This is an unedited preprint of a review article that was publised in 1999 in Theoria number 93, pages 29-52. This version does not have the footnotes that appeared in the published version - DS.

Lyotard and the Postmodern Misunderstanding of Physics

David Spurrett

Department of Philosophy, University of Natal, Durban, 4041, South Africa.

email: <>

Do I contradict myself?

Very well then .... I contradict myself;

I am large .... I contain multitudes.

- Walt Whitman

In a famous passage in his Philosophical Investigations Wittgenstein compares a language to an ancient city, saying that we can see it as ‘a maze of little streets and squares, of old and new houses, and of houses with additions from various periods; and this surrounded by a multitude of new boroughs with straight regular streets and uniform houses’(1958:paragraph 18). Descartes exploited a similar analogy in his Discourse on the Method, drawn in his case between a city and a system of knowledge. His position, though, was strikingly different. Where Wittgenstein describes, he prescribes, stating, first, that ‘there is not usually so much perfection in works composed of several parts and produced by various different craftsmen as in the works of one man’ and going on to argue that the proper task of philosophy is to show us how, individually, we can ‘get rid of’ the opinions which form our existing epistemological landscape, ‘in order to replace them afterwards with better ones, or with the same ones once ... squared ... with the standards of reason’ (1985:Discourse 1).

These two visions are not necessarily incompatible. A Cartesian could accept the Wittgensteinian image, as indeed it is clear that Descartes did, as long as it is taken simply as a description of the status quo. Deep philosophical differences arise here, rather, in the face of the questions whether the bulldozing and rebuilding approach is desirable or even possible. Descartes, of course, thought that it was both possible and morally imperative, since not to reconstruct would be to make improper use of our God-given capacities to reason and to discriminate between truth and falsehood. On his view the disorganised or heterogeneous city of knowledge is a place filled with error, error in turn being a particular kind of sin.

While rationalism itself is no longer popular, the main outlines of Descartes’ programme to set the epistemological city to order live on. As well as giving a central place to the ideal of certainty as the goal of epistemology, and proposing analysis into simple ideas which stand in deductive relationships as its method, Descartes set in motion a detailed working epistemology for the practice of the natural sciences. Briefly, his project was simultaneously to make physics mathematical and mathematics physical. In this he set himself directly against the Aristotelian view that the two subjects are utterly distinct, physics being the study of things which change and exist independently, while mathematics concerned things which neither change nor exist independently. Since Aristotelian explanations are syllogisms starting from first principles relating to essences, the different essences appropriate to his physics and mathematics meant that in scholastic thought the two subjects were irreconcilably divorced. Descartes helped shape one revolution which reconceived mathematics as the pure study of form and proportion, and another which saw the objects of physics as essentially formal and proportional, in his case via the identification of matter with extension. These two transformations made modern mathematical physics possible.

Descartes’ own physics was, though, a picturesque failure and within a century it was irreversibly superseded by a Newtonian conception of how empirical questions could be made mathematical. Where Descartes talked vaguely and unworkably of vortices in a universe completely filled with matter, Newton formulated precise and quantitative statements describing isolated bodies of matter in motion, decisively closing Aristotle’s ancient divide between physics and mathematics. He followed Descartes in helping end a long period in which physics had been founded upon a set of deep heterogeneities between different kinds of matter and different kinds of motion (the distinctions between natural and violent movement, the sub-lunar and super-lunar spheres, corrupt and aetherial substance) these again being largely the heritage of Aristotle. Newton’s first law of motion, thus, states simply: ‘That every body continues in its state of rest, or of uniform motion in a right line, unless it is compelled to change that state by forces impressed upon it’ (1934:Law 1). The revolution here is largely encapsulated in the words ‘every body’ - indicating a uniform world with Newton going on to specify a universal property of that world. With reference to Newton’s great triumph regarding gravity the historian of science Brooke has pointed out that ‘[t]here was now a universal law of gravitation. Because it applied to all bodies everywhere the universe had at last become a universe’ (1991:53). It is worth remembering here that no less an astronomer than Kepler had endorsed a kind of heterogeneous form of attraction which acted in such a way that fragments of the Earth would ‘fall’ towards the Earth, fragments of Mars toward Mars, and so forth. With Newton, in contrast, we see the modern notion of the cosmos as a universe: a single space governed by a unique set of uniformly operating laws. With that we are also given the notion that those laws are the proper object of study of a natural philosophy based on mathematical principles, that is to say, in the modern sense, of physics.

Classical modern physics takes the form of an ambitious project, untroubled by humility or reservation. What it promises is simply that all questions relating to motion or other types of material change, and hence in a sense all questions about reality, may be settled by finding out the relevant mathematical laws and the values of the appropriate quantities. In its mature form its grail is the image of Laplace’s vast intellect which, famously, if given ‘all the forces by which nature is animated and the respective situation of the beings who compose it,’ could ‘embrace in the same formula the movements of the greatest bodies of the universe and those of the lightest atom; for it nothing could be uncertain; and the future as the past would be present to its eyes’ (1951:4).

This challenge is one that the supporters of a conception of language or knowledge as heterogeneous need seriously to consider. It is simply insufficient for a would be citizen of the unruly later Wittgensteinian city to say, for example, that physics on the Laplacian model can be left to its own devices, but that any of a number of alternative and perhaps mutually incommensurable discourses about the world will also be possible or available. Such discourses will be doomed to have a second-class status compared to physics, which will always, so to speak, be holding the epistemological trumps. One might of course choose to ignore physics, or wilfully to proceed with one’s intellectual life using some other approach to knowledge, but in doing so one remains open to the Cartesian objection that this is simply symptomatic of unfortunate but indefensible methodological laziness.

