing returns on investment, eventually resulting in the paradigm causing more problems than it solves. This tendency corresponds sociologically to the degradation of the commitment of later generations to carry on with the paradigm. At that point, the paradigm enters ‘crisis’ mode, soon followed by a ‘revolution’ from which a new paradigm emerges. Kuhn interestingly characterised the revolutionary moment as one in which scientists are briefly open to non-scientific considerations, ranging from theology and philosophy to art and politics, only to be followed by a reorganisation of their self-understanding to make it appear that the new paradigm was where the science had been heading all along. Thus, the next generation of scientific recruits is led to believe that, say, Newton and perhaps even Aristotle would recognise Einstein and Heisenberg as having made significant advances on their original inquiries.

Kuhn quite rightly dubbed such historical revisionism ‘Orwellian’. At the same time, the epithet reveals that Kuhn thought about paradigm change from the standpoint of the historian of science rather than the scientific practitioner. ‘Orwellian’ stresses what is left behind and forgotten, if not actively suppressed, rather than what is enabled in a way that might have previously seemed impossible. In recent work, I have discussed this tension as a struggle for modal power, which I take to be at the heart of our ‘post-truth’ condition. It reflects the gradual abandonment of an assumption that Kuhn's periodically disruptive, paradigm-driven account of science continues to share with the opposing ‘Whig’ historical account favoured by the scientific establishment, which projects a tale of linear progress. Both sides agree that the history of science is ‘temporally asymmetrical’, in the sense that the research frontier moves forward by leaving the past behind. The difference is that Kuhn believed that science's past is sufficiently interesting to warrant its own positive epistemic space, even if it will never pose a challenge to the paradigm of the day. Indeed, Kuhn's ‘separate but equal’ policy to science and its history has inspired historians of science to explore literary (aka ‘postmodern’) modes of analysis that suspend any strong epistemic commitments capable of troubling the practising researcher. In short, the liability that results from losing the status of fact over time is converted to the virtue of being forever available to the tools of fiction.

But what if the past is never truly left behind? As I earlier observed, Kuhn's general historical vision was influenced by Darwin's theory of evolution's sharp distinction between extant and extinct species, the latter belonging to palaeontology, a field renowned for its literary flair, not least in Darwin's own writings. So far, so Kuhnian. However, in recent times, the Harvard medical geneticist George Church and others have proposed a program of ‘de-extinction’ and ‘regenesis’, which re-envisions the fossil record and related palaeontological evidence as sites for DNA extraction that might then be used to resurrect extinct species. Perhaps the most controversial proposal in this vein was Church's published call for women to carry to term a baby containing Neanderthal DNA. This formerly science-fictional prospect became possible once genetic material could be successfully treated as a code that is executable on a range of carbon-based platforms that might cross species boundaries. The sense of ‘impossibility’ suddenly migrated from the ontological to the ethical realm – from what we cannot do to what we ought not do. All this would be familiar ground to Marshall McLuhan, who asserted that old media provide the content for new media. Consider the old radio and television broadcasts already available on YouTube – and, more to the point, the long-term digitisation of print media, access to which is increasingly customised through generative AI. It is on this point I will close.

Forty years ago, Don Swanson, a Berkeley-trained physicist who shifted to the library science faculty at the University of Chicago, managed to discover a hypothesis that provided a cure for Raynaud's syndrome (fingertip numbness) by using computer-based technologies to combine literatures of different disciplines not normally read together by professional academics. Swanson's search protocols on these admittedly primitive systems easily overrode academic reading biases. The result was surprising only because machine learning defied the expectations of its users – and perhaps even programmers – with novel outputs. But it was not surprising from a scientometric standpoint, given the long-standing finding that most academic publications are systematically ignored because they don’t fit easily into the research narrative of a science's dominant paradigm. One can only hope that generative AI, with its much greater data and computational capacity, is allowed to evolve into an engine for alternative ways of thinking and being, including those previously ignored due to various academically institutionalised habits. Swanson originally called this unexplored realm ‘undiscovered public knowledge’, explicitly with an eye to Popper's World Three. It might even open the door – as the history of modern mathematics has demonstrated – to new ontologies to accommodate the new combinations of information that are increasingly being made available to more people. In this respect, it shouldn’t be surprising that Schrödinger opposed Heisenberg's idea that quantum mechanics is a ‘closed theory’. Along with every other currently closed theory, it may soon be open for business again.