UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The troubled meeting of Richard Rorty and Thomas Kuhn Foulkes, Erica 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1999-0189.pdf [ 8.62MB ]
JSON: 831-1.0089093.json
JSON-LD: 831-1.0089093-ld.json
RDF/XML (Pretty): 831-1.0089093-rdf.xml
RDF/JSON: 831-1.0089093-rdf.json
Turtle: 831-1.0089093-turtle.txt
N-Triples: 831-1.0089093-rdf-ntriples.txt
Original Record: 831-1.0089093-source.json
Full Text

Full Text

T H E T R O U B L E D M E E T I N G OF R I C H A R D R O R T Y A N D T H O M A S K U H N by E R I C A F O U L K E S B . S c , The Universi ty of Vic tor ia , 1980 R.T. , Nor thern Alberta Institute of Technology/Edmonton General Hospital , 1983 B . A . , The Universi ty of Bri t ish Columbia , 1995 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF A R T S in T H E F A C U L T Y OF G R A D U A T E STUDIES Department of Philosophy W e accept this thesis as conforming to the r e q u i r e d standard T H E U N I V E R S I T Y OF BRITISH C O L U M B I A A p r i l 1999 <gj Er ica Foulkes, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, 1 agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbt Vancouver, Canada Department Date J£ far<l 7 DE-6 (2/88) Abstract The philosopher Richard Rorty counts himself among the supporters of Thomas K u h n , who has done extensive and important work i n the history and philosophy of science. Rorty bases his support on a number of similarities wh ich he sees between his o w n work and that of K u h n . M y intention i n this thesis has been to demonstrate the distinct lack of s imilari ty between the work of Ror ty and that of K u h n . It is m y contention that Rorty's belief i n the sympathy between their respective programmes is misguided, being grounded i n Rorty's misunderstanding of a considerable number of elements i n Kuhn's programme, as we l l as the intent w i t h w h i c h it was offered. The first chapter is an exegesis of Rorty's general phi losophy of culture, moving into a more specific exegetical look at Rorty's references to Kuhn's work. The second chapter provides an extensive outline of Kuhn's historically sensitive examination of the philosophy of science. The th i rd chapter engages i n a discussion and analysis of comments w h i c h Rorty has made concerning Kuhn's work and the ways i n w h i c h it relates to his own , encompassing both similarities and differences. I believe that I have shown that Thomas Kuhn's w o r k i n the history and philosophy of science is not at a l l applicable to the sort of cultural programme w h i c h Rorty is offering, i n terms of scope, intent, and fundamental content. TABLE OF CONTENTS Abstract ii INTRODUCTION •. 1 CHAPTER I Overcoming Epistemology: An Overview of Rorty's Cultural Programme 4 CHAPTER II Thomas Kuhn: Understanding Science 20 Introduction 20 Normal, or Ordinary, Science 21 Scientific Crisis, and the Transition into Revolutionary, or Extraordinary, Research 29 Empiricism, Kantianism, and the Spectre of Relativism 44 Summary of the Relationship between Normal and Revolutionary Science 53 The Distinction Between Science and Non-science 56 The End of Science 58 CHAPTER III Allies or Adversaries: The Troubled Meeting of Rorty and Kuhn 60 Nature, Mind, and Correspondence Theories 60 Overcoming Epistemology. 63 The Need for an A Priori Framework 64 Rationality and Solidarity 70 Science and Cultural Holism 76 Pragmatism. 81 Truth and Relativism 86 Alternative Discourses/Matrices 89 The Question of Incommensurability 91 Hermeneutics 100 CONCLUSION 108 BIBLIOGRAPHY 110 iii Introduction W i t h some qualifications to w h i c h I w i l l attend i n the body of this work, Richard Rorty counts himself among Thomas Kuhn's supporters. In his Philosophy and the Mirror of Nature, Rorty describes K u h n as one of the "heroes of this book". 1 1 w i sh to argue that Rorty's sentiments i n this regard are grossly misplaced, and that his sense of alliance w i t h Kuhn's project is achieved at the cost of his misunderstanding it. The first chapter is intended as a strictly exegetical section, covering Richard Rorty's general phi losophy of culture. The latter part of this chapter w i l l look i n more specific detail at Rorty's references to the work of Thomas Kuhn ; I intend there to provide evidence for Rorty's attempt to enlist K u h n as someone whose work promotes the direction of his own. The second chapter provides an extensive outline of Kuhn's programme, wh ich latter can be taken as an historically sensitive examination of the philosophy of science. This section is also strictly exegetical. The th i rd chapter engages i n a discussion and analysis of comments w h i c h Rorty has made concerning Kuhn's work and the ways i n w h i c h it relates to his o w n . This w i l l encompass what Ror ty sees as both similarities and differences. It is important to understand that it is not m y intention, i n this work, to evaluate other writers w h o m Rorty invokes or directly quotes; they are merely included as vehicles for the element i n Rorty w h i c h I w i s h to consider, v iz . , the relation of his philosophy to the work of K u h n . I w i l l therefore not be evaluating Rorty's work on the basis of internal consistency or the lack of it; nor w i l l I be considering whether either Rorty's or Kuhn's programme provides a plausible explanation, i n my view, of the way i n w h i c h we ought to consider cultural and /or scientific activities. I believe there to be few similarities between Ror ty and Kuhn , and these are relatively superficial. The relationship between them is characterised pr imar i ly by divergence, al though it is m y contention that Rorty fails to see the degree to w h i c h he and K u h n diverge, on account of his quite substantial misinterpretations of K u h n . Rorty understands his talk of cultural discourses to represent a generalisation of Kuhnian scientific paradigms. He sees K u h n to be engaged i n a compatible project to his own , by virtue of what Rorty takes to be: 1. their mutual rejection of correspondence theories; 2. their common desire to dispense w i t h the not ion of observation; 1 Rorty, PMN, p. 382. 1 3. the credence w h i c h they give to non-rational factors influencing human behaviour; 4. their framing of rationality pr imar i ly i n terms w h i c h describe it as the t r iumph of techniques of persuasion, rather than of force; 5. their rejection of the desire for and adherence to a framework of accepted conceptual and practical commitments, such as we find i n tradit ional epistemology; Ror ty describes this as their having questioned the need to identify a set of rules for direct ing rational agreement; 6. their orientation towards pragmatism; 7. their common intent to direct us to think of science as interwoven w i t h other cul tural activities: to this end, Ror ty sees K u h n as having "softened the dist inct ion between science and nonscience" 2 ; 8. the importance w h i c h they place upon sensitivity to historical circumstance; 9. the fact that neither of their positions is relativist, despite the possibility of their being construed as such; and 10. their common proposal of hermeneutics as a way of "coping" w i th one's o w n and others' cultures. Nevertheless, Ror ty takes issue w i t h Kuhn's relatively greater allegiance to tradit ional epistemology. Further, given what Rorty sees to be Kuhn's project, he believes K u h n to have given undue credence to the idea that phi losophy has a role i n setting put the nature of knowledge. I w i sh to argue that Rorty has misread K u h n as being, overall, inc l ined against t radi t ional understandings of philosophy, epistemology, and rat ional i ty-as Ror ty himself is. W h i l e K u h n does push for a more historically sensitive philosophy of science, he does not intend to negate the function of rational thought, the possibility of some identifiable commonal i ty w h i c h l inks the thought and experience of human beings, or the importance of developing a theory of knowledge. K u h n does not concern himself w i th -e i the r to accept or reject-the question of philosophy's functioning as an all-encompassing discipline w h i c h serves to ground and l imi t other disciplines. Both Rorty and K u h n reject correspondence theories of knowledge. However, Kuhn's retention o f the Kant ian not ion of Nature as provid ing that w h i c h is "given" i n perception leads h i m also to retain observation as an important feature of scientific work. Whi l e both Rorty and K u h n eschew any ontological commitment, Kuhn's sense of the "given" prevents h i m from adhering to a strictly linguistic v iew o f things; Rorty's out look is l inguistic i n orientation. K u h n describes himself as i n some sense a Kant ian. In this light, he is not at a l l inc l ined towards the rejection of a conceptual and practical framework; i n fact, the purpose of his scientific paradigms is precisely to describe the various frameworks by wh ich scientists have operated. He refers to these paradigms as "moving, historically-situated, Archimedean platforms" 3 , w h i c h direct how acceptable scientific thought and practice are to be carried out. Rorty does not 2 Rorty, ORT, p. 38. 3 Kuhn, RSS, pp. 6-7. 2 seem to be adequately aware of this structuralism w h i c h is inherent i n Kuhn's programme. K u h n views science as pr imar i ly a pragmatic endeavour, but he spells this out more clearly than Rorty does for his own cultural "pragmatic hol ism". Final ly, i f what K u h n has said serves to "[soften] the dist inction between science and nonscience" 4 , it is clearly not his intent to dissolve scientific activities into those of the culture at large. Kuhn's work is specifically directed towards clarifying the way i n wh ich science works, and has worked throughout its history. 4 Rorty, ORT, p. 38. 3 1 Overcoming Epistemology: An Overview of Rorty's Cultural Programme Richard Rorty, albeit with some qualifications to which I will in due course attend, counts himself among Thomas Kuhn's supporters, rather than among his critics. It is my purpose in the first section of this chapter to provide a general overview of Rorty's programme and, later on in the chapter, to outline more specifically what Rorty has to say about the relationship of Kuhn's work to his own. This chapter is intended to be exegetical, and not evaluative of either Rorty's philosophical remarks or his understanding of Kuhn. Rorty describes the central concern of philosophy as the instantiation of itself as a general theory of representation, such that its task is to divide the culture into several areas: those which represent reality well, those which represent it less well, and those which fail to represent it at all. In this model, where we have knowledge, we have accurate representation, which is made possible by special mental processes, and made intelligible through a general theory of representation. This representational model, Rorty contends, was born out of the enclosure of mental activity in a special kind of mental substance, one which was to be understood as isolated from non-mental, or physical, substance. As the mind looks upon the world, it receives representations, which are in the mind, and which are therefore able to be viewed by the Eye of the Mind. The mental thus represents the non-mental, and the door is opened to scepticism and the haunting question, "How do we know that anything which is mental represents anything which is not mental?"1 Rorty goes on to attack this notion of mind on the basis of what he sees to be its superfluity. He asks why it should make any difference whether or not we have minds. Behind this lies his view that mental states could sensibly be viewed merely as place-holders for talk of neurons. So, for example, in certain situations, such as contact with a hot stove, a brain-event occurs: we might call this a case of "stimulated C-flbres". An expression of pain results, the communication of which being all we need to understand what has happened. There is no need to formulate a hypothesis which includes anything such as "raw feels". Neither do we need to posit a special mental substance which would accommodate such 1 Rorty, PMN, p. 46. 4 mental entities as these. Moreover, we have so far failed miserably i n our attempts to identify any causal mechanisms w h i c h might serve to describe how particular raw feels could be l inked w i t h particular neural-stimulation events. It may s imply be, that when we report what we misguidedly take to be raw feels, what we are, i n effect, reporting is neural stimulations, the impl ica t ion being that perhaps we ought just to do this directly, wi thout inserting the misleading talk w h i c h promotes an unnecessary and confusing positing of the m i n d . 2 Ampl i fy ing this difficulty that we w i l l never be able to explain the difference between feeling a pain and s imply reacting to, say, a stimulated C-fibre, is the point, crucial for Rorty's philosophical view, that there is already no detectable difference between the two "from the outside". A n observer of someone i n pain w i l l have no means of dist inguishing whether the v ic t im is experiencing a peculiarly mental entity or s imply reacting to the physical s t imulat ion of certain neurons. It is i n the examination of this latter point that we are confronted, once we set foot on epistemic ground, w i t h the seductive appeal of metaphysics. If we take seriously the possibility of such private mental entities as raw feels, a long w i t h our abil i ty to speak of these inner states i n the absence of any talk of their behavioural accompaniments, then there is every reason to consider raw feels as serious candidates for ontological status. Then the door is open to the not ion of a mental substance, of w h i c h these entities serve as modes. Moreover, an environment is created i n w h i c h scepticism about the existence of other minds can flourish. But Rorty believes that he can furnish us w i t h a solution to this incommodious state of affairs. W e should drop the not ion that to have knowledge of something requires our being acquainted w i t h its "special, felt, incommunicable qualities". To have knowledge of the presence of a pain is to understand the meaning of "pain", and this requires being taught to label a certain state (i.e., C-fibre stimulation) as a pain-state, and to behave i n a certain manner w i t h respect to that state. Unless there were such a th ing as typical pain-behaviour, we w o u l d never be able to teach a chi ld the meaning of the w o r d "pain". But such behaviour belongs to the public domain . Accord ing to Rorty, our assumption that we learn what pain is by "casting linguistic garb over...our direct acquaintance w i t h special, felt, incommunicable qualities" 3 is i n error. To deepen the problem, the not ion that language is used to name that w h i c h is naturally already given (known by the mind's Inner Eye) leads us into scepticism: since knowledge of felt mental entities is not only privileged (i.e., incorrigible), but incommunicable , we end up being irremediably sceptical w i th respect to whether or not others are naming the same incommunicable quality that we are when speaking of pa in . For the sceptic, facts about behaviour and environment are irrelevant to the essence of pa in . Since we cannot tel l whether what we are accustomed to cal l ing "raw feels" are truly experiences of mental objects or stimulations of certain neurons, then we ought, by a principle of philosophical economy, to abolish this excess explanatory 2 Ibid., pp. 70-83. 3 Ibid., p. 110. 5 talk of raw feels (which are hopelessly private and incommunicable , anyway), and speak instead i n materialist language, w h i c h is to say i n terms of neural s t imulation. But this does not mean that we should embrace eliminative materialistic mind-body identity theory, for the materialist is guilty of making metaphysical claims, and this is surely something w h i c h Rorty is concerned to avoid. Rorty's c la im is that it is pointless to pursue the issue of whether a sensory report signals something mental or neural, "not just because nobody has any idea how to resolve the issue, but because nothing turns upon it". N o predictive, explanatory, or descriptive power w o u l d be lost i f we were to describe things l ike pains i n terms of stimulations of particular neural fibres. What we are concerned wi th , after a l l , are these very powers, i n their practical sense. What is more, we do not possess sufficiently clear ideas of what the mental and the physical are to be able to get m u c h use out of a philosophical dis t inct ion between them. He contends that "[o]nly a philosopher w i t h a lot invested in the not ion of 'ontological status' w o u l d need to wor ry about whether a corr igibly reportable pain was ' rea l ly a pain or rather a stimulated C-fibre". Insisting on an ontological divide between mental states and neurons can only generate further unnecessary epistemic and metaphysical problems. 4 Rorty views our separation of m i n d and matter as ul t imately an unfortunate linguistic development-one wh ich has been made a l l the more unpleasant by the slide into metaphysical th inking w h i c h it facilitates. The terminologies of "sensations" and of "brain processes" are nothing more than two ways of ta lking about the same thing. Rorty here anticipates the question: "Two ways of ta lking about what?" But he advises us to resist our natural metaphysical urge, to "abandon argument and fall back on sarcasm, asking rhetorical questions like 'What is this mental-physical contrast anyway? Whoever said that anything one mentioned had to fall into one or other of two (or half a dozen) ontological realms?' " 5 The most that I can make out of Rorty's answer to this question "Two ways of ta lking about what?", is that we are deal ing w i t h alternative vocabularies for ta lking about the w o r l d construed i n something l ike an experiential sense, rather than about the w o r l d as metaphysically characterisable. Hence, it makes no difference per se w h i c h vocabulary we choose to employ, since we cannot tell whether what we are experiencing are mental entities or stimulations of neural fibres. However, given the unpalatable philosophical confusions and endlessly fruitless debates generated by the creation of a definite, ontologically conceived split between the mental and the physical, it w o u l d be best to dispense w i t h one of these lexicons. Rorty opts for the abolishment of the mentalistic sort of vocabulary, on the grounds that it fosters the v iew of knowledge as accuracy of representation, an account w h i c h holds that certainty can only be had rationally about representations. H e sees this sort of v i ew as mak ing the emergence of scepticism inevitable, w h i c h w i l l lead us into the isolation of an inescapable 4 Ibid., p. 120. 5 Ibid., p. 122. 6 solipsism. For h im , this is intolerable, as it must be for anyone whose Utopian vis ion consists largely i n what can be achieved through dialogue w i t h others. Our "abil i ty to say obvious things to ourselves alone" 6 cannot therefore be a mark of philosophical t r iumph. So Rorty's solution entails r idd ing ourselves of the representational model of knowledge, thereby divesting ourselves of the image of m i n d as the mirror of Nature, w h i c h functions as the receptacle for these representations. This reinforces the fact that we can do without the notion of m i n d altogether. What w o u l d serve us better, i n Rorty's estimation, w o u l d be a not ion of "personhood" 7 , based not o n philosophy's description of the human m i n d and its operations, but rather on our mora l intuit ions. It follows from the image of m i n d as mir ror ing N a t u r e - w h i c h rests upon the not ion that the m i n d is naturally "given" to i tself-that we can understand and improve knowledge by pol ishing the mirror : understanding as best we can the workings of the m i n d . The Rort ian imperative is to set aside this offending image of m i n d as mir ror of Nature. In giving this model up, we give up the not ion of philosophy as a discipline committed to the construction of a permanent, neutral framework for enquiry, and thus as a guiding light for the culture. This sort of discipline arose, according to Rorty, i n the scientisation of philosophy w h i c h was promoted through the development of a methodology designed to achieve epistemic certainty. Thus, we can point to a tradit ional pattern: "an attempt to escape from history--an attempt to find non-historical conditions of any possible historical development". The natural quest for understanding w h i c h characterised pre-modern philosophy came to be run together w i t h an unnatural quest for certainty. 8 For Rorty, this development signalled a misguided evolution of phi losophy into epistemology. Rectifying it entails a re-conception of certainty as "a matter of conversation between persons, rather than a matter of interaction w i t h non-human reality". In this, we see a prime issue for Rorty i n his attack on the val id i ty of correspondence theories of knowledge. A n y dist inction between necessary and contingent truths w i l l thereby be erased, to be replaced by a simple differential i n the ease w i t h w h i c h interlocutors find themselves able to disagree w i t h our assertions. This w o u l d place us i n a s imilar posi t ion to the Sophists. "Rational certainty" then becomes a matter of victory or persuasiveness i n argument, rather than relation to an object known. We w o u l d seek explanations for things among the parties to the discussion, rather than among putative faculties of the m i n d . 