It is for this reason that, for example, Rorty’s approach to physics in Philosophy and the Mirror of Nature is so deeply flawed. Rorty is quite willing to say that ‘[p]hysicalism is probably right in saying that we shall some day be able, ‘in principle,’ to predict every movement of a person’s body (including those of his larynx and his writing hand) by reference to microstructures within his body.’ Untroubled, though, Rorty maintains that the practical difficulty of performing actual demonstrations of this would mean that except for the ‘occasional pedagogical exercise’ we would in fact remain ignorant of the causal details, and hence that the ‘danger to human freedom of such success is minimal’ (1980:354). Thus liberated by the poverty of our data we can supposedly pursue our hermeneutical endeavours unimpeded, with the crucial imperative being, always, to keep the conversation going. Bhaskar points out pithily that ‘this is disastrous’ and that on Rorty’s view ‘we would have to reckon a falling man free in virtue of his ignorance of ... gravity’ (1991:50).

More to the point, as long as no restrictions on the capacity of physics with respect to its objects of study are acknowledged, then no matter what other non-physical description of anything that we do or say we can come up with, we will always be open to having that description displaced by the ‘one true’ account in terms of mathematical laws relating physical quantities. Anything in ‘our’ account of the phenomena in question which is equivalent to a proposition in physics will simply be physics in translation or disguise, and anything with no counterpart in the physical description will be at best falsehood, at worst nonsense. (Rorty’s own position is hardly saved by his frequent attempts either to collapse the distinction or invert the priority between scientific and literary discourses, since his pragmatist approval of the effectiveness of science undermines his hermeneutically inspired anti-realism about its content at every turn.)

The reason for this is clear: modern physics began as a project which offered itself as the source of a unified and commensurating description of the world which was epistemologically superior to all others. This project is deeply connected to the image of the rational, Cartesian city of knowledge. It follows that if, for whatever reason, we are loyal to a Wittgensteinian image of the city of language and thought, and are unwilling to consign all discourses which are not part of physics to a second-rate position, we need to take the philosophical battle to physics itself, and with the vision of physics which it is still appropriate to call Cartesian. This is exactly what Lyotard does in The Postmodern Condition, a striking feature of which is the central position of a number of arguments concerning mathematics and the natural sciences, especially physics. Indeed it is at least reasonable to say that section thirteen of The Postmodern Condition, entitled ‘Postmodern Science as the Search for Instabilities’ in which Lyotard argues that the pragmatics of contemporary scientific practice show an appropriately post-modern openness to paralogy and incommensurability, underwrites much of the wider argument of his ‘report’.

I want to argue that while Lyotard is entirely correct to take up the question of the natural sciences in The Postmodern Condition, his arguments are inadequate for their purposes and that in consequence the claims of natural science, especially physics, as a commensurating discourse are undamaged, with potentially disastrous consequences for Lyotard’s programme as champion of heterogeneity.

Lyotard’s work documents what he sees as the collapse of faith in homogeneity, and develops a critical understanding of the ways in which master discourses were and are essentially repressive and totalitarian. Lyotard’s postmodernity can be seen as a creative fusion of aspects of Kant and Wittgenstein in which the sublime is reconceived as a primarily ethical category and a paradigm example for problems of justice, but played out on the terrain of language games and in the field of communication as performance. For Lyotard master-narratives and discourses which force commensurability are essentially totalitarian, their necessarily procrustean ways being paradigmatic breaches of justice and acts of violence (or ‘terror’) against whatever differences between discourses do not admit of translation. The possibility Lyotard identifies and argues for is one where difference is accepted without attempted translation, where we attempt to rethink the dignity of communication and justice without the easy templates and stencils he so abhors. To keep a metaphor used above alive, Lyotard is adamant both that in point of fact we do, and that in the interests of justice we should, play our language games in ‘no trumps’.

The Postmodern Condition was published in French in 1979, and first published in English translation in 1984. The work has taken on something of the status of a manifesto for postmodernity, even in the face of Lyotard’s explicit and emphatic rejection of that reading. Very briefly Lyotard’s position is that the cultural (and also epistemological) postmodern condition is the counterpart of the economic post-industrial condition. Postmodernity, which he conceives of as continuous with modernity in some ways, involves a ‘change in the status of knowledge’ which Lyotard considers with a particular emphasis on the question of legitimation, the processes by which statements are accepted into or excluded from some or other discourse. Within this problem field Lyotard defines the cultural modern condition, itself related to the industrial economy, as a circumstance in which a science ‘legitimates itself with reference to a metadiscourse ... making an explicit appeal to some grand narrative, such as the dialectics of Spirit, the hermeneutics of meaning, the emancipation of the rational or working subject, or the creation of wealth’ (PMC:xxiii). This being the modern, the ‘post’-modern is quite clearly and simply that which comes after the period of appeal to metanarratives, so, famously, Lyotard states that ‘[s]implifying to the extreme, I define postmodern as incredulity toward metanarratives’ (PMC:xxiv).

Lyotard contends that various factors, notably including ‘progress in the sciences,’ (PMC:xxiv) have contributed to the loss of confidence in metanarratives, and proposes that in the period following this loss of faith knowledge will circulate in new ways, being considered more as a product, or even as a form of currency. He states emphatically that ‘[k]nowledge is and will be produced in order to be sold’ (PMC:4) and goes so far as to suggest that: ‘It is not hard to visualise learning circulating along the same lines as money, instead of for its ‘educational’ value or political … importance; the pertinent distinction would no longer be between knowledge and ignorance, but rather … between ‘payment knowledge’ and ‘investment knowledge’ - in other words, between units of knowledge exchanged in a daily maintenance framework … versus funds of knowledge dedicated to optimising the performance of a project’ (PMC:6).

I suggest that there are a number of ways in which these claims are little short of ridiculous. Strong and deep disanalogies between knowledge and currency are simply pushed aside by any proposal to identify the two. No matter how much we come to view knowledge as a product and hence a kind of property, as it is evident the postmodern patent and copyright obsessed world does, we cannot evade the fundamental problem of epistemology: knowledge always has to be identified and justified whereas money can be defined or performed into existence, often by the proverbial ‘promise to pay the bearer on demand’. In any event facts and dollars obey different principles of circulation: the results of showing someone how to do something and showing them my salary are quite distinct, since if I give someone information or a training I do not thereby lose it myself. We simply should not seriously expect future financial updates to quote the latest rates of the ‘Geisteswissendollar’ against the Yen. Lyotard is, of course, not quite that idiotic, and his more detailed and careful remarks on epistemology are discussed below. What is worth noting at this point is the fact that Lyotard’s vision of a pragmatic economy of circulating statements, narratives and money is profoundly Wittgensteinian.