9 Rorty argues that the idea of foundations of knowledge~the desire for an epistemology-is merely the most recent product of a particular choice of perceptual metaphor. Alternatively, we could think of knowledge i n terms of merely relations of propositions, w i t h no reference being made to perception at a l l . Justification then becomes strictly a relation of inference between 6 Ibid., p. 123. 7 Ibid., pp. 126-7. 8 Ibid., p. 9. 9 Ibid., p. 157. 7 propositions. He acknowledges that this could lead to an infinite regress of propositions invoked to defend other propositions, but assures us that this is of no consequence, since pragmatic conversation requires no more than that we consider such a chain of justification only unt i l a l l the participants i n the discourse are satisfied. To think of knowledge as presenting a problem about w h i c h we ought to have a theory is, according to Rorty, a product of our v iewing knowledge as an assemblage of representations. But i f we see this way of th inking as optional , then epistemology itself is optional, as is any conception of philosophy w h i c h is grounded i n it. Rorty describes the process of justification w i t h i n the epistemological t radit ion as reductive and atomistic. This, because it attempts to isolate pared-d o w n foundations of knowledge w i t h i n the ind iv idua l ; Ror ty has described these foundations as privileged representations. His preferred v is ion of human thought is of something w h i c h he terms "holism", i n w h i c h the justificatory process is "conversat ional" 1 0 . Here, we f ind that justification, rather than being a matter of the delineation of a special relation between ideas (or words) and objects, is instead a matter of conversation as social practice. In this, we see pragmatism confronting the unreality of tradit ional epistemological concerns. A n understanding of knowledge entails an understanding of the social justification of belief. H o l i s m rejects the tradit ional quest for certainty i n w h i c h philosophy has involved itself, and turns away from such terms as "conceptual", "apodictic", or "foundat ional" 1 1 . Epistemology becomes "naturalised", by virtue of becoming itself a matter of empir ical discovery. It is sufficient for us to learn about each other's inner states from reports of them. W e do not refrain from questioning pain-reports on account of their primit ive incorrigibil i ty; rather, the certainty of a pain-report reflects the fact that no one cares to question it. This attempt to explain rationality and epistemic authority by means of reference to societally imposed l imits on linguistic practice illustrates an attitude of "epistemological behaviourism", a species of hol i sm. Understanding the rules of a language-game is the only requirement w h i c h remains "to understand.. .why moves i n that language-game are made (all , that is, save for the extra understanding obtained from inquiries nobody w o u l d cal l epis temological- into, for example, the history of the language, the structure of the brain, the evolut ion of the species, and the pol i t ical or cultural ambiance of the players)". The shift into this version of understanding renders the delineation of necessary and sufficient conditions for knowledge an impossible task. 1 2 It seems that much of this emphasis on linguistic practice is motivated by the difficulty of dist inguishing between a response to language and a response to experience. I presume this to mean that, whi le we can expect the evolut ion of 1 0 Ibid., p. 170. 1 1 Ibid. 1 2 Ibid., p. 174. 8 response to experience to be aided by the development of a theory of knowledge, framing things i n terms of a response to language does not require such a theory, but can get by w i t h a treatment of the conventions of l inguistic practice: a study of the workings of language-games. The emphasis, i n hol ism, upon collective social practice as the guide for what is to count as rational, and the correspondent refusal to a l low for privileged acquaintance w i t h sensory appearances or meanings, is an indicat ion of the absence of any mediat ing element between the impact of the environment upon h u m a n beings and their reports about i t . W e can make no reference to "inner entities" as premises upon w h i c h to base our inferred knowledge of outer entities. Notions of "mind" or "stream of consciousness" are obviated. We remove the tradit ional epistemological bias towards grounding our practices of justification i n "fact"; indeed, we have to ask whether or not the search for foundations of knowledge even makes sense, "whether the idea of epistemic or mora l authority having a 'ground' i n nature is a coherent one". F r o m the pragmatic perspective, it is not possible to see what it might be l ike for social customs to be so g rounded . 1 3 Since the epistemic concern w i t h what is true, and the mora l concern w i t h what is right are more sensibly recognised to be matters of social practice than of any essential facts about humani ty or Nature i n general, the justificatory practice cannot operate outside reference to already accepted beliefs and language-practices. This renders coherence the test of what it is justifiable to assert. Rorty remarks that the latter point may look questionable, since we are used to conceiving of coherence as a by-product of philosophy's extraction of a permanent, neutral matr ix w h i c h functions across the board as a foundation for a l l enquiry, and throughout a l l of history. The privi leged status w h i c h this foundation, on this latter view, w o u l d accord to particular sets of scientific or mora l v i e w s - a l l o w i n g them to be seen as more "rational" than the al ternat ives-w o u l d support the notion that the more relativistic stance of hol ism must rule out coherence theories of justification. A l l that w o u l d be left to the holist engaged i n discourse w o u l d be the abil i ty to construct piecemeal and partial criticisms. But Rorty objects that we cannot help ourselves to anything beyond justification based on hol ism, given that the tradit ional epistemological v i ew rests upon appeal to behaviouristically unveriflable episodes (viz. , raw feels, i n w h i c h the m i n d recognises its o w n direct acquaintance wi th , for example, an instantiation of blueness). W e saw this complaint earlier i n Rorty's argument that we cannot dist inguish between a report of pain as an expression of the presence of a genuine inner entity (a "raw feel"), present immediately to consciousness, and a report of pain as an expression of a st imulation of C-fibres. Because the holist is not dr iven by fear of the epistemological sceptic, as is the foundationalist philosopher, he is able to let go of a justificatory process w h i c h the foundationalist deems necessary, as the ground beneath our feet, i n favour of a process w h i c h grants credence to the more f luid behaviourist hol ism w h i c h justification has t ruly Ibid., p. 178. 9 always been, anyway. "Observation" becomes just what is intersubjectively agreeable these days. An account of the nature of knowledge, then, can be no more than a description of human behaviour. Going back to our example of pain-reports, that we know what pain is merely reflects our ability to link the concept of pain with other concepts, in order that we be able to justify claims about pain; this, I take to be reliant on the framework of coherence described by the rules of the particular language-game in which the conversationalists are engaged. Knowing what pain feels like, in some sort of sense in which pain is immediately present to the mind, is irrelevant. Talk of sensory experience invokes insufficient and unnecessary causal conditions for knowing what sorts of things we are talking about when we make reports about pains, colours, etc.: [They are] insufficient for the obvious reason that we can know what redness is like without knowing that it is different from blue, that it is a colour, and so on. [They are] unnecessary because we can know all that, and a great deal more, about redness while having been blind from birth, and thus not knowing what redness is like. It is just false that we cannot talk and know about what we do not have raw feels of, and equally false that if we cannot talk about them we may nevertheless have justified true beliefs about them.1 4 The traditional notions of "givenness" have confused the ability to discriminate with the having of raw feels. It is Rorty's contention that we can discriminate without having raw feels, and indeed only do discriminate in this way, by virtue of the linguistic community which is the sole true source of epistemic authority. In fact, it is on the basis of the individual's expected membership in the speech-community which prompts us to grant that she is the sort of being who is party to knowledge by raw feels in the first place. Rorty observes that, while we may balk at the claim that the child begins to possess knowledge only at the time that he learns language (but not before), we are not similarly troubled by the claim that the adolescent begins to possess previously absent rights and responsibilities upon attaining adulthood. Rorty sees these claims as analogous: in neither situation has there been some essential shift from within the individual; rather there has been a shift in the individual's relations with others. What has changed is the social context. The question of the attribution, to other beings, of knowledge of what things are like has, for Rorty, more to do with morality than with justified true belief: [N]obody except philosophers of mind cares whether the raw feel of pain or redness is different for koalas than for us, but...we all care quite a bit about a koala when we see it writiiing about. This fact does not mean that either our or the koala's pain is "nothing but its behavior"; it just means that writhing is more important to our ability to imagine the koala asking us for help than what is going on inside the koala. Pigs rate much higher Ibid., pp. 184-5. 10 than koalas on intelligence tests, but pigs don't writhe in quite the right humanoid way, and the pig's face is the wrong shape for the facial expressions which go with ordinary conversation. So we send pigs to slaughter with equanimity, but form societies for the protection of koalas. 1 5 From this, we can see that our mora l sense is governed not by some sort of recognition of certain objective qualities i n beings, but instead on the imagined possibility of conversation w i t h i n a given speech-community. Our misguided attribution of feelings to other beings is nothing more than a reflection of this. The "inside" of l iv ing beings is thus to be explained by what goes on on the "outside", especially i n relation to their place i n the community . The interior state of the organism is merely something posited to explain its observed behaviour. In this light, it is not irrational to "send pigs to slaughter w i t h equanimity, but form societies for the protection of koalas". Since rationality derives from the dictates of the communi ty and its language, and not from anything "internal" to a being, rational allotment of mora l concern is based on our expected inclus ion of a particular being i n our communi ty . This means that other animals w h o most plausibly resemble humans i n some relevant way w i l l inspire greater mora l concern i n us than those w h o resemble us less convincingly. Rorty favours the rejection of the not ion that language is the expression of something "inner", w h i c h must be discovered prior to our being able to te l l what an utterance means. This entails the abandonment of concepts and meanings, inc luding such things as beliefs and desires. Their dispensability derives from the lack of behaviouristic equivalents for them. If we v iew concepts and meanings as a special source of truth or authority, they become harmful; but they assume a more beneficent aspect when s imply posited to explain our behaviour. This beneficence is the reward of our recognition that explanatory power rightfully follows pragmatic dictates. Philosophy then drops out as a discipline w h i c h guards against "irresponsible reification" and w h i c h systematises "our scruples about what objects one may assume". The different vocabularies w h i c h we use to talk about things (e.g., about actions and beliefs, or about movements and neurons) is not a mark of the difference between the real and the ontologically disreputable, or between the factual and the mythical . Al though these divergent vocabularies can be mixed i n our utterances (e.g., i n the sentence "If we had just stuck i n an electrode i n the right place i n the cortex, he w o u l d never have decided he was Napoleon"), we cannot hope to develop them into laws w h i c h are parts of comprehensive theories. Yet this is not because they belong to different ontological realms, such as the mental and the physical . Neither can we reduce one vocabulary to another, i n the interest of hi t t ing upon a vocabulary w h i c h reflects the true and ultimate structure of reality. A l l that is open to us is to judge the various vocabularies we have i n respect of their Ibid., p. 190. 11 pragmatic or aesthetic virtues. There is nothing to be gained, then, f rom differentially va lu ing explanations as either "scientific" or "unscient i f ic" . 1 6 So, to understand the superiority of the N e w Science to Aristotel ian investigations, we need to turn "outward", towards the social context of justif ication of the day, rather than seeking privi leged items i n consciousness w h i c h could serve as the touchstone for truth. The latter move is embodied i n the mistake o f the epistemic t radi t ion: confusing the causal process of acquir ing knowledge w i t h the justificatory process. If we turn away from this mistake, we w i l l f ind that there is nothing left for epistemology to be. W e re-describe " truth" i n terms of what it is good for us to believe, removing it as the gate-keeper of assertions i l lustrating genuine contact w i t h reality. This opens the w a y for a culture i n w h i c h there is no all-encompassing discipline w h i c h grounds and sets l imitations o n the claims of the others. Rorty believes that he finds support for this stance i n Kuhn's The Structure of Scientific Revolutions. In Rorty's pragmatist v is ion, objectivity ( w h i c h he characterises as h u m a n beings try ing to describe themselves i n relation to a non-human reality) is to be reduced to solidarity, an empathic recognition of community , i m b u e d w i t h reciprocal relations of responsibility. This shift obviates both metaphysics and epistemology. If we are determined to retain a not ion of "truth", then it must be relegated to the status of s imply what i t is good for us to believe. T o say that something is "true" is nothing more fancy than an expression of commendation, a n d this term w i l l have an identity of meaning i n a l l cultures (viz., that o n w h i c h intersubjective agreement has been achieved), but a diversity of denotation (i.e., different cultures w i l l come to different intersubjective agreements f rom those i n other cultures). To doubt the truth of what we believe at present is merely to expect that somebody may at some point come up w i t h a better idea, or description. O n questions where agreement has been found to be relatively easy to attain, we c la im knowledge; where agreement is more elusive, we can only c l a i m to be expressing opinions. This focus o n agreement renders truth and rationality dependent u p o n the famil iar procedures of justi f ication w h i c h a given society uses i n an area of enquiry. In evaluating questions w h i c h arise w i t h i n our culture, we must necessarily begin from what Rorty terms an "ethnocentric" position, where, i n practice, we privilege the beliefs and values of our o w n group, even though there can be no non-circular justif ication for d o i n g so. In both evaluating our o w n culture from w i t h i n , and i n testing our beliefs against those of another culture, a l l w e can do is play off some beliefs against others, i n the effort to generate more fruitful conversation, w h i c h is to say, conversation w h i c h aims towards a more successfully pragmatic picture. This brand of ethnocentrism is provis ional: it provides a natural starting-point~the only one available to u s - f r o m w h i c h to expand our conversational efforts outwards, towards other cultures, casting the net of solidarity ever wider . As such, it is preferable to the unrecognised, or at least unacknowledged, ethnocentrism of the tradit ional philosopher, w h o Ibid., pp. 206, 208. 12 attempts to take refuge i n buttressing his o w n perspective w i t h fixed notions of correspondence to reality, or of a peculiarly human mora l sense or set of cognitive faculties. One of Rorty's chief aims is to suggest the establishment of a l iberal Utopia, characterised by a universal incl inat ion towards "liberal irony". W e are called upon to recognise that cruelty is the worst th ing that we can do, and to face up to the contingency of even our most central beliefs and desires (thereby demonstrating "irony"). There is nothing i n this v is ion w h i c h depends upon the recognition of genuine facts, to be gleaned from a process of proper enquiry. Instead, it speaks of human solidarity, w h i c h is fostered by imaginat ion and empathy. Creation, and not discovery, is therefore the mode w h i c h we are t rying to get at i n this society. What this amounts to is that the final victory of poetry over philosophy w i l l occur as the t r iumph of metaphors of self-creation over those of discovery. This focus on self-creation~and i n turn, the creation of the culture by self-creating individuals-emphasises freedom as the goal of thought, at the same t ime that it banishes truth. Self-creating individuals are not merely to make do w i t h an inherited w o r l d . The a i m is to achieve distance from the desire to br ing particulars under general principles, and the attempt to f ind necessary truths. W e need only judge "particular present situations and options as s imilar to or different from particular past actions or events". Any th ing w h i c h we encounter i n l i fe -" f rom the sound of a w o r d through the color of a leaf to the feel of a piece of skin"~can "dramatize and crystallize a human being's sense of self-identity". A n y such thing can play the role i n ind iv idua l life w h i c h philosophers have thought c o u l d - a n d shou ld -on ly be played by things w h i c h were universal. But again, solidarity w i l l require that reflective human beings give sense to their freely created indiv idual lives by placing themselves into a larger context. They do so by te l l ing the story of their contr ibution to their c o m m u n i t y . 1 7 After a l l of this, i f we sti l l insist on the relevance of the not ion of internal representations, then we can only do so by restricting ourselves to something on the order of " 'innatist' views common to Chomsky and Fodor" 1 8 , where we encounter the not ion of a "wired-in" language and meta-language of thought: If one gives up the notion that empirical psychology is going to do what the British Empiricists failed to do-show how a tabula rasa gets changed into a complicated information-processing device by impacts upon peripheral sense-organs-then one wi l l not be surprised that half of the adult's subroutines were wired into the infant's brain on instructions from the chromosomes. Further, it wi l l not strike one as important to our understanding of the nature of man or his mind to discover just which were wired into them and which came along later. 1 9 Rorty, CIS, p. 37. Rorty, PMN, p. 251. Ibid., p. 241. 13 Knowledge of how the human mind works is consequently put on a par with knowledge of how glands or molecules work. We need invoke no explanations to be uniquely applied to human beings, and not to things in the world. To grasp this sort of meta-linguistic vocabulary is not to grasp something general, but rather something particular. We are not engaged in the business of setting out the terms under which we can recognise someone to be a rational enquirer. Judgments of rationality are not to be related to elements in the mind, but rather to the ways in which means are adjusted to ends, as well as to the particular ends which are sought. This is to say that rationality attaches to particular behaviours, rather than to a general mental structure. So any indulgence in the premise that there is a fixed language of thought should not be allowed to slide into the premise that our knowledge of the nature of that language is itself immune to correction on the basis of experience. On the assumption that there are a priori constraints in the form of inner, mental truths, we are led into the notion of foundations of knowledge and theories of representation, with all the attendant problems described thus far. On the other hand, Rorty quotes Fodor as claiming "that the discovery of the language of thought will be a long-drawn-out empirical process", and that that "has a corollary that we may always be quite wrong about what this language is, and thus wrong about what is a priori". Because of the uncertainty into which this places the nature of the a priori, and because its empirical character puts it on a par with the sorts of investigations we find in other disciplines, philosophy is ousted from its position of aggrandisement as the tribunal of reason which sits in judgment upon the activities of other human endeavours. "Accurate representation" should come to be seen as simply a belief which helps us to do what we want to do. Grounding "rationality" and "objectivity" in traditional notions of accurate representation "is a self-deceptive effort to eternalise the normal discourse of the day", when all that discourse truly embodies is agreed-upon criteria for reaching further agreement.20 To sum up, Rorty's is an extended attempt to "deconstruct the image of the Mirror of Nature". It is essential to this effort that one see the human self as nothing more than a continually re-woven web of beliefs. There is nothing prior to socialisation or history which can serve to define the human. The desire for a permanent, neutral, ahistorical vocabulary is itself a historical phenomenon. In making this historicist turn, we resist the temptation to seek an escape from time and chance. The world and the self must become de-divinised, so that we can see both in terms of "an expanding repertoire of alternative descriptions rather than The One Right Description". We no longer think of the world or the self as speaking to us. Epistemology is to be replaced by a new way of looking at things, which will count as "an expression of hope that the cultural space left by the demise of epistemology will not be filled-that our culture should become one in which the demand for constraint and confrontation is no longer felt". This new hope is embodied in hermeneutics, and Rorty cautions us that we are not to view it Ibid., pp. 11, 252. 14 as a discipline, a programme of research, or a method of achieving the same sort of results w h i c h epistemology failed to achieve . 2 1 I w i l l return to this question of hermeneutics i n Chapter Three. I w i s h n o w to turn to specific references w h i c h Rorty makes to Kuhn's work, i n order to demonstrate Rorty's belief i n the general s imilari ty of Kuhn's programme to his o w n . But intel l igibi l i ty of this account w i l l be facilitated i f I first give a few quite general statements elucidating Kuhn's programme for understanding the history and philosophy of science. The standardly accepted model of science paints a picture of a fairly steady accumulat ion of ind iv idual discoveries and inventions, w h i c h accumulat ion constitutes the modern body of scientific knowledge. Periods w h e n science does operate i n this wa y are described, i n Kuhnian terms, as exemplifying "normal science". Kuhn's contention is, however, that, i f we take seriously an historical v iewpoint of science, we see, rather than this cumulative picture, one i n w h i c h not only scientific practice, but the entire way of envisioning the scientific field concerned^ has undergone a series of variably radical transformations. As a result of such transformations, not only new answers to problems are generated, but the problems themselves are not the same after the shift as they were before. The periods of upheaval between the dissolution of the o ld scientific order (for reasons I w i l l explain i n the next chapter) and the functional instantiation of the new one are characteristically conceptually and methodological ly unusual periods, the work done i n such a climate being refferred to as "revolutionary science". These are times w h e n the scientific v is ion w h i c h guides research is i n transition, and hence somewhat unclear. This situation reveals a need to pu l l back from conceptions of. scientific t ruth and knowledge w h i c h entail the correspondence of belief to some discoverable absolute state of reality. Kuhn's programme is intended to describe, i n the absence of such conceptions, how we might make sense of scientific activity. Rorty states that "Wittgenstein, Heidegger, and Dewey have brought us into a period of ' revolut ionary philosophy ( in the sense of Kuhn's 'revolutionary' science) by introducing new maps of the terrain (viz. , of the whole panorama of human activities) w h i c h s imply do not include those features w h i c h previously seemed to dominate." In the interest of promoting this "revolutionary" outlook, his Philosophy and the Mirror of Nature "is a survey of some recent developments i n philosophy, especially analytic philosophy, from that point of v i ew of the anti-Cartesian and anti-Kantian revolut ion" described at the beginning of the introduct ion to the book. This implies that Rorty views his project as w o r k i n g to achieve much the same for phi losophy as Kuhn's project d i d for science. Indeed, Rorty goes on to say that "most of the particular criticisms of the t radi t ion w h i c h [he] offer [s] are borrowed from such systematic philosophers as Sellars, Quine, Davidson, Ryle , M a l c o l m , Kuhn , and Putnam." Further, Rorty "hope[s] to Rorty, PMN, pp. 315, 390; CIS, pp. 39-40. 15 convince the reader that the dialectic w i t h i n analytic philosophy...has carried.. .philosophy of science from Carnap to Kuhn" . 2 The basis of what Rorty takes to be his c o m m o n ground w i t h K u h n is the shift w h i c h he is advocat ing-and w h i c h I have out l ined earlier i n this chapter -from the more personally based abil i ty to have knowledge w h i c h we see i n tradit ional philosophy (whether it be grounded i n the having of clear and distinct ideas or the having of raw feels), to knowledge based i n the social sanctioning of belief. Hence, "to understand...the superiority of the N e w Science to Aristotle, the relations between this science and mathematics, common sense, theology, and moral i ty~we need to turn outward, toward the social context of justification rather than to the relations between inner representations." It is significant that Rorty notes that "this attitude has been encouraged" by, among other things, 23 "Kuhn's Structure of Scientific Revolutions". As evidence of this, Rorty says that so-called "observation", according to Kuhn , is "just a matter of what we can agree on these days" . 2 4 It appears to me that he reads K u h n as consequently preferring to drop the notion of observation altogether 2 5 , a move wh ich w o u l d appeal to Rorty. In taking this stance, K u h n has helped to "exhibit the sterility of attempts to give sense to phrases like 'the way the w o r l d is', or 'fitting the facts' " 2 6 . Rorty remarks that there has been a temptation i n tradit ional phi losophy to see a new problematic as "the o ld one r ightly seen" 2 7 , w h i c h is to say that we get, through our contemplations, a progressively clearer picture of the o ld problematic, and gradually come to terms w i t h it better; on this view, we find new ways to deal w i t h o ld problems, and thus become increasingly adept at solving them. Rorty looks to Kuhn ian history and philosophy of science as a parallel case w h i c h serves to illustrate the inaccuracy of this tradit ional general philosophical v iew. In his o w n field, K u h n has helped to dispel "the not ion of increasingly accurate representations of nature being found i n its mi r ro r " 2 8 , v i z . the m i n d . He has done so by essentially demonstrating, i n his "cri t icism of the 'textbook' approach to the history of inquiry", that the above "temptation should be resisted". Importing this Kuhn ian orientation into his discussion of philosophy i n general, Rorty advises that, instead of the tradit ional, cumulative understanding of philosophy, we have "new philosophical paradigms nudging o ld paradigms aside", such that "a new set of problems emerges and the o ld ones begin to fade away" 2 9 . So Kuhn ian "revolutionary science" provides a model , for Rorty, for the f luid discourses w h i c h shape the sort of culture w h i c h he has advocated i n his PMN, pp. 6-7. Ibid., p. 210. Ibid., p. 227. Ibid. p. 225. CIS, p. 20. PMN, p. 264. Ibid., p. 276. Ibid., p. 264. 16 o w n ph i l o sophy . Hence , the R o r t i a n d i s t inc t ion b e t w e e n " n o r m a l " a n d " a b n o r m a l " d i scourse is to be t aken as genera l i s ing Kuhn ' s d i s t inc t ion b e t w e e n " n o r m a l " a n d " r evo lu t i ona ry " sc ience. R o r t y descr ibes this para l le l thus : i n " the p ragmat i c a p p r o a c h to k n o w l e d g e suggested b y ep i s temo log i ca l b e h a v i o u r i s m " 3 , n o r m a l d i scourse is that w h i c h c a n be r e n d e r e d c o m m e n s u r a b l e (i.e., " ab le to be b rough t u n d e r a set o f ru les w h i c h w i l l tel l us h o w ra t iona l a g reement c a n be r e a c h e d o n w h a t w o u l d settle the issue o n every po in t w h e r e statements s eem to c o n f l i c t " 3 1 ) . In contrast, a b n o r m a l d i scourse is that w h i c h c a n n o t be r e n d e r e d c o m m e n s u r a b l e (i.e., it lacks ru les t h r o u g h w h i c h ra t iona l ag reement m i gh t be reached ) . F o r Rorty , the ques t ion o f the presence o r absence o f ru les g u i d i n g ra t iona l a g reement is not at a l l a c ruc i a l one for a cu l ture . R o r t y observes that K u h n has a distrust o f " t he posit iv ist ic i d e a that ra t iona l i ty is a mat te r o f a p p l y i n g c r i t e r i a " 3 2 . Thu s , K u h n "g ives us reason to say that there is n o deeper d i f f e rence " b e t w e e n the "patterns o f a r g u m e n t a t i o n " o f sc ience ("as the d i s covery o f w h a t is rea l l y out there i n the w o r l d " ) a n d those o f "d i scourses for w h i c h the n o t i o n o f ' co r re spondence to r e a l i t y seems less appos i te (e.g., pol i t ics a n d l i terary c r i t i c i sm) " t h a n " that b e t w e e n w h a t happens i n ' n o r m a l ' a n d i n ' abno rma l ' d i s c o u r s e . 3 3 In o ther words , the patterns o f a r g u m e n t a t i o n o f sc ience are not s ign i f icant ly d i f ferent f r o m those o f a n y o ther a rea o f the cu l ture ; ne i ther n e e d re ly u p o n rules de s i gned to gu ide ra t iona l ag reement . W e see this re f lected i n the f o l l o w i n g passage i n Rorty : As Kuhn argues in The Copernican Revolution, we did not decide on the basis of some telescopic observations, or on the basis of anything else, that the earth was not the center of the universe, that macroscopic behaviour could be explained on the basis of microstructural motion, and that prediction and control should be the principal aim of scientific theorizing. Rather, after a hundred years of inconclusive muddle, the Europeans found themselves speaking in a way which took these interlocking theses for granted. Cultural change of this magnitude does not result from applying criteria (or from "arbitrary decision") any more than individuals become theists or atheists, or shift from one spouse or circle of friends to another, as a result either of applying criteria or of actes gratuits. We should not look within ourselves for criteria of decision in such matters any more than we should look to the world.34 T h e s imi lar i ty b e t w e e n the activit ies o f sc ience a n d those o f o ther cu l tu ra l endeavour s is e c h o e d i n the s imi lar i ty b e t w e e n p h i l o s o p h y a n d o ther cu l tu ra l endeavour s . Bu t this latter u n d e r s t a n d i n g runs against t rad i t i ona l p h i l o s o p h y s v i e w o f itself as u n i q u e l y p r o v i d i n g foundat ions for know ledge , v i a the m ind ' s capac i ty to m i r r o r Na ture . In the tradi t ional i s t c l imate, " the expl ic i t h o p e o f p re -Ibid., p. 320. Ibid., p. 316. ORT, p. 25. PMN, pp. 332-3. CIS, p. 6. 17 Kuhnian philosophers of science was to have an account of 'the nature of science' w h i c h no future scientific revolut ion could dis turb" . 3 5 The Rort ian understanding of hermeneutics (which I w i l l flesh out i n Chapter Three) is, roughly speaking, the relinquishment of the desire for an epistemology: the desire to unite people under a common rationality, the desire that there be "a special set of terms into w h i c h a l l contributions to the conversation should be put" . 3 6 Rather, hermeneutics unites people through c iv i l i ty and the hope of agreement; other than this, there is neither common goal nor common ground. Rorty's promot ion of what he terms hermeneutics is another facet of his philosophy i n w h i c h he sees himself as indebted to the work of K u h n : In recent years, debates about the possibility of epistemology as opposed to hermeneutics have gained a new concreteness as a result of the work of T.S. Kuhn....Since the Enlightenment, and in particular since Kant, the physical sciences had been viewed as a paradigm of knowledge, to which the rest of culture had to measure up. Kuhn's lessons from the history of science suggested that controversy within the physical sciences was rather more like ordinary conversation (on the blameworthiness of an action, the qualifications of an officeseeker, the value of a poem, the desirability of legislation) than the Enlightenment had suggested. In particular, Kuhn questioned whether philosophy of science could construct an algorithm for choice among scientific theories. Doubt on this point made his readers doubly doubtful on the question of whether epistemology could, starting from science, work its way outward to the rest of culture by discovering the common ground of as much of human discourse as could be thought of as "cognitive" or "rational". 3 7 Rorty goes on to say that Kuhn's examples of revolutionary change i n science are "cases of the sort w h i c h hermeneutics has always taken as its special assignment". Hence, " K u h n wished to oppose the tradit ional c l a im that 'what changes w i t h a paradigm is only the scientist's interpretation of observations that themselves are fixed once and for a l l by the nature of the environment and of the perceptual apparatus' ". Kuhn's embracing of the hermeneutic method is disturbing to those philosophers of science, oriented towards the epistemological tradit ion, who were looking for a "neutral scheme", or neutral observation-language. They saw K u h n as "endanger [ing] the not ion of theory-choice i n science", since philosophers of science had "envisaged [themselves] as providing an algori thm for theory-choice". But "Kuhn was right i n saying that 'a philosophical paradigm init iated by Descartes and developed at the same time as Newtonian dynamics ' needed to be over thrown" . 3 8 Rorty describes both himself and K u h n as sharing the v iew of "pragmatism": an anti-realist posit ion w h i c h "gives up the attempt at a God's-eye v i ew of things, the attempt at contact w i t h the nonhuman w h i c h [Rorty] has been Ibid., p. 25n. PMN, p. 318. Ibid., p. 322. Ibid., pp. 323-5. 18 cal l ing 'the desire for object ivi ty" ' . He takes it that K u h n w o u l d agree (as Rorty, himself, does) w i t h Putnam's admoni t ion that we "drop the not ion of a God's-eye point of v i e w " . 4 0 He describes K u h n as espousing a " 'pragmatist' h o l i s m " 4 1 , explaining that "[w]hat [he is] ca l l ing 'pragmatism' might also be called ' left-wing Kuhnian ism' " 4 2 He states that the "general pragmatist c l a im" is that "there is no permanent ahistorical metaphysical framework into w h i c h everything can be fitted". A Kuhn ian corollary of this Rort ian thesis is that "there is no single commensurat ing language, k n o w n i n advance, w h i c h w i l l provide an i d i o m into w h i c h to translate any new theory, poetic id iom, or native cul ture" . 4 3 Referring to his Philosophy and the Mirror of Nature, Rorty observes that "Dewey, Wittgenstein, Sellars, Kuhn , and the other heroes of this book a l l have their o w n ways of debunking 'truthfulness to reality i n the sense postulated by philosophical realism' " , 4 4 Rorty observes that Kuhn's "pragmatist friends (such as [himself]) routinely congratulate h i m on having softened the dist inct ion between science and nonscience". 4 5 In the t ime "[bjefore the arrival of Kuhn , Tou lmin , Feyerabend, and Hanson, it was often thought that the physical sciences were. . .paradigmatically rational areas of culture". In the "pre-Kuhnian philosophy of science...rational enquiry was a matter of putt ing everything into a single, w ide ly available, familiar context-translat ing everything into the vocabulary provided by a set of sentences w h i c h any rational inquirer w o u l d agree to be truth-value candidates". 4 6 I w i l l move now to a detailed exegetical examinat ion of the Kuhn ian programme, taking as its source writ ings supplied by K u h n , himself. This w i l l prepare us for the analysis, i n the th i rd chapter, of h o w w e l l Rorty's expectations of how Kuhn's work might aid his o w n project measure up against the ways i n w h i c h K u h n describes his o w n work . ORT, p. 24. Ibid., p. 27. Ibid., p. 65. Ibid:, p. 38. Ibid., p. 2l5. PMN, p. 382. ORT, p. 38. Ibid., p. 95. 19 2 Thomas Kuhn: Understanding Science Introduction In his groundbreaking Structure of Scientific Revolutions, Thomas Kuhn was concerned to replace what he saw to be a misguided conception of the scientific enterprise with one he believed more accurately to describe the way in which science actually proceeds. According to the standardly accepted model, science develops and progresses through a process of accumulation of individual discoveries and inventions. This accumulation constitutes the modern body of scientific knowledge. The history of science is thus presented as an account of "at what point in time each contemporary scientific fact, law, and theory was discovered or invented", and by whom. Kuhn contends, however, that the "[c]umulative acquisition of unanticipated novelties proves to be an almost non-existent exception to the rule of scientific development." If we take the historical viewpoint seriously, we see "that science does not tend toward the ideal that our image of its cumulativeness has suggested." In fact, "most new discoveries and theories in the sciences are not merely additions to the existing stockpile of scientific knowledge."1 Kuhn was alerted to the inadequacy of the standard view when it struck him that, in a typical contemporary study of such theories as, for example, Aristotelian dynamics, phlogistic chemistry, or caloric thermodynamics-all now considered to be out of date-these discarded theories would be deemed incorrect or even unscientific, i.e., not yet science. But to do so is to ignore the presence and importance of the actual historical development of science. The mere fact of its having been replaced by a newer theory is insufficient evidence by which to hold an older theory to be unscientific. The key to understanding science lies not in our seeking to determine the contributions of an older science to our present science, but rather in our attempting to retain the integrity of that older science in respect of its own time. This means that a particular science of any age ought to be studied from a viewpoint wherein it possesses "the maximum internal coherence and the closest possible fit to nature." Such a historical view entails the study of particular theories within the context in which they achieved prominence.2 1 Kuhn, SSR, pp. 2, 96, 140; ET, p. 226. 2 SSR, pp. 2-3. 20 That out-of-date theories cannot be accurately deemed simply unscientific points to the lack of a set of overarching conceptual and practical directives which might be used to dictate the conclusion to any scientific question. Instead, the framework of conceptualisation and practice has undergone a series of transformations throughout scientific history. A researcher at any point in time brings a certain scientific background to her investigations. Further, scientific conclusions are influenced by such factors as accidents and the individual make-up of the researcher. These are all temporal elements. So there is no neatly packaged knowledge of "what it is to be scientific" which can be applied to scientific problems. If there were, there would be neither need nor place for a specifically historical study of science. Nevertheless, out-of-date theories can be recognised as having been produced by a roughly similar sort of activity to, and for the same sorts of reasons as, current scientific knowledge and research. The difficulty of the cumulativist view of science is underscored when we look more closely at the two possible ways in which we can consider out-of-date theories: as myths or as science. If they are to be relegated to the mythical, then their rough similarity to current science is puzzling. If we then decide that they are indeed to count as science, then we must conclude that science has, in the past, included bodies of belief which are strikingly incompatible with present beliefs. That there is some contextual set of received beliefs which the researcher brings to his study of scientific problems indicates that, while observation and experience~the foundations of the empiricist view of investigation-must have a hand in directing the range of admissible scientific belief, they do not alone determine the body of that belief. Instead, "research [is] firmly based upon one or more past scientific achievements, achievements that some particular scientific community acknowledges for a time as supplying the foundation for its further practice". This "tradition-bound" conduct is the essence of science as it is generally practiced, i.e., during times of scientific stability, and it is what Kuhn has labelled "normal science."3 Normal, or Ordinary, Science The most obvious feature of Kuhnian normal science (as it is described in his earlier writings) is the presence of a "paradigm", which Kuhn defines generally as a situation in which accepted examples of scientific practice (including law, theory, and instrumentation) provide a model leading to a coherent tradition of scientific research. The paradigmatic nature of what is today familiar to us as science can be seen in sharper focus when we consider what Kuhn (in earlier writings) referred to as "pre-paradigm science". This is exemplified by the state of physical optics before the time of Newton. In these early stages of a science's 3 Ibid., pp. 4, 6, 10. 