Wittgenstein called for many abstract, a priori or generalising forms of enquiry into meaning to be replaced by careful attention to the ways words are actually used in various language games. In this he attacked the notion that there was some peculiarly philosophical problem about what and how words meant over and above any question about the ways in which words were normally deployed. Similarly, in absence of modern appeals to metanarratives such as ‘the dialectics of Spirit, the hermeneutics of meaning, the emancipation of the rational or working subject, the creation of wealth’ (PMC:xxiii) which might be expected to locate and evaluate a knowledge claim in a larger context, Lyotard has it that the question of the worth of any item of knowledge is to be decided on more local and pragmatic grounds, and goes so far as to describe his approach in The Postmodern Condition as making use of the notion of a language game as a ‘method’ (PMC:27). Where Wittgenstein called upon us to set aside the philosophically inspired notion that language in its entirety should be a unified and consistent whole, and instead take account of its variety and differentiation, Lyotard calls on us to realise that the age of unifying programmes for knowledge has ended and given way to an epistemological landscape which we should recognise as being much like the city Wittgenstein described in the Philosophical Investigations.

Of course, unless it can be shown that the classical modern account of knowledge has been decisively displaced, it is always possible that present and patchy knowledge could plausibly be seen as no more than a half-way house on the way to a more integrated and fundamental account of the type envisioned by Descartes, so that in terms of the city-metaphor Rome could still end up like Milton Keynes. I will return to this question, but need first to note a potentially dangerous ambiguity: Lyotard defines the postmodern in terms of an attitude to metanarratives, but, as the opening discussion of modern science above makes abundantly clear, the challenge of modern science is not so much that it is a dangerously powerful metanarrative as that it represents the rejection of narrative as such. Lyotard partly concedes this point, calling the relationship between narrative and scientific knowledge ‘unequal’ and noting that the scientist ‘questions the validity of narrative statements and concludes that they are never subject to argumentation or proof’ and then ‘classifies them as belonging to a different mentality: savage, primitive, underdeveloped, backward, alienated, composed of opinions, customs, authority, prejudice, ignorance, ideology’ (PMC:27).

Lyotard plausibly relates this scientistic attitude to a peculiarly western form of cultural imperialism. This particular approach towards narratives and hence metanarratives has, of course, the same history as modern physics, and it will be worth looking at its background in more detail to make clear how modern science relates to narratives. In an illuminating passage in The Order of Things Foucault sums up the ‘modifications’ to the ‘episteme of Western culture’ which make the idea of knowledge as representation possible. (It is this conception of knowledge, common to Descartes and Newton, which is largely constitutive of modern mathematical physics.) The final item on Foucault’s list points out that ‘since to know is to discriminate, history and science will become separated from one another’ and that the united world of renaissance learning will divide into, on the one hand, ‘erudition, the perusal of written works, the interplay of their author’s opinions’ and, on the other, science ‘lacking any common unit of measurement’ with erudition, and composed instead of ‘the confident judgements we are able to make by means of intuitions and their serial connection’(1970:54-5).

Foucault’s point here is to contrast the classical approach with its dominant western predecessor, renaissance Neoplatonism. The Neoplatonist approach emphasised the ways in which the world could be considered to be an interrelated and meaningful network of connections in which any deployment of signs or symbols, including narratives, was an important part. While it is commonplace that mercury was used as a treatment for syphilis, the Paracelsean reasoning is not so well known: ‘Syphilis is signed by the market place, where it is caught; the planet Mercury has signed the market place; the metal mercury, which bears the same name, is therefore the cure for syphilis’(Hacking 1975:42). To a modern ear this ‘argument’ sounds like a dangerous tissue of confusions between literal and metaphorical similarities and relationships. Foucault reports that Buffon, faced with Aldrovandi’s extraordinary collection of topics in his work on serpents, including heraldic applications, use in human diet, dreams, mating habits and fables as sub-headings in a single chapter, was moved to exclaim with an eighteenth century frustration we share today ‘Let it be judged after what proportion of natural history is to be found in such a hotch-potch of writing. There is no description here, only legend.’ Foucault’s reply was that for Aldrovandi there was no distinction in the form Buffon saw one, that ‘it was all legenda - things to be read’ (1970:39).

In the face of such evidence we must simply reject Lyotard’s sweeping assertion that science and narrative have ‘always been in conflict’ (PMC:xxiii). Such a position trades on a version of the distinction between literal and metaphorical language which is one of the issues actually at stake in any debate over the relationship between science and narrative. This mis-identification of ‘science’ with modern science is the first of the significant misunderstandings of physics in the argument of The Postmodern Condition, and it stands in the way of a clear understanding of the particularity of modern science. If we eschew making assertions about science and narrative as such, then what becomes painfully clear here is that modern natural science is constituted in terms of a general hostility to narratives. (Indeed it is precisely this kind of scientistic aversion which makes it seem so straightforward to group non-scientific discourses together under the heading of ‘narratives’ at all.) This hostility arises because in the terms of the classical modern episteme narratives are not even a genuine form of knowledge. Given this we should not expect science to compete with metanarratives in the same ways as narratives might compete with one another, and neither should we expect the collapse of faith in metanarratives to have especially significant consequences for science. (Indeed it is surely possible that such a collapse could even favour a part of culture which at least protests its radical independence from narrative.)

Lyotard recognises and partly concedes this point, but has it that science enters the contest of metanarratives when it ‘does not restrict itself to stating useful regularities and seeks the truth’ (PMC:xxiii). When this happens apparently a ‘discourse called philosophy’ comes into being which champions science in the lists of legitimation by generating suitable (meta)narratives containing science.