21 development, one sees at most the existence of various "schools". Since, for K u h n , the interpretation of data is only possible where there is at least some foundation of theoretical and methodological belief, each of these schools w o u l d have to possess its o w n foundation, and it was i n the context of such a matrical foundation that theory-selection, -evaluation, and -cri t icism w o u l d go on . Because such an underlying body of belief is thus impl ic i t i n the collection of facts, more than "mere facts" are at hand. Al though there was, for each school, a grounding of agreed-upon scientific practice and vis ion w i t h i n w h i c h work was carried on, "fact-gathering" i n pre-paradigm science was st i l l a more random activity than it has been i n paradigm-based science (because of the plural i ty of research matrices i n the former). Since there was no single standard conceptual and methodological corpus adopted by the scientific communi ty either as a whole or i n a particular area of study at any given time, a scientist's choices i n terms of experimentation and observation were made relatively freely. In the absence of a universally accepted paradigm (at either the whole scientific or the discipl inary level), a l l of the facts that could possibly pertain to the development of a given science could seem equally relevant. 4 In the history of science, however, this in i t ia l divergence i n schools of thought was eventually to disappear i n the various scientific disciplines, as a result of the t r iumph, i n each discipline, of one of the pre-paradigm schools. In order to accomplish this, the body of theory and practice of the successful school had to become accepted as a paradigm w h i c h was to guide the work of a l l researchers i n a given field. It had therefore to seem better than its competitors, al though it need not explain all the facts it was to encounter. The acquisit ion of a paradigm is a sign of maturi ty i n the development of a particular scientific field; it is an indicat ion of a new level of confidence w h i c h encourages researchers to undertake more detailed, precise, and esoteric work . This is because the constant re-examination of fundamentals w h i c h dogged pre-paradigm science is obviated by the more t ightly directed fact-collection and theory-articulation demonstrated i n paradigm-based science; energy previously directed towards the continual setting of foundations could be shifted into increased technical efficiency and more specialised problems. Scientific work could afford to become more esoteric and oriented towards puzzle-solving once its foundations could be taken for granted by experimenters. This signalled a new attitude wherein current theory came to be more exploited, and less cri t icised. The transit ion to a mature science is marked by the diminishment of cr i t ical discourse. Hence, as each discipline, or area of research, developed sufficiently, non-dominant schools gradually disappeared, through the conversion of seasoned practitioners, and the attraction of most of the next generation's practitioners, to the newly ascendant paradigm, as w e l l as through those schools being ignored by the new scientific mainstream. Paradigmatic dominance is exemplified by Ptolemaic or Copernican astronomy, Aristotel ian or Newtonian dynamics, and corpuscular or wave optics. 4 Ibid., pp. 10, 13, 15-17. 22 One of the basic characteristics of a paradigm, then, is that its achievements are sufficiently unprecedented so as to attract adherents away from competing models. But this novelty cannot be a cont inuing focus of researchers i f the paradigm is to fulfil its essential role i n the developmental progress of the field. K u h n stresses that normal science does not a i m to produce unexpected novelties, and scientists normal ly do not a im to invent new theories. W h e n normal science is successful, novelties of fact or theory are not discovered; i n fact, i f any such novelties threaten to disturb paradigm stability, they are typically suppressed. The congealment of scientific attention into a successful paradigm is a result of that paradigm's abil i ty to achieve anticipated research results i n a new way, a way w h i c h constitutes an improvement over the collective abilities of the various schools of the pre-paradigm period, or, as we shall see later, over the abilities of the previously enshrined paradigm ( in the case of an already-matured discipline). As I have mentioned, it is i n Kuhn's earlier wri t ings that he refers to the science typified by the various schools of thought extant prior to the coming into dominance of a single discipl inary matr ix as "pre-paradigm" science. Somewhat later, he sought to qualify this by observing that discipl inary matrices are indeed possessed by any scientific communi ty , inc luding these schools of the so-called pre-paradigm period: "a rigorous t raining i n convergent thought has been intrinsic to the sciences almost from their or igin ." Paradigm-function i n relation to research-problems is a crucial feature of science. Scientists are fundamentally "puzzle"-solvers, the dominant cri terion for most of them being the abil i ty to recognise and solve puzzles posed by Nature. Were it not for the stabilising effect of the paradigm, the requisite security for the scientist's tackl ing problems, especially those of a more esoteric type, w o u l d be lacking. W h e n engaged i n a normal research problem, the scientist must premise current theory as the "rules of his game". The work of science is thus composed of these instrumental, conceptual, and mathematical puzzles, to w h i c h it must always be expected that there w i l l eventually be a solution, else commitment to the scientific enterprise w o u l d wane. The corpus of accepted current theory bestows meaning on the problems of normal research, lending support to the assurance that a solution to the puzzle w i l l be found. Problems w h i c h appear potentially insoluble are therefore not considered to be scientific problems: they may be rejected as metaphysical, as part of another discipline, or as being too problematic to be wor th the effort. 6 The paradigm, then, functions to provide criteria for choosing problems that can reasonably be assumed to have solutions. W i t h i n it, the range of anticipated and assimilable results is narrowed from the much larger field supplied by the experimenter's imaginat ion. It serves to l imi t the nature of acceptable solutions and the steps by w h i c h they might be obtained. Research-outcomes w h i c h fail to fall w i t h i n this normal range w i l l , under the conditions of 5 ET, pp. 228, 295. 5 SSR, pp. 35-7, 205; CGK, p. 4. 23 normal science, be deemed research-failures. So, what is being tested i n normal research is the ind iv idua l scientist's abil i ty to solve his puzzle; what is not being tested is the corpus of current science, w h i c h is instead required to form the background against w h i c h puzzle-solving is done. N o r m a l science thus "seems an attempt to force nature into the pre-formed and relatively inflexible box that the paradigm supplies", and new sorts of phenomena w h i c h do not fit into this box are often not seen at a l l . Paradigm procedures and applications, laws and theories serve to "restrict the phenomenological field accessible for scientific investigation at any given time", thus a l lowing the scientific endeavour to have a conceivable method of research. Nevertheless, the body of admissible procedure cannot be over-specified: the paradigm must remain sufficiently open-ended so as to leave problems for it practitioners to resolve. But behind the issue of problem-choice, the strictures of the paradigm provide guidance i n respect of favoured conceptual, theoretical, and methodological commitments, as w e l l as of preferred types of instrumentation and the ways i n w h i c h accepted instruments may legitimately be employed; it may be necessary, for instance, that an experimental apparatus be redesigned i n order that it produce results of a better fit w i t h accepted theory. 7 The ini t ia l success of an adopted paradigm, however, is generally neither complete nor widespread. It is instead largely a promise of success i n some areas. The settled regime of normal science is typified by the actualisation of a significant amount of what has been promised by the new paradigm. In this way, the paradigm becomes more accepted, but as it does so, it also becomes more elaborately articulated. Scientific problems are increasingly model led upon previous achievements. M u c h scientific research w i t h i n a t radi t ion is an attempt to adjust existing theory and observation into progressively closer agreement. It also involves extending existing theory into areas w h i c h it is expected to cover, but to w h i c h it has not yet been applied. Some research w i l l thus be directed towards f inding new applications of the paradigm, or towards increasing the precision of applications that have already been introduced. In this process of elaboration, a paradigm may be somewhat reformulated. The significant point here is that none of these problems or adjustments w o u l d be considered wor th undertaking by anyone who had basic doubts about the val idi ty of existing theory, or who leaned towards a new theory. As the discipline matures through the further art iculat ion and success of the adopted paradigm, we see increasing professional!sation and specialisation, accompanied by a growing resistance to paradigm-change. Moreover, the development of an in i t ia l paradigm spells the end of any research done outside of a paradigm. A l l future research w i l l be contained w i t h i n one such matr ix or another. K u h n observes that the sciences acquired something l ike paradigms "at precisely the point when the field began to make rapid and systematic progress". 8 It is crucial , i n understanding Kuhn , to recognise that he regards science as an importantly communal enterprise. The paradigm achieves its success by virtue 7 SSR, pp. 10, 24, 35, 38-40, 42, 60; CGK, p. 5. 8 SSR, pp. 64, 79; ET, p. 230. 24 of its acceptance by the relevant scientific communi ty as a whole , the group of practitioners of a scientific specialty. Indeed, the paradigm may be said to govern not a subject-matter, but rather a group of practitioners. Paradigms are a demonstration of consensus. Scientific communities, of course, exist at different levels. W i t h i n the overarching group of natural scientists, we can identify the divergent specialisations represented by astronomers, physicists, chemists, and zoologists, for example. At a level of further professional refinement, we can distinguish, say, solid-state from high-energy physicists, and organic from inorganic chemists. Yet further split t ing up of scientific domains occurs as disciplines mature and their paradigms become ever more esoteric and narrowly focused. The ma in body of each group, at whatever level, is united around the subject of its members' highest educational degree, membership i n professional societies, and publications i n journals. Consequently, paradigms w i l l also exist at different levels. Some w i l l guide the work of a broader group of researchers; others w i l l be of relatively more narrow scope. For instance, w i t h i n the large and diverse communi ty of physical scientists, a l l w i l l have learned the laws of quantum mechanics, but al though they share the basic matrix, they do not a l l learn the same applications of these laws. Solid-state physicists are not w o r k i n g w i t h i n quite the same paradigm as are chemists. Beyond this, even w i t h i n the confines of the more precisely specified paradigm, two indiv idual researchers may differ i n the judgments w h i c h they d raw from it . Scientists can agree on their identification of the paradigm i n general without agreeing on their interpretation of it. Nevertheless, this disparity does not interfere w i t h the paradigm's abil i ty to guide research: its guidance w i l l just be manifested i n slightly different ways. In order to look w i t h greater resolution at the paradigmatic framework of conceptualisation and practice i n the sciences, we must turn to Kuhn's later revisions of his o w n terminology, init iated i n the wake of criticisms that his not ion of a paradigm, as presented i n The Structure of Scientific Revolutions, was not sufficiently clear. Kuhn's later programme refines this not ion by substituting the terms "disciplinary matrix" and "exemplars" for the term "paradigm". The discipl inary matr ix of a professional discipline is composed of three basic elements: 1. symbolic generalisations: the formal (or readily formalisable) components of the matrix; 2 . models: components w h i c h provide the permissible and preferred analogies and metaphors, or even an ontology; 3. exemplars: concrete problem-solutions; the professional community 's standard examples, to w h i c h students are heavily exposed i n their t ra ining ( in the form of textbooks, lectures, and laboratory exercises); and 4. values: e.g., theories are expected to be accurate, precise, simple, plausible, and compatible w i t h other current theories, as w e l l as being self-consistent; they must permit puzzle-formation and -solution, and are ideal ly relatively 25 more quantitative than qualitative; unanimity w i t h i n the communi ty is a paramount value; the scope of a theory is also to be considered. Symbolic generalisations a l low the application of logic and mathematics to problems. A n example w o u l d be the formal equation / = ma, or the verbal expressions "action equals reaction" and "elements combine i n constant proport ion by weight". Less specifically interpreted, these belong to the broader scientific community . But different subgroups of that communi ty w i l l more specifically interpret and apply them i n different particular ways. Models wh ich express analogies are of clearly heuristic usage, such as when one describes a gas as behaving l ike b i l l ia rd balls i n random mot ion . A n example of a model intended to provide an ontology is found i n the explanation of perceptible phenomena as being due to the mot ion and interaction of atoms i n the vo id . Exemplars provide a demonstration of and practice i n problem-solving wh ich help the student set the Gestalt w h i c h is typical of the discipl inary matrix. They a l low the student to assimilate "a time-tested and group-licensed way of seeing" 9 . The student "is learning the language o f a theory and acquir ing the knowledge of nature embedded i n that language" 1 0 . This w i l l prove to be essential t ra ining for the prospective scientist, since researchers solve puzzles by model l ing them on previous puzzle-solutions. Scientists "never learn concepts, laws, and theories i n the abstract and by themselves...[i]nstead, these intellectual tools are from the start encountered i n a historically and pedagogically prior unit that displays them w i t h and through their appl icat ions" 1 1 . Thus, we see that the cognitive content o f science is not strictly embedded i n theory and rules or criteria, w i t h exemplary problem-sets serving as a means through w h i c h students can merely gain facility i n the application of the theory. A great deal of this content is also carried i n exemplary problem-sets. Values tend to be more wide ly shared among various scientific communities than are symbolic generalisations and models. Al though they may be shared generally throughout a given scientific community, they may not be applied i n the same way w i t h i n that communi ty . K u h n insists that, rather than encountering the w o r l d i n simplistic empiricist fashion, equipped only w i t h a fresh tabula rasa, the science student is instead "programmed", i n her training, to recognise what her prospective communi ty already knows. She learns that the meaning of her community 's endeavours inheres i n its members' abil i ty to apply labels unequivocally and draw conclusions as to what might be further expected. The matrix w h i c h is learned i n scientific education is then reinforced by subsequent life i n the profession. 1 2 W e might be tempted to characterise this structure as a collection of four fundamental types of rule-based activities. O n this v iew, symbolic generalisations 9 SSR, p. 189. 10 CGK, p. 272. 11 SSR, p. 46. 1 2 Ibid., pp. 43, 46, 182-90; CGK, pp. 21, 241, 272; ET, pp. 227, 297, 312. 26 w o u l d provide one set of rules, models another, and so on . But K u h n contends that the existence of a discipl inary matr ix need not imply the presence of rules. Instead, scientists possess something more l ike an intuit ive knowledge of what w i l l count as research w h i c h fits the discipl inary matrix. Indeed, i f the philosophers whose interest it is to pursue the abstract structure of science ask scientists to come up w i t h the "correspondence rules" w h i c h they use i n their business of problem-solving (a l lowing them to attach symbolic representations to Nature), scientists often deny the relevance of such rules, and i f they do not, the "rules" that they do provide may w e l l vary from one ind iv idua l to another, and may, i n any case, be defective. This is not to say that there is no place for rules, or criteria, i n the explanation of h o w science works . It is just to say that, i f we are able to identify such criteria, it is because they have been abstracted from an already functioning discipl inary matrix, rather than being the prior determinants of that matrix. In Kuhn's estimation, we have no access to elements more m i n i m a l than sense-data reports, such as might be exemplified by "green there". In some sense, we can v iew these as "the given"; i n this respect, they are construed experientially. But understood theoretically, it is rather s t imuli w h i c h rank as given (although we have access to these only indirectly, v i a scientific theory), since we can consider sensations to be the result of a vast amount of neural processing of s t i m u l i . 1 3 H o w does this experiential fundamentality of sensations relate to science as non-criterial? K u h n offers an analogy i n w h i c h one is teaching a ch i ld to discriminate among ducks, swans, and geese. The ch i ld w i l l come to be able to pick out swans as a l l s imilar to each other, for instance, but may w e l l be unable to tell y o u what a swan is, w h i c h is to say that she w i l l be unable to specify criteria for its identification and discr iminat ion from ducks and geese. Her learning has been pr imar i ly by ostension. Similar ly , learning by ostension is essential to the reconstruction of scientific knowledge: i n both cases, such learning involves neural re-programming through reinforcement and correction. Elements w i t h i n the discipl inary matrix are related to each other on the basis of their common famil ial characteristics, by the "network of overlapping and crisscross resemblances". The prospective scientist undergoes a s imilar process to the child's, i n learning to work w i t h the exemplars of his prospective scientific community; "exposure to a series of exemplary problem-solutions teaches [him] to see different physical situations as l ike each other". The practice of normal science depends partly upon this learned abil i ty to group objects and situations into similarity-classes, wi thout an answer to the question "Similar w i t h respect to what?"14 Seeking out criteria for p ick ing out the similarities and differences o n w h i c h any interpretation of data w i l l depend is thus obviated by this focus on the acquisit ion of exemplars, w h i c h provides the student w i t h t raining i n seeing the ET, p. 308. SSR, p. 46; CGK, pp. 273-5; ET, pp. 312-13. 27 Gestalt w h i c h is embodied i n the discipl inary matrix. The science student's basic qualification i n solving exemplary problems "is a perception of similari ty that is both logical ly and psychologically prior to any of the numerous criteria by w h i c h that same identification of s imilari ty might have been made". Further, "an acquired abil i ty to see resemblances between apparently disparate problems plays i n the sciences a significant part of the role usually attributed to correspondence rules." Through her practice i n work ing w i t h exemplars, the science student w i l l come to see new problems as analogous to previously solved problems. There is a form of correspondence here, i n that she sees i n the new problems the attachment to Nature of the symbolic consequences of o ld problems. But despite this emphasis on exemplars, K u h n advises that it is not " l ikely that very much human knowledge is acquired and stored w i t h so little recourse to verbal generalizations"; the effective ostension provided by exposure to exemplars must be combined w i t h model l ing and symbolic generalisations. St i l l , exemplars and models tend to be more effective determinants of communi ty substructure than are symbolic generalisations, perhaps because the former are more completely specified, whi le the latter remain relatively more sparse. Overal l , it is largely i n the shared examples of the communi ty that scientists assimilate and store knowledge . 1 5 A similarity-based scientific programme is preferable to one based on criteria partly because the stricter class-boundaries of the latter (that is, between famil ial groups identified and dealt w i t h by the current theory) tend to foster a situation i n w h i c h class-boundaries w i l l have to be altered more frequently, i n order to accommodate new data. The more "primit ive" resemblance-programme handles cases of new phenomena more often unproblematically, and wi thout continual adjustment. However, "[tjhere are appropriate occasions for switching to the we l l -known strategy that relies upon boundaries and rules": rules become advantageous when the need to resolve a part icularly troublesome quandary arises. 1 6 It is pr imar i ly philosophers w h o tend to want to study scientific examples and derive correspondence rules from them. But this practice has the effect of distorting the v i ew of the way i n w h i c h the scientific communi ty gathers knowledge, by substituting one means of data-processing (rule- or criterion-based) for another (similarity-based). Hence it generates an inaccurate picture of the way i n w h i c h scientific knowledge is acquired and stored. Knowledge develops differently through discipl inary matrices i n general, and exemplars i n particular, from the way it develops through rules. If the cognitive process were t ruly to fol low a fundamentally rule-based pattern, it w o u l d result i n a considerably weaker endeavour. Manipu la t ion of exemplars may be construed as practice i n the application of rules, but i f it is, then we must recognise it as an unconscious sort of practice, since no rules are made explicit i n the passing on of these problem-sets to the student. These shared examples provide a sort of tacit ly embedded knowledge. Rules are actually isolated after the fact, pr imar i ly as an activity of ET, pp. 306-8, 313. Ibid., pp. 316, 318. 28 philosophers wish ing to erect an abstract procedural structure. They should not be v iewed as the hal lmark of objectivity, as concrete puzzle-solutions should not be v iewed as "subjective" or "intuitive"; the latter are undeniably systematic and corrigible. A typical and misguided w a y of understanding exemplars is as evidence for the theory of w h i c h they are applications. But, al though "some of them were part of the evidence at the t ime actual decisions were being made...they represent only a fraction of the considerations relevant to the decision process." B y the time many of these exemplary applications were devised, the decisions pertaining to theory-choice of w h i c h they may appear to be evidence have already been made. Supplying evidence is not their pr imary pedagogic function. If they were offered to science students w i t h such an intent, the composers of textbooks w o u l d be guilty of "an extreme bias", through having failed to offer alternative theories for equal consideration, as w e l l as any evidence against the accepted matrix. In addit ion, as anyone who has been a science student knows, students do not accept theories on the basis of having favourably weighed the evidence for them as presented i n exemplars, but rather on the authority of teacher and text. Clearly, students have not the scientific competence to do o therwise . 1 7 Scientific Crisis, and the Transition into Revolutionary, or Extraordinary, Research If normal science were a l l there were to the scientific endeavour, we w o u l d have the cumulative model w h i c h I have indicated above that K u h n has laboured so strenuously to dispel. A n accurate study of the history of science reveals that the stability of the matr ix of many scientific disciplines has, from time to time, been upset, a new matrix subsequently eventually emerging. Such upsets are generally caused by the accumulat ion of a collect ion of anomalies, discoveries w h i c h are at odds w i t h the current body of theory. A n anomaly is detected when it is recognised "that nature has violated the paradigm-induced expectations that govern normal science"; something has occurred w h i c h cannot "be aligned w i t h professional expectation". This may take the form of a problem w h i c h ought to be solvable by k n o w n rules and procedures resisting solution, or of the failure of a piece of equipment. Init ial ly i n the period of stability, generally on ly the anticipated and the usual are experienced. Researchers at first tend to see only the types of phenomena for w h i c h previous experience w i t h i n the matrix has equipped them; hence, the occasional discrepancy w h i c h crops up w i l l not at this point be noticed. As the same discrepancies appear repeatedly, researchers w i l l begin to become aware of them. Nevertheless, they are usually resolved by normal means. This may involve a minor adjustment of the matrix, so that the anomalous becomes the expected. SSR, p. 80; ET, p. 327. 29 Some discrepancies w i l l resist resolution, however. Such problems may be ignored, i f the researcher deems it more scientifically profitable to move on to attempt to solve other problems, setting aside recognised counter-instances for later w o r k . 1 8 If unsolved anomalies cannot reasonably be placed aside-for instance, i f too many of them have accumula ted - then a certain amount of energy w i t h i n the scientific communi ty concerned w i l l be directed towards their being eventually resolved. At some point, however, the accumulat ion of discrepancies w i l l reach its cr i t ical mass, so to speak, at w h i c h point it is fair to say that a state of crisis has hit the discipline. There is the reappearance of a significant amount of discourse of a cr i t ical flavour. Scientists begin to lose faith i n the adequacy of their discipl inary matrix. Al though they tentatively consider alternatives to it, they continue to resist its full-scale renunciation. Anomalies are sti l l not treated even at this t ime as genuine counter-instances of the matrix. Increasing numbers of researchers ra l ly to devote their professional attention to these difficulties; as a result, anomalies come to be isolated more precisely. A n y rules w h i c h may have been abstracted and identified as exemplary of the matr ix loom larger as its security begins to show strain; they provide a relatively more substantial handle w h i c h adherents to the o ld matr ix are able to grasp, as they labour to stretch that matr ix around persistent anomalies. If scientists demonstrate a tendency to venture into the terrain of philosophical analysis as a help w i t h their p rob l ems-that analysis being, as mentioned previously, occupied largely w i t h the abstraction of a cri teriological skeleton for scientific research-they w i l l do so part icularly at this time. K u h n states that [i]t is no accident that the emergence of Newtonian physics in the seventeenth century and of relativity and quantum mechanics in the twentieth should have been both preceded and accompanied by fundamental philosophical analyses of the contemporary research tradition. Nor it is an accident that in both these periods the so-called thought experiment should have played so critical a role in the progress of research....the analytical thought experimentation that bulks so large in the writings of Galileo, Einstein, Bohr, and others is perfectly calculated to expose the old [matrix] to existing knowledge in ways that isolate the root of crisis with a clarity unattainable in the laboratory. 1 9 Looking, for example, at the Copernican Revolut ion, we may note that, as discrepancies were discovered i n Ptolemy's astronomical system, astronomers were at first able to eliminate them as problems by making adjustments i n the system. But as this process continued, the system had to become ever more complex, i n order to encompass, as newly solved problems, what had been discrepancies w i t h i n it. W i t h t ime, the complexity of the system outstripped its accuracy. Further, manipulat ion of the system i n one place often distorted it i n SSR, pp. 5-6, 52, 64, 81-2, 112-13, 186. Ibid., p. 88. 30 another, g iving rise to yet more new anomalies. Ptolemaic astronomy had reached a state of c r i s i s . 2 0 This account of the increasing manifestation of anomaly as a prelude to crisis clearly depends upon an acceptance of the description of normal science offered above. Without this special apparatus (viz. , normal science) w h i c h is constructed for the purpose of describing and deal ing w i t h anticipated functions, the results that lead ul t imately to novelty w o u l d not be recognised as such. The detection of anomaly occurs because the scientist knows, w i t h precision, what to expect i n his investigations. Anomaly must be highlighted against the background provided by the disciplinary matrix of normal science. Thus, the more precise the matrix, the more sensitive an indicator of anomaly it is. Paradoxically, then, normal science prepares the way for its o w n temporary demise, leading to change i n the regime. Dur ing the crisis-period, the discipl inary restrictions w h i c h are essential to normal science are naturally relaxed, as the previously accepted matrix functions increasingly less effectively. A typical symptom of matrix-crisis is the proliferation of a number of different "articulations and ad hoc versions" of the extant theory; these are offered as attempts to stretch the theory such that it might be capable of contributing a solution to intransigent anomalies. Otherwise, anomalies threaten to remain outside of any structure whatsoever, w h i c h w o u l d render scientific progress considerably more elusive. The result is increasing theoretical vagueness and decreasing theoretical uti l i ty, as the proliferations make it progressively more difficult to identify what the theory actually is. The decaying matr ix guides research w h i c h more and more resembles that characteristic of the period prior to the dominance of any one matrix, what K u h n had earlier called "pre-paradigm" science. K u h n notes that "by proliferating versions of the [matrix], crisis loosens the rules of normal puzzle-solving i n ways that ult imately permit a new [matrix] to emerge." This is the domain of "extraordinary", or "revolutionary" science w h i c h is clearly of a more turbulent nature than that produced i n normal per iods . 2 1 At this t ime, scientific attention is concentrated on a narrow area of trouble, and anomalies are more consciously recognised as such than they were before. Norma l science, i n fact, leads to revolutionary science m u c h more surely than w o u l d pre-paradigm science, because it "[i]solates for continued and concentrated attention those loci o f trouble or causes of crisis upon whose recognition the most fundamental advances i n basic science depend" . 2 2 S t i l l , the impending full breakdown of the matr ix is on ly very rarely explici t ly recognised as such. The atmosphere is rather one of at-times-overwhelming confusion: about h o w Nature is to be viewed, and about the conceptual framework and methodology w h i c h guide research. The relaxing of the o ld order allows new things to be seen, and /o r o ld things to be seen i n a new light. Commitment to the Ibid., p. 68 Ibid., pp. 24, 71-2, 78, 80, 87. ET, p. 234. 31 o ld paradigm has weakened, thus freeing up the vis ion of researchers, increasing its plasticity so that new visions may be forged. Progressively greater numbers of researchers become estranged from what has been tradit ional, and behave i n an ever more eccentric manner i n relation to that t radit ion. The speculative and less fully articulated theories of the crisis-period thus play an important role, i n that they point the way to discovery, whether or not such discoveries are indeed the ones anticipated by the speculative theories. Sometimes what w i l l ul t imately w i n out as the new dominant matr ix is foreshadowed fairly quickly and directly i n the shreds of structure w h i c h scientists are attempting to bu i ld through their extraordinary research: [0]ften a new [matrix] emerges, at least in embryo, before a crisis has developed far or been explicitly recognized...one can say only that a minor breakdown of the [matrix] and the very first blurring of its rules for normal science were sufficient to induce in someone a new way of looking at the field. What intervened between the first sense of trouble and the recognition of an available alternate must have been largely unconscious. 2 3 For that matter, a new matrix may be anticipated long before a crisis point even appears on the hor izon. This was the case when Aristarchus proposed a heliocentric astronomical system i n the Th i rd Century B . C . , dur ing the period w h e n Ptolemaic geocentric astronomy was the norm. Actua l crisis and matrix-transformation d id not occur unt i l Copernicus' t ime. Generally, however, new theories are not glimpsed prior to the c rumbl ing of the o ld , and further, the final stage of the emergence of the new way i n w h i c h ordered data w i l l be produced is inscrutable for quite some t ime even after the discipline has entered into crisis. In general terms, what w i l l eventually be revealed as a superior matr ix to the o ld is that w h i c h exhibits the most coherent body of observation and theory, still being connected to the sensory input (the "given") w h i c h forms part of the grounding of observations. In simple terms, we might say that the preferred theory is the one w h i c h best fits the facts, al though this may be thought to put the case i n an overly empir ical way . I w i l l have more to say on this, i n relation to considerations of what constitutes truth and knowledge, later. The new order must therefore promise to resolve problems left unresolved by the o ld matrix, whi le at the same time, promis ing to preserve to a significant degree the problem-solving abil i ty that has accrued to science through previous theories. Dur ing this transition period, there w i l l be a large, al though incomplete, overlap between the problems that can be solved by the o ld and the new orders. Counter-instances to the o ld matr ix help to permit the emergence of a new one i n w h i c h they are no longer a source of trouble: w i t h i n the new order, they may even appear tautologous. But although, to scientists, the abil i ty of the new matr ix to solve problems more effectively than the o ld is generally "the most significant and persuasive" factor, it may not itself be a compel l ing force i n theory-change. Sometimes, when it is first proposed, the new matr ix has st i l l not solved many of SSR, p. 86. 32 the problems confronting i t , and it may not have dealt convincingly, or effectively, even w i t h those it has solved. More decisive arguments for the new matr ix tend to present themselves after it has had t ime to become more w e l l -developed. These arguments w i l l then pass into the body of normal-scientific theory . 2 4 Alternatively, values other than problem-solving abil i ty may be applied, w i t h variable amounts of weight. A shift may be facilitated i f the new matrix displays a strikingly better quantitative precision than the o ld one, or i f it is s imply more quantitative. Again , i n the in i t ia l stages of the shift, there may be benefits of the new theory w h i c h are not consciously appreciated. W e see this, for instance, i n general relativity's abil i ty to account w i t h precision for the anomaly w h i c h had emerged i n the mot ion of Mercury 's perihelion, but Einstein had not indeed anticipated this. Later revelations of this sort (which result from routine performance of research w i t h i n the replacement-matrix) stimulate a somewhat delayed vote of confidence for that matrix. Sti l l another value w h i c h may be invoked is aesthetic appeal: the new matrix may be neater, simpler, or considered more suitable or appropriate than the o ld . This tends, at times, to be a rather weaker factor than some of the others, however, on account of the crudeness of the early versions of most matrices; aesthetic appeal is not in i t ia l ly particularly well-developed. B y the time the aesthetic force of a theory is appreciated, other means of persuasion have already taken effect. This is not to say, though, that the power of this appeal ends up to be insignificant; it may sti l l advocate strongly for the new theory, albeit w i t h some delay, and it can be helped i n this respect i f scientists who respond positively to aesthetic advantages are themselves important w i t h i n the scientific communi ty . Recourse to these and other values i n judging the appropriateness of a new matr ix are crucial to the progress of science. K u h n states that " i f a new candidate for a [matrix] had to be judged from the start by hard-headed people who examined only relative problem-solving ability, the sciences w o u l d experience very few major revolutions". Sometimes, it is "only personal and inarticulate aesthetic considerations" that can "make at least a few scientists feel that the new proposal is o n the right track." K u h n cites two examples: w h e n first introduced, Copernicus' astronomical theory d id not have many grounds of appeal other than this, and Einstein's general theory of relativity drew mathematicians pr incipal ly on aesthetic grounds. So acceptance of a new matr ix grows through the interplay of a number of values wh ich , i n any given case, w i l l exhibit different strengths i n persuasiveness. 2 5 A new interpretation of Nature, w h i c h w i l l be offered up as a superior successor to the o ld and now-inadequate theory, emerges first i n the mind(s) of one or a few individuals . Since new matrices are born from the soil of o ld ones, they do incorporate much of the vocabulary and apparatus of the o ld . At least part of the achievement of any matrix w i l l thus prove to be more permanent. Ibid., pp. 78, 169. Ibid., pp. 157-8. 33 Addi t ional ly , a revolutionary change may not affect the whole matrix, but on ly one application of i t . Overal l , though, new matrices w ie ld the tools of the o ld i n a new way, setting up novel relationships between the elements involved. In this way, the new matr ix solves anomalies, revealing new regularities not previously known . In order to appreciate the magnitude of change instituted by a matrix-shift, and thereby dissociate it from the tradit ional picture of s imply cumulative science, we need to look more closely at the state of "incommensurabil i ty" w h i c h characterises the gulf between successive matrices. After a shift to a new theoretical and experimental structure, we can say, i n some sense, at least, that scientists f ind themselves i n a different w o r l d from the one i n w h i c h they were work ing before. A n example w i l l illustrate this: Aristotelians saw a swinging suspended stone finally coming to rest according to their belief that a heavy body is moved by its o w n nature from a higher posit ion to a state of natural rest at a lower one. The stone swung back and forth because it was fall ing w i t h difficulty, being constrained by the string into this tortuous back-and-forth pathway. Gal i leo, however, saw i n this arrangement a pendulum: a body that almost succeeded i n repeating the same mot ion ad infinitum. M u c h of his new system of dynamics was constructed around the properties of this pendulum, and was governed by a considerably different way of seeing than that w h i c h Aristotle had possessed: where Aristotle had seen a fal l ing stone, Galileo 1 had seen a pendulum. The new Gal i lean order was born out of a "flash of intuit ion", w h i c h revealed a novel w a y of seeing the w o r l d . The researcher w h o offers tip a new matr ix for consideration must then to i l (along w i t h any who choose to risk fo l lowing h i m at the outset) to develop this programme to the point where it may function as a convincing instrument for scientific progress. K u h n cautions, though, that the changes wrought through a matrix-shift, being less than Jtotal, ensure that the scientists after the shift are, to some degree, st i l l looking at the same w o r l d as before. M u c h of the language and instrumentation remain the same, al though they are often employed differently from before. 2 6 It should be clear from this example that the proponents of different matrices (e.g., Aristotle and Gali leo) effectively set different standards for what is to count as proper science. W e can achieve finer resolution of this point i f we look at stil l further shifts i n the evolution of dynamics. For example, under Aristotel ian and Cartesian dynamics, a theory of mot ion was expected to explain the cause of the attractive forces between particles of matter. But Newton's dynamics impl ied that this sort of explanation was unnecessary, since a l l the theory had to do was note the existence of such attractive forces. Hence, w i t h the adoption of Newton's theory, the question of the explanation of the attractive forces was essentially no longer considered to be a scientific concern. [ The radical nature of the difference of these standards renders pre- and post-revolutionary normal-scientific traditions (i.e., the o ld and the new matrices) "incommensurable", w h i c h is to say that one matr ix cannot flatly be translated Ibid., pp. 118-9, 122,129. ; 34 into the other, for they carve up the wor ld , as it were, i n different ways. Al though the vocabularies of alternative matrices superficially consist largely of the same set of terms, they apply many of those terms to Nature i n remarkably different ways. Similar i ty relations between phenomena change, such that objects are grouped into similarity-sets (learned v ia exemplars) differently i n the new matrix. At one point, for example, metals were transferred from the set of compounds to that o f elements. Further, the content of the Copernican statement, "planets travel around the sun" cannot be expressed in a statement that invokes the celestial taxonomy of the Ptolemaic statement "planets travel around the earth". The difference between the two statements is not simply one of fact. The term Splanet' is a kind term in both, and the two kinds overlap in membership without cither's containing all the celestial bodies contained in the other. A l l of which is to say that there are episodes in scientific development which involve fundamental change in some taxonomic categories and which therefore confront later observers with problems like those the ethnologist encounters when trying to break into another culture. 