This, though, is surely an image of philosophy which concedes too much to a restrictive form of positivism. Without abandoning or relaxing the standard modern notion of what it is to know a physical fact, it is possible to envisage a number of different relationships between science and philosophy, and more importantly a number of different approaches to the legitimation of science. So, for example, (meta)narratives could be seen by supporters of science in its modern form as having merely didactic utility, but as being essentially dispensable like the Discourse on the Method, Descartes’ personal Bildungsroman, or the ladder of statements leading to the end of Wittgenstein’s Tractatus Logico-Philosophicus. In this case there is a certain kind of convenience which metanarratives can bring, but in principle the same work could be done by other means. A second possibility is that metanarratives could serve a kind of propaganda function for science, generating positive publicity for those too busy or soft-brained to read current physics journals, or who might object to what scientists were doing if they knew what it was. In such cases the supporters of science might think it, quoting Kant, ‘not only prudent, but permissible, and indeed even commendable, to further the good cause through sophistical arguments’ so as to bring about more quickly the same result as would have been obtained ‘by entirely sincere and honest procedure’ (1929:600). Here too the non-literal function is in principle dispensable, although pragmatically useful nonetheless.

In the face of this, Lyotard’s evidence for his contention that the narrative function is indeed indispensable in the processes by which science legitimates itself is simply inadequate. He proposes, for example, that the philosophy of Descartes can ‘only demonstrate the legitimacy of science through what Valéry called the story of a mind, or else in a Bildungsroman, which is what the Discourse on Method amounts to’ (PMC:29). This is certainly an unacceptable reading of the purpose of Descartes’ Discourse which, even if it does have many of the crucial features of a modern novel, did not set out to ‘legitimate’ science by means of narrative, or even to stand in support of Descartes’ entirely non-narrative works such as the Principles of Philosophy or the Rules for the Direction of the Mind. A clear indication of this is that while Lyotard explains elsewhere in The Postmodern Condition that it is part of narrative pragmatics that anyone who has been told a story is thereby legitimated as a possible (re)teller of the same tale (PMC:20f), Descartes is adamant that to make proper use of his developmental tale it is imperative to think through the arguments for oneself from beginning to end. At most, then, we can allow that the Discourse, deliberately written in French, was intended to play the didactic role described immediately above. Similarly Lyotard’s observation that when scientists appear on television they immediately ‘recount an epic of knowledge’ does not even meet his own standard of being a ‘crude proof’ (PMC:27) of the claim that science relies on narrative. That scientists tell stories on television (even if it were always the case that this is what they did) tells us far more about television than it does about science. The public ridicule heaped upon those who tell great stories but cannot deliver the goods, such as the prophets of cold-fusion in a test-tube, or of new ‘cures’ for HIV infection, is ample evidence of this. The plain fact is that the stories do not legitimate science at all, but function rather as a stand in for an entirely non-narrative explanation which must be assumed to be both available and satisfactory for the story to sell.

I am certainly not trying to suggest here that modern science has no political dimension, or that it has successfully placed itself beyond the reach of social or philosophical criticism. The point is rather that modern science is political in its own particular ways because it is partly constituted by the idea of there being a kind of epistemological ‘bottom line’ to which it has the best possible kind of access, and that consequently it is at least relatively immune to criticism from discourses which do not have this kind of access. Not only is this claim made, but it has some force. Feyerabend’s protest in Science and a Free Society that modern societies are Copernican by a kind of coercion, since the question of heliocentrism was never put to a vote or referendum (1978:74), is absurd for this very reason. Heliocentrism is supported because it is true, and we can no more submit the question whether the earth travels around the sun to political deliberation than Canute could decree that the tides turn back. What is important about the way this claim of epistemic superiority on behalf of science functions is that it is not a competing (meta)narrative claim, but an anti-narrative one, made in the name of a language which supposes itself entirely to renounce the metaphorical.

To the extent that this is so, we must be wary of Lyotard’s suggestion that ‘It is therefore impossible to judge the existence or validity of narrative knowledge on the basis of scientific knowledge and vice versa: the relevant criteria are different’ (PMC:26). Modern style natural science precisely is that kind of knowledge which claims both to eschew narrative as a source of legitimation and to be able to trump the legitimating attempts of narrative or non-scientific discourses. In both cases it does so via a kind of Procrustean violence of just the sort Lyotard abhors, a point to which I return in the discussion of postmodern science below.

Before postmodern science can directly be discussed, though, I need briefly to make good on my earlier claim that natural science and mathematics play the central role that I have suggested they do in Lyotard’s Report. We can note the following as evidence. First, the argument of The Postmodern Condition, almost from beginning to end, observes a distinction between ‘science’ and ‘narrative’ which is, if anything, more pronounced than that between ‘modern’ and ‘postmodern’. The only two kinds of pragmatics which Lyotard discusses in detail are those of science and story telling in sections 6 and 7 of his report and, as noted, he maintains that science and narrative have ‘always’ been at odds. Second, in a variety of ways Lyotard makes use of the cybernetic identification of knowledge or information with negative entropy, and often operates with a specifically thermodynamic notion of efficiency and control. This is so much the case that in a discussion of determinism he refers to Laplace’s fictional ‘demon’ as the ‘ideal fulfilment’ (PMC:55) of the goal of maximised thermodynamic performance. (This is a revealing slip, since, as can be seen from the quotation above, Laplace’s hypothetical epistemologist has equally transparent access to the past and the future on the basis of its grasp of the present facts and the laws of nature. Since almost none of the systems studied by thermodynamics are reversible, retrodiction is generally impossible.) Third, it is a central claim in Lyotard’s argument that technical features of the ways we store and transmit data and information impose constraints on the ways we will think of information and knowledge. The constraints he has in mind are essentially mathematical, relating to the demand for commensurability apparently required to allow knowledge to be translated into digital media, and circulated in various ways (PMC:3-4). Fourth, and finally, all three of these strands come together in the crucial section 13 of The Postmodern Condition, where both the practice and the logic of legitimation of modern science are characterised in thermodynamic terms and arguments from natural, formal and information sciences used to defend the conclusion that science now stands in a new relationship to narrative. Whatever else is going on in The Postmodern Condition, then (and there is plenty more) science and mathematics are major elements in Lyotard’s strategy to commend postmodernity, and significant sources of statements for his philosophical language game.