2 7 W i t h these redistributions, researchers i n competing matrices may respond to the same st imuli i n incompatible ways. These incompatabili t ies w i l l "cluster most densely about the phenomena upon w h i c h the choice of theory most centrally depends". 2 8 Because o f this, " [c jommunicat ion across revolutionary divides is inevitably par t i a l " . 2 9 It is important to see that problems of incommensurabil i ty are not of a merely linguistic nature. This means that inter-matrix translation cannot be resolved through a simple re-stipulation of terms. The difference between matrices, al though reflected i n their language, is actually prior to the applicat ion of those languages. K u h n believes that a language-metaphor for describing alternative matrices w o u l d be too inclusive, given that we are really deal ing w i t h what he calls a "lexical taxonomy", w h i c h describes the meanings of a restricted class of terms: natural kinds, artifactual kinds, social kinds. K u h n cautions that we might more appropriately think i n terms of concepts, rather than of words. This w o u l d make it something ak in to a conceptual scheme, w h i c h is not a set of beliefs, but "a particular operating mode of a mental module prerequisite to having beliefs, a mode that at once supplies and bounds the set of beliefs it is possible to conceive." This taxonomy must therefore be i n place before a description o f the w o r l d can begin, because unproblematic communica t ion is on ly made possible by shared taxonomic categories: " [ i ]n discussion between members of communit ies w i t h differently structured lexicons, assertability and evidence play the same role for both only i n areas (there are always a great many) where the two lexicons are congruent". It is these areas of overlap w h i c h provide the necessary "bridgeheads" from w h i c h they endeavour to evaluate their differences, RSS, p. 5. SSR, p. 201. Ibid., p. 149. 35 and w h i c h permit the members of one group to attempt to acquire the lexicon of the other i n order to assess its viabi l i ty . P inpoint ing areas of crisis, then, is tantamount to localising incommensurabil i ty i n particular areas i n w h i c h two lexical taxonomies d iverge . 3 0 That new matrices characteristically permit predictions different from those a l lowed by the o ld matr ix demonstrates the logical incompatabil i ty of two alternative matrices. W h i l e this is not to say that the logical inclusiveness of two matrices is, i n principle, impossible, it is, according to Kuhn's observations, "a historical implausibi l i ty". Matrices at least appear to be logical ly exclusive. Their "incompleteness of logical contact" ensures that proponents of different matrices "talk through each other". 3 1 The recovery, by the historian of science, of older meanings cannot therefore be a process of translation, since this is not possible where there is incommensurabil i ty. The shift cannot, as a result, be made one step at a time, but must occur a l l at once, a reflection, also, of the matrix as a fundamental unit for scientific research, one w h i c h cannot be fully reduced to logical ly atomic components. W e see, for instance, pockets of inter-definition of te rms-exhib i t ing "local ho l i sm"~which render i t necessary that many o f the terms o f scientific lexicons be learned i n clusters. Newtonian "force" and "mass" exemplify this; moreover, these terms must be learned i n concert w i t h Newton's Second Law of Mot ion , w h i c h guides h o w one is to pick out Newtonian forces and masses. This renders the shift a process more l ike a GestaZt-shift, the sort w h i c h produces "bilinguals", al though not through translation-efforts. 3 2 It is difficult for the proponents of alternative matrices to decipher how it is that they differ. They must first recognise each other as members of different lexical communit ies. In the in i t ia l stages of theory-conversion, they w i l l attempt to translate each theory into the other. They w i l l do so by first isolating the terms wh ich are used unproblematical ly w i th in each community, but w h i c h are troublesome between the groups. They w i l l attempt to predict what the proponents of the competing theory w o u l d see and say w h e n presented w i t h a particular stimulus. They become more skilful at predict ing each other's behaviour. Dur ing this period, it is important that they refrain from judging anomalous responses as incorrect. They w i l l also attempt to demonstrate to each other the technical advantages of their o w n theory. The demonstration of impressive concrete results of the new theory w i l l persuade some adherents to the old to try to ascertain how those results were achieved. But they w i l l do so sti l l operating i n the mode of translation, attempting to foster the process by reading papers, ta lking w i t h proponents of the new theory, and watching them and their students work . If the proponents of the o ld theory are unable to produce results as impressive as those of the new theory, the new theory begins to take hold , and those w h o had c lung to the o ld theory find that, at some point, they have ceased RSS, pp. 4-5, 9, 12. SSR, pp. 97, 110. Ibid., pp. 11, 85; RTC, p. 566; RSS, pp. 4-5. 36 to translate, and instead have begun to speak the new language. "Exploring an alternative theory by techniques [such as these], one is likely to find that one is already using [the new lexicon] (as one suddenly notes that one is thinking in, not translating out of, a foreign language)." The depth of the taxonomic divide means that, if proponents of alternative matrices fail in their discussions to achieve bilingualism, they are bound to end up talking through each other somewhat.33 But a matrix-shift is not strictly similar to a Gestalt-shift, on account of the inability of the scientific enterprise to allow its practitioners to shift back and forth between Gestalten. We may see what appear to be reversals in the direction of shift, as in the greater resemblance of the explanation of gravitational attraction offered by Einstein to those of Newton's predecessors than to those of his successors. But this is, of course, not a genuine reversal, since the matrix promoted through Einstein's work was nevertheless a new one overall. Periods where this kind of instability does occur are notably periods of crisis (e.g., when light was sometimes believed to be a wave, and other times believed to be a particle). Further, the scientist does not see the matrix-shift in the same light as the psychological subject, for example, sees a Gestalt-shift. For instance, the scientist does not say that, looking at the moon, she once saw it as a planet, but now sees it as a satellite; her judgment is instead that, while she once saw the moon as a planet, she was mistaken. This speaks of a more solid commitment to a chosen matrix than the GestaZt-image would suggest. That we see incommensurability at all is an indication of the absence of an overarching language-or, more accurately, in Kuhnian terms-lexical taxonomy "capable of expressing, in its entirety, the content of [all lexicons] or even of any pair". This entails the absence of a neutral observation language, as well as of external standards by which one might judge a matrix to be inherently unscientific: "what occurfs] [is] neither a decline nor raising of standards, but simply a change demanded by the adoption of a new [matrix]." Adherents to the old and the new differ with respect to the desired "institutional matrix", and there is no "supra-institutional framework for the adjudication of revolutionary difference". This lack of an overarching scientific standard entails that whole theories in competition with each other for preferred-matrix status cannot be evaluated using the same sort of criteria that we can isolate as being used in judging a single theory's individual research-applications. We might put this another way, by saying that there is no such thing as an algorithm for theory-choice-a set of rules through the application of which scientists might make a clear determination of a superior theory. If there were such a thing as an algorithm for theory-choice, it would "eliminate all need for recourse to group values, to judgments made by minds prepared in a particular way". The presence of an algorithm would thus contradict Kuhn's emphasis on exemplars as an important means of training in the collective practice of scientists through 34 exposure to concrete scenarios. SSR, pp. 148, 201-2; CGK, p. 277; ET, p. 339. SSR, pp. 93-4, 200-1; CGK, pp. 12, 238. 37 In truth, there is an inaccuracy i n describing theory-choice i n terms of decision at a l l , since theories are research-programmes to be evaluated over t ime, and there is no "conscious deliberation on each issue prior to the assumption of a research stance". At no point is the scientist aware of having reached a decision or made a choice. Thus, the transfer of allegiance from one theory to another is better described as "conversion". Scientists come to evaluate the posit ion i n w h i c h they f ind themselves as a result of their having made the conversion. It is on ly after the fact of conversion to a new theory that terms such as "choice" and "decision" are applied; the true "nature of the change [is] disguised i n later reports". 3 5 Having asserted a l l of this, however, K u h n insists that the lack of a neutral lexical taxonomy does not point us to the intrinsic irrat ionali ty of scientific theory. Against the charges of some of his critics, he is not "trying to make science rest on unanalyzable ind iv idual intuitions rather than on logic and law". He argues that one can sti l l pick out good reasons w h i c h take part i n the conversion from one theory to its r i v a l . 3 6 The lack of an algori thm for theory-choice means that arguments deployed by scientists i n favour of particular theories are not logical ly compel l ing. Some of Kuhn's readers have sharply crit icised this notion, seeing it, i n one account, as leading to a situation governed by "mob psychology" 3 7 . This is a characteristically philosophical complaint, since, "this mode of development [which] requires a decision process w h i c h permits rational men to disagree" (i.e., revolutionary science as K u h n has described it) w o u l d be "barred by the shared algori thm w h i c h philosophers have generally sought". K u h n has taken pains (more than should have been required) to emphasise that rationality is essential to the scientific enterprise, i n times of both crisis and stability. For h i m , "rationality" and "justification" are inter-defined terms. Incommensurabil i ty is not a threat to the rational evaluation of truth-claims. The lack of logical compuls ion w h i c h we see i n inter-matrix arguments does not mean that scientists ever refrain from using logic and mathematics i n these arguments: "[t]o say that, i n matters of theory-choice, the force of logic and observation cannot i n principle be compel l ing is neither to discard logic and observation nor to suggest that there are not good reasons for favouring one theory over another". There is no impl ica t ion "that scientists may choose any theory they l ike so long as they agree i n their choice and thereafter enforce it": "[o]ne scientific theory is not as good as another for do ing what scientists normal ly do". It is perfectly possible for arguments to be logical , and yet uncompel l ing. There may be a disagreement over premises, or over the manner i n w h i c h they are to be appl ied-both being a common feature among scientists debating theory-changes. At some point, one theory must assume dominance over the others; for scientists at times such as these, "their recourse is to persuasion as CGK, pp. 238, 276-7; ET, p. 338. SSR, p. 191; CGK, p. 235; ET, p. 328. Lakatos, CGK, p. 178: "Thus, in Kuhn's view scientific revolution is irrational, a matter for mob psychology." 38 a prelude to the possibility of p r o o f . Thus, decision-problems between alternative matrices cannot be resolved by recourse to proof-rather, they involve "techniques of persuasion, or...argument and counter-argument i n a situation i n w h i c h there can be no p r o o f . This does not refer only to proofs of a logical sort; the competi t ion between matrices cannot be solved through appeal to empir ical proofs, either, since proponents of competing matrices are generally unwi l l i ng to grant the non-empirical assumptions of the others . 3 8 The arguments offered for the adoption of a matr ix by its proponents are necessarily circular: each group must use the tools of its o w n matrix to argue i n its defence: The resulting circularity does not of course make the arguments wrong or even ineffectual. The man who premises a paradigm when arguing in its defense can nonetheless provide a clear exhibit of what scientific practice wi l l be like for those who adopt the new view of nature. That exhibit can be immensely persuasive, often compellingly so. Yet, whatever its force, the status of the circular argument is only that of persuasion. It cannot be made logically or even probabilistically compelling for those who refuse to step into the circle. The premises and values shared by the two parties to a debate over [matrices] are not sufficiently extensive for that. As in political revolutions, so in [matrix] choice-there is no standard higher than the assent of the relevant community. To discover how scientific revolutions are effected, we shall have to study not only the impact of nature and of logic, but also the techniques of persuasive argumentation effective within the quite special groups that constitute the community of scientists. 3 9 The rationality involved i n theory-choice can be seen i n the struggle w h i c h scientists undergo when applying the values mentioned above to the question of theory-choice-e.g. , accuracy, precision, simplicity, plausibil i ty, broad scope, enhanced qualitative function, internal consistency, compatibi l i ty w i t h other theories, and unanimous (or, at least, as nearly as possible) acceptance throughout the relevant scientific community . These values are important i n that they provide the shared basis for theory-choice, i n a communi ty where a shared grounding is v i ta l to its achievements. Because they are imprecise, however, ind iv idua l scientists may legitimately differ about their application to concrete cases. Different scientists may disagree about the relative weights to be accorded to various values when several of them are to be deployed. Accuracy is the most nearly decisive of a l l the criteria: it is the least equivocal, and the predictive and explanatory powers so important to scientists depend upon it. But theories cannot always be discriminated i n terms of accuracy. For example, Copernicus' system was not more accurate than Ptolemy's unt i l Kepler later drastically revised it, and this only occurred because Kepler chose Copernicus' system for reasons other than accuracy. Individual differences i n respect of relative emphasis on different values need violate no accepted scientific practice. So values may conflict: accuracy may SSR, pp. 148, 158, 198-9; CGK, pp. 234, 261-2, 264; ET, pp. 320, 332; RTC, p. 569; RSS, p. 3. SSR. p. 94. 39 dictate the choice of one theory, while other considerations-of scope, say-favour its competitor. Two different theories may each be more accurate or simple than the other, but in different areas of their operation or application. For instance, considering simplicity, the computational labour involved in predicting the position of a planet was pretty much the same within the Copernican and Ptolemaic theories, but Copernicus' system was much simpler in terms of the mathematical apparatus used to explain retrograde motion. No one value is sufficient on its own for theory-choice. So values function not as rules which determine theory-choice, but instead merely influence the choice. The revolutionary shift to a new matrix may involve a shift in the relative weighting and application of values. This constitutes a feedback loop. As I mentioned before, Kuhn stresses that science is inherently a group activity. A matrix shift, then, will require the consensus of the scientific community affected by it. Nevertheless, it is absurd to expect that all members of the community will become convinced of the wisdom of the change at the same rate. So rather than a single group-conversion, there will be "an increasing shift in the distribution of professional allegiances". Gradually, the number of experiments, instruments, and publications based oh the new order will multiply, and more researchers "will adopt the new mode of practicing normal science, until at last only a few elderly hold-outs will remain". But "after the last hold-outs have died, the whole profession will again be practicing under a single, but now a different, [matrix]".40 This last statement reflects a thesis which has caused Kuhn much difficulty with critics. It is that, in both the pressure for matrix-transformation which brings about crisis and the theory-choice which is demanded of scientists in times of crisis, there are factors at play of which notice is not typically taken in traditional accounts of science. Returning to the crisis leading to the Copernican Revolution, we see that, although the breakdown in normal technical puzzle-solving activity was the primary stressor on the old Ptolemaic system, there were other significant factors, as well: social pressure for calendar reform, medieval criticism of Aristotle, and the rise of Renaissance Neo-Platonism~i.e., a number of historical elements, which could not be counted as scientific in themselves. Similarly, there are a number of unscientific influences on theory-choice by researchers. For instance, those who arrive at a candidate for a new matrix are almost always "either very young or very new to the field". They are less "committed by prior practice to the traditional rules of normal science", and more likely to see the inadequacy of those rules as it begins to appear. On the other hand, those scientists who are immersed, through time and practice, within the old matrix tend to resist its demise, demonstrating greater faith than the new-comers that their time-tested matrix will eventually solve all of its problems. Theirs is a necessary trait for the progress of normal science. But in times of crisis, and especially at the later stages of crisis, and when a new matrix-candidate has Ibid., pp. 152, 158-9. 40 reached the point of serious consideration by the relevant community , this attitude may prove to be unreasonably s tubborn . 4 1 Other non-logical and arguably unscientific influences may sit further afield from mere attachment to that to w h i c h one has become adjusted through thought and practice. In this regard, K u h n refers to "the sun worship that helped make Kepler a Copernican". Concern about one's reputation wi th in the scientific communi ty may also influence the energy w i t h w h i c h one pursues the promot ion of a new matr ix. Differences i n the environment where science is practiced-e.g. , different countr ies-can also have an effect on research and theory-choice. 4 2 Matrix-debates, then, are not strictly about relative problem-solving abili ty. The central issue concerns w h i c h matr ix should guide future research on problems, and this question is raised at a t ime when neither competitor is capable of resolving al l of its problems. The promise of future success of the matrix directs the choice somewhat more than does the past success of the o ld order. It is thus largely on faith that the scientist w i l l embrace a new matrix. But this is not to say that such faith is b l ind : it requires a basis, al though this basis "need be neither rational nor ul t imately correct". A t bottom, significant factors w i l l be whatever scientists w i l l a l low to persuade them to change their minds . Certainly, as noted above, argumentation plays a considerable role i n this process, but Kuhn's controversial point is that choices are not dictated by logic and experiment alone: socio-psychological imperatives are also involved. The Zeitgeist has a role i n intellectual affairs. Considering the nature of the scientific enterprise and the attitude of its practitioners, however, one w o u l d not expect such factors to be merely frivolous i n their nature. Understanding what scientific progress is requires an understanding of the value-system, or ideology, employed by scientists. This system is transmitted and reinforced through the institutions w h i c h characterise science, and so the explanation of progress w i l l , " i n the final analysis, be psychological or sociological". One of Kuhn's "irreducibly sociological principles" is that "[wjhatever scientific progress may be, we must account for it by examining the nature of the scientific group, discovering what it values, what it tolerates, and what it disdains." This posit ion is "intr insically sociological", or "ideological", departing from "dogmatic and naive" "justificationism and falsificationism". K u h n is not here intending to invoke an individual 's psychological idiosyncrasies, but rather "the common elements induced by nurture and training i n the psychological make-up of the licensed membership of a scientific group". These elements are, at least i n principle, analysable. The "social psychology" to w h i c h K u h n is referring is "quite different from indiv idual psychology reiterated n times". But neither is it an idealised m i n d w h i c h practices normal science; there are, i n Kuhn's estimation, no such m i n d s - o n l y shared ideals w h i c h affect behaviour . 4 3 Ibid., pp. 69, 90, 144, 151. ibid., p. 152; ET, p. 333. SSR pp. 79, 152, 157-8, 191; CGK, pp. 21-2, 237, 240; RSS, p. 3. 41 W i t h i n this social-psychological net of influences offered to the scientist as part of his training, there is, nevertheless, room for variabil i ty based upon divergence i n the indiv idual emphasis placed upon particular factors. Some of the factors relevant to theory-choice are: 1. the individual 's previous experience as a scientist, e.g., h o w successful he has been i n his work thus far, i n what particular field he was work ing when confronted by the need to choose, and how much of his work depended upon concepts and techniques affected by the crisis; 2. the individual 's personality, e.g., the drive towards originali ty w i l l be stronger i n some researchers than i n others, leading to both a relatively greater wil l ingness to take risks and less ego-attachment to the theory; and 3. the individual 's involvement i n or influence by social movements, e.g., Kepler's immers ion i n Neo-Platonic and Hermetic movements, German Romanticism's predisposing effect towards the acceptance of energy conservation, and nineteenth-century Bri t ish social thought affecting Darwin 's concept of the struggle for existence. Every ind iv idua l choice w i l l thus be the result of the applicat ion of a mixture of relatively more shared and relatively more ind iv idua l values. But even the indiv idual values are heavily t inged by the t raining w h i c h prepares the student for eventual acceptance into the scientific group, e.g., education and prior pattern of professional research. So, for instance, conce rn -wi th in reason, that is~for the state of one's professional reputation could plausibly be seen as heavily influenced by these shared ideals, or common elements, since that reputation is an important key to the legit imation of the body of one's work, and therefore of one's contr ibution to the progress of the scientific endeavour. This progress is obviously an essential concern for scientists. It w o u l d be a rather dramatic overstatement, i n Kuhn's estimation, to label these socio-psychological influences as "irrational". He has insisted that "[n]o process essential to scientific development can be labelled ' irrational ' wi thout vast violence to the term". He prefers the mi lder term "arational" to describe these factors wh ich , whi le not introducing a w i l d card into the scientific enterprise, nevertheless depart somewhat from strict logical considerat ions. 4 4 Considering the variety of influences w h i c h tug at the scientist confronted w i t h a situation eventually demanding theory-choice, creative scientists must be able to tolerate l iv ing w i t h an "essential tension", l iv ing " i n a w o r l d out of joint". This tension exists between "convergent" and "divergent" thinking, and the abi l i ty to support it, even when it "can occasionally become unbearable is one of the pr ime requisites for the very best sort of scientific research". Convergent th inking typifies consensus-bound research, whi le the divergent tendency has to do w i t h the flexibili ty and open-mindedness w h i c h a l low the imaginat ion to play w i t h unl ikely possibilities, refraining from becoming too attached to apparently self-evident facts or concepts. The judgment called for at this t ime of confusion mediates between the conservative retention of the o ld order and the riskier leap CGK, p. 235; Horgan, p. 42. 42 to the new. The scientific group must simultaneously display the characteristics of both tradi t ional ism and iconoclasm. Even i f a new theory does begin to take ho ld --and most do n o t - m u c h theoretical and experimental work is usually required "before the new theory can display sufficient accuracy and scope to generate widespread convict ion". A t this t ime, research guided by both the o ld and the new matrices w i l l be conducted. The scientist w i l l w i s h to preserve as many of her group's puzzle-solutions as possible (constituting a force w h i c h speaks i n favour of guarding the integrity of the o ld matr ix) , but she w i l l also wi sh to maximise the number of puzzles that can be solved, and this may be better served by the shift to a new ma t r ix . 4 5 The individual i ty of scientific value-judgments w i t h i n the communi ty (meaning the relative weight ing of the values listed above) is one of the tangled roots of the scientific crisis. It plays an important role i n fomenting the essential tension w h i c h mediates theory-change, i n terms of both openness to imaginative alternatives to o l d theories and moderat ion of the acceptance of new ones. A n identification of the general characteristics w h i c h typify good scientific theories is not by itself sufficient to determine the decisions of ind iv idua l scientists. Individual choice i n respect of w h i c h canon to follow must also be taken into account: "[o]ne can explain, as the historian characteristically does, w h y particular men made particular choices at particular times". That the members of the group do not a l l apply the shared basic values of the communi ty i n the same way ensures that ind iv idual researchers respond to anomalies i n different ways: some w i l l detect the roots of crisis where others w i l l see problems satisfactorily solvable by the o ld ways. Some w i l l see crisis at a point at w h i c h others see "only evidence of a l imi ted talent for research". If researchers were to respond i n s imilar ways, they wou ld , on the one hand, collectively see each anomaly as a source of crisis, w h i c h w o u l d impel them to embrace each new theory that came along. There w o u l d be none left to try to make the existing theory account for current apparent anomalies. Such fickleness w o u l d spell the end of normal science, having severe repercussions on scientific progress. Moreover, that most proposals for new theories prove to be unwise, and ult imately insupportable, w o u l d deepen the folly of too free and easy an attitude towards the taking of theoretical risks. O n the other hand, it is necessary for the health of the scientific enterprise that somebody i n the communi ty react to anomalies and new theories i n a higher-risk sort of way; otherwise, the revolutions w h i c h have proved essential to the developmental progress of science w o u l d be given no opening to occur. W e might v iew this ind iv idual divergence as a "hedging of bets", "the community 's way of distr ibuting risk and assuring the long-term success of its enterprise". Individual variabi l i ty thus functions as a strength: i f a decision must be made under circumstances where there is a significant possibility of an error in judgment , it may be important that individuals judge i n different ways. SSR, p. 186; CGK, p. 21; ET, pp. 226-7, 332. SSR, pp. 185-6; CGK, pp. 21, 241, 248, 262; ET, p. 226-7, 324-5, 332. 43 In the end, the highly important value of unanimity w i l l require that the relevant scientific communi ty w i l l converge around a new matr ix deemed acceptable as that w h i c h determines the future tone and direct ion of normal scientific research. Decisions as to the preferred matr ix are made by the communi ty of specialists, not by individuals . So the group behaviour w i t h respect to theory-choice " w i l l be affected decisively by shared commitments" . 4 7 Empiricism, Kantianism, and the Spectre of Relativism N o r m a l science, w h i c h is what occurs most of the t ime, operates on the assumption that the scientific communi ty knows what the w o r l d is l ike. This assumption is necessary for the success of the enterprise, being closely connected to the suppression of novelty w h i c h typifies normal science, since a sense of security i n one's knowledge of the w o r l d requires a certain amount of stability. The usual v i ew of what occurs i n a matrix-shift is that there is an alteration i n the scientist's interpretation of observations w h i c h are fixed by both the nature of the environment and human perceptual abilities. O n this v iew, an inherently empir ical one, proponents of different matrices see the same w o r l d differently. But K u h n argues that data are not s imply stable i n this way. Within a given matrix, the researcher is confident i n his knowledge of what counts as a datum, what instrumentation is needed to reveal it, and what concepts are relevant to its interpretation. Therefore, whi le we are indeed justified i n making reference to interpretation of data as being part of what is going on i n science, we can only do so w i t h i n the confines of a particular matrix. Aristotle interpreted his observations of the fal l ing stones defined by his matrix; Gali leo interpreted his observations of the pendula of w h i c h his matrix spoke. Interpretation can only provide an articulation of a matrix; it is incapable of correcting that matrix. Interpretation is a logical exercise, and as such, w o u l d be logical ly l inked to the experience of a given normal scientific matrix. But i n revolutionary science, the "flash of intui t ion" through w h i c h the beginnings of a new matrix are born does not embody such logical l inkages . 4 8 In revolutionary science, then, we see a difference i n what constitutes data between the o ld order and the new. (Again, this w i l l not be total: at least some of the o ld matr ix w i l l be carried through the shift.) So, for instance, it is not the case that a scientific observer sees a swinging stone, and then interprets it as either an object i n constrained fall or a pendulum. For the Gal i lean researcher, seeing a swinging stone is no more elementary than seeing a pendulum: the observation is dependent upon the matrix w h i c h acts as a prior defini t ion of observations. The observation of a fall ing body is a v is ion through one matrix; that of a pendulum is a v is ion through another. So measurements or retinal imprints cannot be taken as SSR, p. 200; CGK, p. 241. SSR, pp. 120-3. 44 foundational, since they must already be part of a matr ix i n order to be intelligible at a l l . This is the only way that we can pose any questions about such measurements or imprints . If the matr ix changes, so do the questions we ask: these questions "presuppose a w o r l d already perceptually and conceptually subdivided i n a certain way". In order for a scientist even to enquire what measurements or retinal imprints make the pendulum what it is, she must first be able to recognise a pendulum. This connects to Kuhn's assertion that we do not learn to see the w o r l d item-by-item, since the recognit ion of something as a particular th ing occurs w i t h i n an accepted framework of similarity-relations (i.e., 49 w i t h i n a context). It must be understood, though, that the a priori structure, set out by the operative matrix, w h i c h defines the scope and meaning of normal research does not, for K u h n , provide a criterial conceptual framework w h i c h w o u l d a l low the scientist to specify i n advance whether or not each imaginable observation w i l l fit or falsify the theory. The necessary presence of a framework of intel l igibi l i ty~one w h i c h allows the recognition and, therefore, continued observation of pendula, for instance-must not be construed as obviating amendment of the framework by observation. For example, one cannot seek a cri terion w h i c h w o u l d specify, at the outset, that a l l swans are white , such that the discovery of a black swan falsifies that theory. The discovery of a black b i rd w h i c h strongly resembled swans i n other respects w o u l d eventually cause a crisis for the theory. There w o u l d be an ensuing focus on research i n this troublesome area, and this w o u l d include the search for more black birds of this type. In the end, either the theory w i l l be amended to include black swans, or a new natural family made up of these birds w i l l be introduced. W h i c h occurs w i l l depend largely upon the strength of the theoretical belief that colour is important for characterising natural families such as this. The point to note here is that both logical considerations and observation of concrete particulars w i l l play a significant part i n the outcome, whether this be the adjustment of the existing theory or the creation of a new taxonomic group. The lack of a criterial basis for theory reconstruction dovetails w i t h Kuhn's description, outl ined previously, of the growth of theory through the bu i ld ing of similarity-relations, based on ostension. Some of these relations w i l l be altered through revolution, so that objects and situations are grouped differently from how they were before. So the o ld empir ical picture of sensory experience that is fixed and neutral no longer functions effectively. The scientist only sees his data once "his research is w e l l advanced and his attention focused" 5 0 . M a k i n g sense of fixed and neutral sensory experience probably w o u l d require the discovery of a pure observation-language, and this w o u l d depend upon the construction of a satisfactory theory of perception and of the mind , to w h i c h we are not even close. Very different s t imul i can produce the same sensations i n different individuals, the same stimulus can produce very different sensations, and the route from stimulus to sensation is Ibid., pp. 127-9. Ibid., p. 126. 45 condit ioned, i n part, by education and socialisation. " [S] t imul i receive much neural processing before anything is seen or any data are given to the senses." 5 1 If the members of two divergent research-groups have systematically different sensations upon receipt of the same st imuli , then, in some sense, they l ive i n different worlds . These worlds not on ly diverge among communities, but also alter w i t h t ime. Individuals of the same group share a c o m m o n background i n education, language, experience, and culture. Thus, we have good reason to suppose that their sensations are similar . They also exhibit coherently connected communicat ion and behaviour. But, w i t h the differentiation and specialisation of groups, we have no evidence for the immutabi l i ty of sensation. Hence, it may be parochial to assume that the route from stimulus to sensation is the same for everyone. W e do not even have direct access to s t imuli , since these are only known, themselves, through elaborate theory. O n pa in of solipsism, though, we must regard the st imuli to w h i c h observers are exposed as the same for everyone. 5 2 It is along the pathway from stimulus to sensation that the various groupings of observations w h i c h distinguish alternative matrices w i l l be learned. K u h n says, I think it likely myself that much or all of the clustering of stimuli into similarity sets takes place in the stimulus-to-sensation portion of our neural processing apparatus; that the educational programming of that apparatus takes place when we are presented with stimuli that we are told emanate from members of the same similarity class; and that, after programming has been completed, we recognize, say, cats and dogs (or pick out forces, masses, and constraints) because they (or the situations in which they appear) then do, for the first time, look like the examples we have seen before.53 At any rate, K u h n asserts that we should not over-emphasise the differences i n experience between either individuals or communit ies. W e do, after a l l , consider ourselves to share i n a common sort of general neural apparatus, however divergent the "programming" w h i c h influences this biological structure may be: "[d]oubtless some aspects of that lexical structure are biological ly determined, the products of a shared phylogeny." Beyond even the biological , much of the "programming" of individuals must also be similar, given their sharing of a significant amount of their history, perhaps of a language and an everyday wor ld , as w e l l as most o f their scientific w o r l d . Really, even i n great revolutions, the area under dispute is relatively narrow. Given what they share, and employing sufficient w i l l , patience, and tolerance of the threat of ambiguity, scientists occupying divergent viewpoints can nevertheless come to find out much about how they differ. Their commonal i ty provides a resource upon w h i c h they can rely w h e n venturing into the uncertain territory wrought by a state of c r i s i s . 5 4 CGK, p. 276. SSR, pp. 126, 128, 193, 196; CGK, p. 276; ET, p. 308; RSS, p. 10. CGK, p. 276. Ibid., RSS, p. 10. 46 Even though the proponents of different matrices practice, in many respects, in different worlds, "that is not to say that they can see anything they please...or not at all". Kuhn insists that there is an integrity to perception, one which is born of the selection process which is driven by survival value; the experience and knowledge of Nature are embedded in the stimulus-to-sensation route. Thus, the neural process that transforms stimuli to sensations has a number of built-in characteristics: 1. it is transmitted through education; 2. it has been found, through trial-and-error, to be more effective than its historical competitors; and 3. it is subject to future change through further education, and through the discovery of a lack of fit with the environment. Clearly, then, given his reference to stimuli and goodness-of-fit with the environment, Kuhn is not proposing an understanding of knowledge as strictly constructed by humans. The stimulus is that which is "given" by Nature, although it can only be rendered intelligible to us when we perceive it through a human template of some sort. Although the acts of judgment which feed into scientists' rejection of a previously accepted theory and acceptance of a new one involve the comparison of theories with each other, they are always based upon more than this: they also entail a comparison of those theories with the world. Empirical observation is an essential part of the scientific enterprise: "[m]ost of the puzzles of normal science are directly presented by nature, and all involve nature indirectly". Both normal and revolutionary science, in their respective ways, strive to bring theory and fact into closer agreement. While the nature of facts is, to some degree, determined by the matrix through which they are observed, they include some measure of the "given". It is in this respect that observation is so important. Different matrices should then be understood as telling us different things about the entities of which Nature is composed, and how they behave. "They are not simply about names or language but equally and inseparably about nature." In scientific progress, then, it is fair to say that we see an increasingly detailed and refined understanding of Nature. The student's contemplation of the exemplars of her prospective profession is an essential part of her learning how the currently accepted laws of science attach to Nature: in short, how the world behaves.55 We must take care, then, in understanding the reconstructive nature of science to refrain from falling into the belief that the world is somehow mind-dependent, perhaps an invention or construction of the creatures which inhabit it, and in recent years such suggestions have been widely pursued. But the metaphors of invention, construction, and mind-dependence are in two respects grossly misleading. First, the world is not invented or constructed. The creatures to whom this responsibility is imputed, in fact, find the world already in place, its rudiments at their birth and its increasingly full actuality during their educational SSR, pp. 77, 80, 125, 145, 150, 170, 195-6; CGK, p. 262, 274, 276; RSS, p. 10. 47 socialization, a socialization in which examples of the way the world is play an essential part. That world, furthermore, has been experientially given, in part to the new inhabitants directly, and in part indirectly, by inheritance, embodying the experience of their forebears. As such, it is entirely solid: not in the least respectful of an observer's wishes and desires; quite capable of providing decisive evidence against invented hypotheses which fail to match its behavior. Creatures born into it must take it as they find it. They can, of course, interact with it, altering both it and themselves in the process, and the populated world thus altered is the one1 that w i l l be found in place by the generation which follows. The point closely parallels the one made earlier about the nature of evaluation seen from a developmental perspective: there, what required evaluation was not belief but change in some aspects of belief, the rest held fixed in the process; here, what people can effect or invent is not the world but changes in some aspect of it, the balance remaixiing as before. In both cases, too, the changes that can be made are not introduced at wi l l . Most proposal [s] for change are rejected on the evidence; the nature of those that remain can rarely be foreseen; and the consequences of accepting one or another of them often prove to be undesired. 