The method of language games is, however, an at least questionable instrument when it comes to the study of science, perhaps especially so for natural science. Lyotard explains that his method, which we could call his philosophy of science, involves ‘emphasising facts of language and in particular their pragmatic aspect,’ (PMC:60) which means attending to the ways in which language games and practices function to make some statements permissible and others prohibited because of what is said, how it is said or who says it. The main weakness of such an approach is that it is insufficiently discriminating with reference to the question of the truth of any given scientific theory or assertion. This question is important precisely because, as we have seen, it is just such a truth claim which underwrites the anti-narrative trumping to which modern natural science pretends the right. In order to engage with this question, though, we need to face up to what Bhaskar calls the ‘central paradox’ of philosophy of natural science, which is that ‘men [sic] in their social activity produce knowledge which is a social product much like any other ... [but] that knowledge is ‘of’ things which are not produced by men at all’. Bhaskar marks this distinction by referring to the social aspect of science (which exists in virtue of the fact that it is thought about in certain ways) as its transitive dimension, and to the fact that science is about something independent of social activity, which is to say that its object is independent of what is thought about it, as its intransitive dimension. He goes on to point out that ‘If we can imagine a world of intransitive objects without science we cannot imagine a science without intransitive objects’ (1978:21-2). This enables the fatal weakness of the language game approach clearly to be expressed: it focuses solely on the transitive aspects of natural science. Thus, even though Lyotard does make reference to the question of scientific truth, what he says is inadequate, since he is reduced simply to insisting that it is part of scientific pragmatics that ‘scientific’ statements be shown to be true, going so far at one point as to agree with Medawar that a scientist is one who ‘tells stories’ but going on to claim that the significant difference between the stories of the scientist and those of other story tellers is that the scientist is ‘duty bound to verify them’ (PMC:60).

Expressed in this way it seems as though verification is merely something scientists do, internal to their own language games, and having no important consequences for the integrity or legitimacy of other practices, or for science considered with respect to those practices. But this is exactly what is at stake in the questions of the legitimacy of science and of the status of science under conditions of postmodernity.

The failings of Lyotard’s approach can most clearly be seen with reference to the question of ‘postmodern’ science. Given the central place of science in the argument of The Postmodern Condition, Lyotard’s main discussion of the question, which occurs in a dense and difficult section called ‘postmodern science as the search for instabilities’ has a lot of work to do. It has to show us a convincing picture of natural science operating without either legitimation by metanarrative or the expectation of commensurability, but also to show that such expectations are themselves no longer legitimate in the discourses of science. To the extent that they are legitimate, even if they are less popular than before, science simply remains modern, and hence capable of terrorising other discourses and practices in the ways described above. It is my argument, though, that everything that Lyotard says about so-called postmodern science is entirely compatible with the standard image of modern science. If this is so then the ways in which modern science trumps narratives, and hence legitimates itself, will consequently be unaffected.

Lyotard’s discussion is supposed to show that science does not advance by means of what he calls the ‘positivism of efficiency’, but, rather by means of counterexample, paradox and unintelligibility. Specifically he maintains that the ‘pragmatics of scientific research ... emphasizes the invention of new ‘moves’ and even new rules for language games’ (PMC:53) and refers approvingly to Feyerabend’s anarchism (or dadaism) concerning scientific method. To the extent that any genuine progress, or even change, in science has to involve a break with existing expectations, Lyotard’s assertion is a bland truism. That said, it is worth warning against a possible and serious misconception. It would be an error to equivocate the scientific imperative to novelty with anything like the imperative to innovation which is characteristic of the avant garde, which we know Lyotard considers to be an important element of the postmodern attitude.

This is especially dangerous with reference to the question of legitimation. The fact that some candidate for being approved as a scientific statement, rule or ‘move’ is new indicates at best that it might represent an interesting break with received wisdom. It will, however, more than likely prove to be an epistemological dead end, and in the event that this is the case its demise will not be slowed in the least by any sentimentality over its novelty. On the other hand any as-yet-undiscovered and liberating break with the great stifling mass of things which have ‘been done’ is, from an avant garde perspective essentially legitimate. The bald fact of the matter is that nothing is ever legitimate in science because it is new, unexpected, paradoxical or unintelligible, nor is anything in science ever illegitimate because it is old, consistent, intelligible or expected. The very latest dictionaries of physics list Kepler’s laws, yet Fleischmann and Pons don’t even get to be on Oprah.

Lyotard relates the ‘positivist philosophy of efficiency’ to classical determinism, saying that determinism is ‘the hypothesis upon which legitimation by performativity is based’ and that it presupposes that ‘the system into which the input is entered is stable; that the system must follow a regular ‘path’ that it is possible to express as a continuous function possessing a derivative, so that accurate prediction of the output can be made’ (PMC:53-4). The role determinism plays in legitimation is simple, since it is deterministic systems which hold out the promise of accurate prediction (or better still, successful control) being given the status of a powerful and direct confirmation of theory. In a deterministic world a well supported theory just is one which produces accurate predictions.

There is a danger here, though, which is to slip from noting the importance of determinism to natural science in its modern form to imagining that modern science stands or falls with the fortunes of determinism. The kinds of self-legitimating and anti-narrative moves characteristic of modern science are certainly facilitated by determinism, but they by no means require it. A physical description can still trump a non-physical one as long as the physical version can justify the claim that it is the best account, even if it falls short of being deterministic. It is worth noting that Papineau’s definition of the completeness of physics requires only that ‘physical events are determined, or have their chances determined, by prior physical events according to physical laws’ (1993:16, first emphasis mine).

In any event, Lyotard’s procedure is to attack determinism and the ‘positivist philosophy of efficiency’ by citing ‘a number of prominent examples as evidence against it’ (PMC:54). These citations will apparently show that postmodern science seeks out the unintelligible and the paradoxical, which is then legitimated via the invention of new rules in an ongoing heterogeneous process, Lyotard’s main exhibits being quantum mechanics, thermodynamics, fractal geometry and chaos. The target throughout is the notion of a stable system which underwrites the ideal of legitimation via performance, and of which notion Lyotard sees Laplace’s fiction of the vast intellect as the ‘ideal fulfilment.’ Lyotard asserts that this fiction ‘is sustained by the principle that physical systems ... follow regular patterns, with the result that their evolution traces a regular path and gives rise to ‘normal’ continuous functions...’ (PMC:55). We can say, then, that Lyotard hopes to show that developments in the physical sciences themselves have undermined the conditions for the viability of physics in its modern form, a reading supported by his remark near the opening of The Postmodern Condition to the effect that it is partly ‘progress in the sciences’ which explains the collapse of faith in metanarratives, and the onset of postmodernity.

Lyotard’s first citation comes from microphysics, and he states that the ‘advent of quantum mechanics and atomic physics has limited the range of applicability of this principle [i.e. that of legitimation by performativity] in two ways, the respective implications of which differ in scope.’ The first of these limitations is, apparently, that ‘a complete definition of the initial state of a system (or all the independent variables) would require an expenditure of energy at least equivalent to that consumed by the system to be defined’ (PMC:55). It is far from clear what kind of problem this is supposed to be, or for who. The way Lyotard puts things (if we allow his reading of quantum mechanics to stand) it is evident that anyone who wants the information should just spend the required amount of energy. This would certainly be uneconomic, although the point to be made here is surely that expensive determinism is still determinism, and what is at stake is legitimation by determinism rather than cost. Further, even where cost is considered to be an important factor it is clear that the simple fact of determinism being expensive does not point to any non-physical factors being relevant to deciding what will actually happen, and less so to their being in any being in a position to trump any physical claim, or to resist being trumped by one. That scientists need huge amounts of money to deliver predictions which they claim are in principle deliverable does not seem to threaten the legitimacy of science.

What is perhaps more striking about Lyotard’s argument, though, is that it is based on an unjustifiable interpretation of quantum mechanics. There is every reason to believe that Heisenberg’s uncertainty relations should not be interpreted as setting limits only to our knowledge, but rather as expressing genuine indefiniteness in reality. There are two pairs of uncertainty relations in Heisenberg’s account, each of which can be represented by an inequality. The best known states that product of the uncertainty in position and the uncertainty in momentum of a particle must be greater than or equal to a fraction of Planck’s celebrated quantum of action. The second indicates that the same state of affairs obtains for the respective uncertainty in the time at which some event takes place and the quantity of energy involved. Realist readings of these relations permit elegant explanations of various phenomena, such as the width of spectral lines or quantum tunnelling, without unpleasant ad hoc hypotheses being necessary to account for limitations in our knowledge if we suppose that the uncertainty is an epistemological condition only. Such an approach shows that the ideal of a complete definition of a physical system in the classical modern sense becomes strictly unattainable, since, for example, an electron with a very definite momentum just does not have a well defined position. Other more powerful arguments against classical physics can be derived from the quantum mechanical crisis over locality, or by considering the fact that quantum mechanics has as its central object discontinuous changes (hence the very term ‘quantum’) of a sort anathemic to classical determinism. Note that none of these approaches have anything significant to do with the expenditure of energy required to meet classical objectives, but rather with the abandonment or revision of the very terms in which those objectives were formulated. Instead, though, Lyotard imagines that arguments from quantum mechanics are only a limited objection, since they allow that ‘within the framework of the unreachable - but conceivable - limit of the total knowledge of a system’ (PMC:56) classical determinism remains viable!

For other reasons, though, Lyotard thinks that quantum mechanics forces us to face up to a ‘far more radical revision of the idea of a continuous and predictable path’ which supposedly follows because of considerations which arise when we try to measure the density of a system (his example is a volume of air) at various scales. If we assume a kinetic molecular model of gasses, where a gas is composed largely of empty space containing many rapidly moving (and colliding) particles and also an early 20th century image of atoms as themselves mostly empty space, with a very dense core around which extremely small electrons orbit, then we can follow Lyotard’s argument. When we measure density at scales large compared to the size of the particles comprising the gas we will discover values close to classical predictions. When our sampling volume is comparable to the size of a gas particle, though, at most times we will measure a density of zero or close to it, while at others (when a molecule, or even the nucleus of an atom, happens to be ‘in’ the sampling space) we will find values ‘spectacularly higher than those obtained previously’ (PMC:56, quoting Perrin). Careful contemplation of this fact will, Lyotard says, leads to the conclusion that ‘[k]nowledge of the density of air thus resolves into a multitude of absolutely incompatible statements’ (PMC:57).

It is very difficult to be charitable towards this argument. To begin with, and directly contrary to Lyotard’s bald statement to that effect, it has no important connection with quantum mechanics. Secondly, the results described are in no way radical or paralogistic. Thirdly, and most importantly, the argument has no consequences at all for the legitimation of science. That measurements of density are sensitive to scale and location can hardly be news when density just is a relationship between mass and volume. The illicit move, or even confusion, which underpins Lyotard’s argument is that he is contemplating density measurements made at different times with respect to a particular volume of space which is only periodically occupied by the bearers of mass under consideration. If we measure density at a specific time and the sample volume is empty or close to it, then a very small or zero result is entirely correct, just as a later measurement of the same volume when it is nearly ‘full’ should give a comparably very high value. Most importantly of all, averaged over a large number of measurements, a series of these instantaneous samples will give a result equal to the very classical prediction Lyotard thinks he has shown to be inconsistent with observation. Simply put, the only way to arrive at a ‘multitude of absolutely incompatible statements’ is to ignore facts which we have no business ignoring.

Such details aside, though, what is abundantly clear here is that Lyotard is party to a deep and comprehensive misunderstanding of quantum mechanics. There are two particularly important aspects to this. The first is that, for all its unexpectedness from a narrow classical point of view, quantum mechanics is not in the least avant garde or paralogistic, or at least not any more. It is a highly successful, well established, thoroughly tested part of mainstream physics. It has textbooks, journals, conferences and is transmitted through precisely the same processes and mechanisms Lyotard describes in his account of the pragmatics of scientific education elsewhere in The Postmodern Condition. Its findings and methods are at the heart of a wide range of technological achievements which show that it has become an instrument of that most dull and stable forms of legitimation by performance: engineering. To be a radical about the question of the basic constituents of matter today does not mean supporting quantum mechanics, it can only mean attacking it. The second difficulty is that, at least as Lyotard understands it, quantum mechanics does not indicate any break within the hegemony of physics, at least not in virtue of the manifest lapses of determinism he describes. So that no matter how unusual quantum mechanical descriptions might be, with all their talk of state reduction and statistically specified outcomes, no-one in physics seriously thinks QM shows that any non-physical factor could possibly have any bearing on the development of a microphysical system, and physical descriptions are consequently still holding trumps. So from the point of view of legitimation the modern state of affairs stands unaffected.

Fortunately for Lyotard his argument has more to it than quantum mechanics. In response to the imagined objection that what he has discussed is not especially important since it applies only to the microphysical world, and leaves determinism at the larger scales unchallenged, Lyotard shifts his emphasis to what he calls ‘contemporary mathematics’ which is to say chaos theory and non-linear dynamics.

As is relatively well known now, what are called in natural science ‘chaotic’ systems are deterministic systems which exhibit unpredictable behaviour. This takes place because the dynamical properties of the systems are such that they evolve in an aperiodic way despite having finite possibilities of action. This means that even a simple system such as a computerised weather model stripped down so that the only significant variables are the speed and direction (restricted to clockwise or anticlockwise) of a single convection cell can follow a path on which the sequence of changes of direction and speed of rotation never repeat (Lorenz 1963). Prediction of the behaviour of such systems is in principle impossible, since the – unavoidable – difference between our measurements and the actual states of the system result in a rapidly increasing difference between the evolution of the simulation and that of the real system. This is the explanation behind the famous and misunderstood ‘butterfly effect’ where such apparently insignificant (and ultimately unmeasurable) factors as the flapping of the wings of a single butterfly on one side of the world cannot be considered irrelevant if we want a reliable prediction regarding the possible occurrence of a typhoon on the other. Which is to say that reliable predictions of typhoons, even assuming that their occurrence in fact follows deterministic laws, is not possible.

The problems for Lyotard here, though, are the same as before. Non-linear dynamics is no more crazy or paralogistic than quantum mechanics. Indeed, since its existence is not even a minor problem for determinism, it is even less so. On the contrary, it is a powerful argument for the credibility of determinism even in cases where prediction is unreliable or impossible. Precisely because chaotic systems are routinely deterministic they can hardly count as a crisis for legitimation by performance. Further, chaos theory neither requires nor facilitates any break with the notion that material phenomena are optimally approached by mathematical physics. Then, finally, chaos theory is quite a safe and respectable part of normal science, with all the trappings of institutional acceptability noted in connection with quantum mechanics.

The same goes for Lyotard’s treatment of the related area of fractal geometry. This branch of mathematics exists because it has been shown quite convincingly that classical geometric concepts of measurement are incapable of dealing with common features of most natural objects. The most famous example is probably that of the measurement of a coastline, where Mandelbrot showed that depending on the sampling length of the survey a different result for the length of a coastline would be found, since shorter samples would have to follow smaller indents and outgrowths not resolved by the larger samples. This means, roughly, that a creature with a one centimetre stride will take more than one hundred times as many steps as one with a one metre stride, to traverse the boundary of a real coastline. (This is not the case with a classical geometrical object such as a rectangle, and differs from the kind of variation we find with regular curves such as circles.) Again, though, rather than being paralogistic or radical, fractal geometry too has been comfortably integrated into mainstream science, with precise methods for quantifying the extent to which a type of surface or boundary is non-Euclidean and consequently clear rules for when to use which mathematical instruments. And this has been done without any serious suggestion that anything from Lyotard’s catch–all category of ‘narrative’ need ever play a role in geometry.

Such arguments as Lyotard offers, including those which follow about catastrophe theory, are simply ineffective against the ‘positivism of efficiency’ despite Lyotard’s bold claim to have found ‘an answer in the debate between stable and unstable systems, determinism and nondeterminism’ or to have given us a convincing picture of a science ‘theorising its own evolution as discontinuous, catastrophic, nonrectifiable, and paradoxical’ and which is legitimated ‘solely on paralogy’ (PMC:61). Novelty, the unusual, and the ongoing search for new results and methods are important in scientific practice, but the arguments discussed show Lyotard to be prone to a fundamental confusion between method and intention similar to that which plagued positivism. In a sense he gives us an upside down Kuhnian position, where revolutionary science is set up as a normative model for all scientific practice. But even a parade of relevant and powerful examples of individual scientific research programmes which were in significant ways radical or in fundamental opposition to received wisdom would not show that science was now in general ‘discontinuous, catastrophic, nonrectifiable, and paradoxical’ (PMC:60). Even less would it show that it was just these features which were centrally important to the legitimation of scientific work, either to scientists or anyone else.

This is so despite the fact, and it is a fact, that the sociological changes Lyotard describes are indeed an important characteristic of much contemporary science. Less work is being done on large scale attempts to unify theories. More work is being done by laboratories and research groups trying to solve specific technical problems with no real interest in how their solution might fit into a larger picture. And there is indeed a pronounced shift in funding patterns away from academic research aimed at scholarly publication and towards commercially oriented research most often directed at registering patents, and at producing results which are destined to be thought of as a kind of private and marketable property. The question whether science is legitimated ‘solely on paralogy’, though, is a philosophical one, and it should not be taken as answered by a few specific instances from the sociology of science.

As I showed at the beginning of this paper, Lyotard could evidently see that the claims of natural science stood in the way of the image of our intellectual landscape which he wished to develop and endorse in The Postmodern Condition. This is evidently one of the major reasons for his paying it so much attention in his argument, in effect by means of an attempt to co–opt contemporary physics and mathematics for his purposes. Briefly those purposes were to argue that the various realms of discourse we might in the past have been tempted to classify as ‘narrative’ or ‘scientific’, and even to legitimate by reference to grand or meta- narratives, now stand in more open relationships worked out by means of an ongoing micro–political process, where no realm of discourse has access to a well-justified right to veto the claims of another. But as we have just seen Lyotard’s arguments that science is along for the postmodern ride are inadequate.

In consequence the situation Lyotard is left in is ultimately no different from that of Rorty, briefly discussed and rejected above. Recall that Rorty underwrites his version of heterogeneity and the possibility of an ongoing conversation between multiple discourses by reference to the difficulty of physicalism producing trumping accounts, while still allowing that such manoeuvres could be carried out in principle, and also as an ‘occasional pedagogical exercise’. If we follow Lyotard’s simplification in calling all non-scientific discourse narrative, then we can say confidently that his arguments leave narrative in the same secondary position as Rorty’s hermeneutics, since both endorse a postmodern hermeneutic enthusiasm for freedom of conversation, while ultimately leaving to physics (although not deliberately in Lyotard’s case) the questions as to what noises we will make and marks we will write.

What this means in both cases is that the condition of postmodernity has to be seen as one restricted to areas not covered by natural science. More significantly, even in its own domain the fashion for postmodernism is undermined by this state of affairs, since every effect to which it attaches a collection of tales and discourses and narratives could in principle be subjected to the commensurating and trumping gaze of natural science. And in this process differences which do not admit of translation would not be accorded the kind of awe-struck respect modelled on Kant’s notion of the sublime which Lyotard defends. Rather they would be lucky to get off with the status of secondary qualities in the style of 17th century natural philosophy. This is an interpretation of matters which would probably find favour with many natural scientists and those elsewhere with a naturalist temperament, but it is one which runs directly against the explicit intentions of Rorty, Lyotard and others who ride with the postmodern brand. In fact it is a clear example of what Lyotard would have us regard as a wrong, a breach of justice, and an excuse for terror which is precisely what the postmodern sensibility is supposed to help us avoid.

This puts me in a position to state very briefly the conclusion of this paper, which has not been directed at coming down on either side of the divide between the philosophical tendencies manifest in the images of the Cartesian and Wittgensteinian cities discussed earlier on, but rather at the claims made on behalf of one influential defence of the Wittgensteinian polis. My conclusion is that Lyotard’s politics are sold short by his analysis of science. What he needs is a displacement of the monolithic voice of modern commensurating natural science. What he offers us reduces, in the face of careful critical examination, to an image of a natural science rather more tolerant of internal conflict and partial solutions than its classical beginnings, but still hugely intolerant of non-science, and in most politically interesting ways entirely modern. If postmodernism is to be what it sometimes pretends to be then it stands in need of a vastly more penetrating and effective engagement with natural science. To the extent that such cannot be provided we have significant reason to be suspicious of the more sweeping and comprehensive claims made on behalf of postmodernism, and none more so than those which assert the ubiquity of the postmodern condition.



Bhaskar, R. 1978. A Realist Theory of Science. (Second Edition) Sussex: The Harvester Press.

Bhaskar, R. 1991. Philosophy and the Idea of Freedom. Oxford: Blackwell.

Brooke, J. H. 1991. Science and Religion: Some Historical Perspectives. Cambridge: Cambridge University Press.

Descartes, R. 1985. The Philosophical Writings of Descartes, 2 vols. (translated and edited by Cottingham, J. G., Stoothoff, R., and Murdoch, D.), Cambridge: Cambridge University Press.

Feyerabend, P. 1978. Science in a Free Society. London: Verso.

Foucault, M. 1970. The Order of Things: An Archaeology of the Human Sciences. (translated from the French) London: Routledge.

Hacking, I. 1975. The Emergence of Probability. Cambridge: Cambridge University Press.

Kant, I. 1929. Critique of Pure Reason. (translated by Norman Kemp Smith), London: MacMillan.

Laplace, P. S. 1951. A Philosophical Essay on Probabilities. New York: Dover.

Lorenz, E. N. 1963. ‘Deterministic Nonperiodic Flow’, Journal of the Atmospheric Sciences (20) 1963, pp130-141.

Lyotard, J-F. 1984. The Postmodern Condition: A Report on Knowledge. (translated by Bennington, G and Massumi, B.), Manchester: Manchester University Press.

Lyotard, J-F. 1988. The Differend: Phrases in Dispute. (translated by Georges Van Den Abbeele), Minneapolis: University of Minnesota Press.

Newton, I. 1934. Principia: Vol. I The Motion of Bodies. (Motte’s translation revised by Cajori), Berkeley: University of California Press.

Newton, I. 1934. Principia: Vol. II The System of the World. (Motte’s translation revised by Cajori), Berkeley: University of California Press.

Norris, C. 1997. Against Relativism: Philosophy of Science, Deconstruction and Critical Theory. Oxford: Blackwell.

Papineau, D. 1993. Philosophical Naturalism. Oxford: Blackwell.

Rorty, R. 1980. Philosophy and the Mirror of Nature. Oxford: Blackwell.

Rorty, R. 1991. Objectivity, Relativism and Truth. Cambridge: Cambridge University Press.

Spurrett, D. 1997. ‘Beyond Determinism’, South African Journal of Philosophy, Volume 16 number 1, pp14-22.

Whitman, W. 1855. Leaves of Grass. New York: Walt Whitman.

Wittgenstein, L (1958) Philosophical Investigations, (translated by G E M Anscombe), Oxford: Blackwell.