5 6 This is perhaps an overly vigourous attempt to defend the retention of a sense of there being a component o f experience w h i c h is "given" by Nature. It should be clear from the discussion of Kuhn's work so far that he does not here mean to i m p l y the absence of a further component of experience, supplied by the m i n d as a formal structure of some sort, one w h i c h organises the "given" into experience w h i c h is understood i n a particular way. This wor ld , even though partially mental ly constituted, and partially "given", can nevertheless rightfully be referred to as the "real wor ld" . After a l l , it provides the environment for life, p lacing constraints on that life, so that continued existence depends upon adaptation to those constraints. Science functions as an important contemporary tool for that adaptation. "What more can reasonably be asked of a real w o r l d ? " 5 7 The part ial ly mental ly constituted, partially "given" nature of the experienced wor ld is reflected common-sensically i n that, when we v iew a communi ty from the outside, we see i t as adapting to its environment, whi le , from a vantage-point inside a community, our emphasis is on the mediat ion of our interactions w i t h our environment through "something l ike a mental representation". In Kuhn's v iew, then, mental representations have, to some extent, a world-consti tuting r o l e . 5 8 Under this scheme, fact and theory are not categorically separable. The assimilation of a new theory requires the reconstruction of both pr ior theory and fact. This is because a fact is the result of the integration of the "given", and the formal mental elements provided by the matrix. In a revolutionary re-formulation of the preceding scientific tradit ion, there is no piecemeal evolut ion of a theory to fit facts that were always there. W e encounter no rules for inducing theories from 56 RSS, p. 10. 5 7 ibid. 5 8 Ibid, p. 11. 4 8 facts, since theories are not induced at a l l , but are rather "imaginative posits". Theories "fit the facts" by transforming previously accessible information into facts that had not existed for the preceding matrix; they emerge together i n a tapestry w i t h the facts they fit. Similar ly, as a result of a matrix-shift, some of the o ld problems may come to be designated as unscientific, and effectively cease to exist as problems, whi le others previously so labelled come to assume significance. The revolut ion changes the standards by w h i c h the profession determines what counts as admissible problems and legitimate solutions. The boundaries between science and metaphysical speculation change. The relationship between science and Nature is knowledge-mediated, changing as matrices die out and are replaced by new ones. The concept of "element", for instance, has not remained the same throughout history. A n y concept gains its full significance only when related to other scientific concepts, procedures, and applications; it is thus context-dependent. In this sense, discovery is a complex process (as opposed to an event); it involves becoming aware of something new, but not fully understood, as w e l l as further experimentation w h i c h defines the properties of that w h i c h has been newly found. It is a situation o f not on ly finding that something is, but also what i t is. Discovering that and what together only occurs w i t h i n an established ma t r i x . 5 9 What is fundamentally at stake here is the correspondence theory of truth. The question at hand is whether or not theories correspond to an external, m i n d -independent w o r l d . Certainly, any Kant ian brand of epistemological programme, i.e., one w h i c h specifies the presence of an a priori structure, acting i n concert w i t h a sensible "given", necessitates the rejection of a simple empir ical correspondence theory. Kuhn's contention is that this latter not ion of truth must vanish, a long w i t h foundationalism i n general. Nevertheless, whatever replaces it w i l l st i l l require "a strong conception of t ru th" 6 0 : we must learn to get along without anything at all like a correspondence theory of truth. But something like a redundancy theory of truth is badly needed to replace it, something that wi l l introduce minimal laws of logic (in particular, the law of non-contradiction) and make adhering to them a pre-condition for the rationality of evaluations...On this view, as I wish to employ it, the essential function of the concept of truth is to require choice between acceptance and rejection of a statement or a theory in the face of evidence shared by a l l . 6 1 To declare a statement a candidate for a judgment of truth or falsity is to accept it as part of a "language-game", the rules of w h i c h forbid the assertion of both a statement and its contrary: v iz . , the l aw of non-contradiction. A n y attempt to conduct discourse against this rule endangers the integrity of the scientific lexical community, contributing to the breakdown of discourse. Language-games conducted without the benefit of logical rules reduce to metaphor, poetry and SSR, pp. 6-7, 55-7, 66, 103, 140-1; CGK, pp. 2, 12; ET, p. 338. RSS, p. 6. Ibid., p. 8. 49 myst ic ism. Adherence to logical rules is thus a presupposition of the sort of discourse demanded by normal science. In order to decide i f a statement is to qualify as scientific, one must first determine whether or not it is a truth-value candidate-whether or not it makes sense to attempt to judge it as either true or false. The answer w i l l be determined by the constraints of the lexicon embodied i n the discipl inary matrix. It is through this lexicon that one assesses the rational assertability of the statement. For this, "something l ike the normal rules of evidence" w i l l be employed. A statement awarded truth-value candidacy i n one matrix may therefore not enjoy the same status i n another. What counts as evidence for an assertion i n one case may not in the other. This w i l l serve to deepen the sense of crisis at the t ime of impending matrix-change, since previously viable statements re-cast i n the new lexicon may appear s imply to be the expression of unscientific nonsense. This is w h y the mul t i -l ingual historian of science must clearly take it upon herself to remember i n w h i c h lexicon she is speaking at any one time, on pa in of severely misunderstanding assertions made i n historical contexts, and thus running the clear risk of 62 caricaturing the scientists of the past. What has been fashioned, then, i n this account, is an "mrra-theoretic" use of the not ion of "truth": members of a scientific communi ty " w i l l generally agree w h i c h consequences of a shared theory sustain the test of experiment and are therefore true, w h i c h are false as theory is currently applied, and w h i c h are as yet untested". This understanding of truth eschews any sense of "metaphysical realism" w h i c h one might be tempted to attach to the assessment of theories w h e n looking at them from the outside. The tradi t ional questions of metaphysics are out of place here because there is no theory-independent w a y by w h i c h to determine what is "really there". W e see this i n moving historically from Aristotel ian to Newtonian to Einsteinian mechanics: there is nothing identifiable as a coherent direct ion of ontological development. The vis ion of successive theories approximating ever closer to the truth is an ontological one. But justification, rather than a iming at a goal external to the historical situation, seeks to improve the available problem-solving tools w i t h i n that situation. The w ay i n w h i c h the historian ought to approach historical theories, then, is dictated by hermeneutics (an attitude and process to w h i c h I w i l l return i n Chapter Three), w h i c h invokes a s imilar pattern to that seen w h e n the proponents of competing theories struggle to understand each other across a revolutionary divide. Looking inter-theoretically, the historian of science cannot judge either theory of an historical pair to be true: he cannot justifiably consider even the later, victorious, matrix to be a better approximat ion to the truth than the earlier, vanquished one. To do so w o u l d risk being ethnocentric, or Whiggish . This becomes clear when we remember the incommensurabil i ty w h i c h separates alternate orders: the lack of "a neutral language adequate to the comparison of...observation reports". Thus, looking back, we cannot say that Galileo's observation of a pendulum exhibited a more accurate or objective v i ew than Aristotle's observation of a fal l ing stone. In ib id , p. 9. 50 considering our sense of progress, "[w]e may...have to rel inquish the notion, explicit or implici t , that changes of [matrix] carry scientists and those w h o learn from them closer and closer to the truth". It is thus not open to us to v iew science as an evolution towards anything, even though "[w]e are a l l deeply accustomed to that goa l " . 6 3 That historical disciplines require description i n their o w n terms, as opposed to being imported into our current context, is underscored by the fact that modern disciplines have not evolved one-for-one from earlier disciplines. For example, even though we may say that, i n Hel lenic society, science and philosophy were one, there was no enterprise quite classifiable as either science or philosophy, w h i c h makes it very difficult for us to conceive of a single discipl ine represented by their un ion . A rather stubborn adherence to an out-going matr ix cannot, therefore, be deemed a failure to admit error i n the face of a confrontation w i t h proof. As I have already noted above, matrix-resistance or -change cannot be justified by an appeal to proof. Resistance to scientific change is neither i l logical nor s imply wrong, al though it may prove to be unreasonable. A revisitation of Kuhn's description of matrix-shift illustrates this: At the start, a new candidate for [matrix] may have few supporters, and on occasions the supporters' motives may be suspect. Nevertheless, if they are competent, they wi l l improve it, explore its possibilities, and show what it would be like to belong to the community guided by it. And as that goes on, if the [matrix] is the one destined to win its fight, the number and strength of the persuasive arguments in its favor wi l l increase. Most scientists wi l l then be converted, and the exploration of the new [matrix] wi l l go on. Gradually the number of experiments, instruments, articles, and books based upon the matrix wi l l multiply. Still more men, convinced of the new view's fruitfulness, wi l l adopt the new mode of practicing normal science, until at last only a few elderly hold-outs wi l l remain. And even they, we cannot say, are wrong. Though the historian can always find men...who were unreasonable to resist for as long as they did, he wi l l not find a point at which resistance becomes illogical or unscientific. At most he may wish to say that the man who continues to resist after his whole profession has been converted has ipso facto ceased to be a scientist. 6 4 In culture, K u h n remarks that the recognition of paral lel ism i n respect of the truth of different theoretical positions may lead to a sentiment of relat ivism. Indeed, it is precisely w i t h manifesting this sentiment that many of Kuhn's critics have charged h i m . But he counters that, i n science, the recognit ion o f this parallel ism, and the consequence that, i n some sense, competing theoretical groups may a l l be correct, is not merely relativistic. K u h n has attempted to describe science as a unidirect ional and irreversible process, where later theories are necessarily judged to be practical and explanatory improvements over earlier ones (considering a l l the values referred to above: accuracy, simplicity, scope, SSR, pp. 119, 170-1, 206; CGK, pp. 264-6, 277; RTC, p. 568; RSS pp. 7, 10. SSR, p. 159. 51 number of problems solved, etc.). If he has succeeded i n doing this, then he has also succeeded in salvaging some not ion of scientific progress, even i f this not entail the tradit ional not ion of an evolut ion towards some ultimate truth. Dur ing periods of normal science, the only viable way, i n fact, for the scientific communi ty to conceive of its work is as progressive. This is, i n part, due to the fact that, i n the absence of the competing schools of "pre-paradigm" science, researchers refrain from questioning each other's aims and standards. Certainly, there is no cal l to do so: the common matr ix has freed the communi ty from constantly re-examining its fundamental principles. (We might s imi lar ly identify a sense of progress, however, within each indiv idual school of the "pre-paradigm" period.) "In its normal state, then, a scientific communi ty is an immensely efficient instrument for solving the problems or puzzles that its [matrices] define": the results "must inevitably be progress". Doubts about progress are i n point only dur ing times of crisis. Adherents to the o ld matrix v iew the prospective one as threatening to deter or set back scientific progress. In contrast, those who promote a new order cannot see their programme as heralding anything other than progress. The rewri t ing of scientific literature and the re-casting of earlier matrices either as incorrect or as special cases of the new programme help to paint the tradit ional picture o f progress, w h i c h has led to the misreading of science as a cumulative process overall . K u h n speaks of a "cognitive evolution", w h i c h manifests i n the discursive exchange of, on the one hand, the statements w i t h i n a relatively stable communi ty (as i n normal science), and on the other, the statements of competing matrices (dur ing the crisis period wh ich spawns revolutionary science). He likens this to the exchange of genetic material w h i c h occurs i n biological evolut ion. Kuhn's is to some extent a Darwin ian v is ion: the scientific communi ty selects the fittest way i n w h i c h future science is to be practiced. The increased refinement i n theoretical articulation, and instrumental design and proficiency is punctuated by the relatively dramatic leaps of revolutionary selection. As Darwin proposed i n his theory of biological evolution, this scientific epistemological process seems to have occurred wi thout the benefit of a set goal (i.e., non-teleologically). But i n this case, that means without "a permanent fixed scientific truth, of w h i c h each stage i n the development of scientific knowledge is a better exemplar." This points to his description of his posit ion as "a sort of post-Darwinian Kantianism". Like the Kant ian Categories, the lexical taxonomy of w h i c h K u h n speaks works to supply the pre-conditions of possible experience. But Kuhn's lexical sort of Categories, unl ike Kant's Categories, transform w i t h the passage from one communi ty to another, whether that be through t ime (i.e., historical shifts of matrix) or across the divide separating contemporaneous disciplines. Because of the fundamental character of these Categories, their movabil i ty through revolut ion can result i n huge effects on the conceptions, practice, and achievements of science, even i n cases where only relatively small adjustments i n matr ix have been made . 6 6 Ibid, pp. 163, 166. Ibid, p. 172; RSS, pp. 11-2; Horgan, p. 44. 52 This shifting set of categories furnishes "a moving , historically-situated, Archimedean platform", the shifting a priori grounding of science. It is this, and not single scientific knowledge-claims, w h i c h is being evaluated i n theory-choice. The shared beliefs of a matrix already i n place provide an Archimedean point for the comparative assessment of competing theories. A n y overal l t ruth or falsity of these shared beliefs is irrelevant to the rationality of the evaluation. For the historian of science, too, the set of beliefs w h i c h define a matr ix provide a point from w h i c h she investigates that matrix. The details of actual scientific practice i n historical cases are not as important for the historian's understanding as is the perspective, or ideology, w h i c h she uncovers i n examining these cases. 6 7 Summary of the Relationship Between Normal and Revolutionary Science. In summation, the major point, perhaps, to be taken away from an encounter w i t h Kuhn's programme i n general is that our usual not ion of science is inaccurate. Revolutions are typical ly v iewed i n the light of simple additions to scientific knowledge, and the progression of science is seen as a much more smoothly cumulative process than it, i n fact, is. N o r m a l science is quite specifically the only phase dur ing w h i c h the scientific enterprise could be properly described as cumulative, al though it is true that normal science is what occurs most of the t ime. But the textbooks of science, and the philosophical and popular works model led on them function as a source of authority w h i c h "systematically disguises...the existence and significance of scientific revolutions". Wha t these tools are t ruly doing is recording the "stable outcome of past revolutions". They tend to supply a m i n i m u m of information as to h o w the current normal-scientific t radit ion came about. Textbooks are re-written, who l ly or partially, after a scientific revolution, since they are "pedagogic vehicles for the perpetuation of normal science". A prime influence here is the "omnipresent and perennial" temptation to "write history backward". This distorted picture is fostered by the tendency of science to depreciate historical fact, as w e l l as by the remarkable security enjoyed by scientific practitioners dur ing the stable periods of normal science, when a deep commitment to tradit ion is highly impor tan t . 6 8 A further covering of the tracks, as it were, occurs because scientists operating w i t h i n a later matr ix (say, a Newtonian one) often re-interpret the work of scientists operating wi th in an earlier matrix (e.g., a Gal i lean one) according to that later matrix. This further disguises the radical difference between their views. We see something of this sort going on i n Einstein's attempt to demonstrate that Newtonian dynamics was properly a special case of Einsteinian dynamics. This w o u l d have saved Newtonian dynamics, but w o u l d have done so by restricting it RSS, pp. 6-7. SSR, pp. 52, 96, 136-8, 160; CGK, p. 2. 53 to applications that satisfied much narrower conditions than were c la imed in i t ia l ly for the Newtonian programme. In order to v iew an out-of-date theory as a special case of its successor, the scientist must specifically transform it for that purpose. (Arguably, one could expect there to be a l imi t to the degree to w h i c h one matrix could be stretched to cover another, anyway.) But this v i ew effectively denies the occurrence of revolutionary periods i n the history of science; it promotes, instead, the cumulative model (against w h i c h K u h n is arguing), where in there is always only one matrix, w h i c h becomes progressively expanded and refined as science advances. Thus, work along lines such as this essentially prevents any established matr ix from being challenged by a prospective one, since it w o u l d prohibit scientists from speaking "scientifically" about any phenomenon w h i c h failed to fit w i t h the established matrix. A matrix could then provide no crit ical anomalies, and this w o u l d signal the end of the research through w h i c h science w o u l d develop. It w o u l d then appear to be the cumulative enterprise for w h i c h it has so commonly been mistaken. As such, it could no longer be considered a research-programme, and w o u l d have become merely a tool for engineering. But no active scientific-research matrix ever solves a l l of its problems. The radical and irreconcilable differences-which often show up i n problem-solving a b i l i t y -between successive matrices are necessary. Revolutions need not be instigated by crisis, but they almost always are. That crisis, however, also need not affect the same professional group w h i c h provoked it: "[n]ew instruments l ike the electron microscope or new laws. . .may develop i n one specialty and their assimilat ion create crisis i n another". Scientific revolutions need seem revolutionary only to those whose matrices are affected by them. The Copernican Revolut ion was one for everyone; the discovery of oxygen was one only for chemists. Upheavals may affect only a small community, several communities, or a l l of science. 6 9 Kuhn's remarks about scientific development are intended to be read as both descriptions and prescriptions; the descriptive and the normative are inextricably mixed . In order to qualify recognisably as mature science, the research endeavour must and does demonstrate an alternating relationship (although not a regular or predictable one) between the stable periods of normal research and the upheaval of the revolutionary crisis out of w h i c h a new normal order is eventually born: "extended periods of convergent research are the necessary prel iminary to" the relatively rare revolutionary shifts i n scientific t radit ion. In the in i t ia l emergence of a mature scientific discipline, theory and technique coalesce into a dominant set of concepts, practices, and instruments w h i c h consistently work together to supply concrete problems and solutions for the practice of the field. There should be room left for refinement and increasing sophistication. The range and precision of existing theory are enhanced. At this t ime, it is important for researchers to focus on the smooth operation of their matrix, and not on all the anomalies or incompletely understood phenomena SSR, pp. 92-3,181; CGK, p. 276. 54 w h i c h are lu rk ing i n the wings, i n order that they be able to mainta in some faith i n their o w n work through their having made some headway i n i t . 7 0 Since no matrix ever sees the solution of a l l o f its problems, there is never any research wh ich does not generate counter-instances to the theory, or anomalies, i.e., a potential source of crisis. The difference between normal and extraordinary science is not, then, that the former does not come up against counter-instances, whi le the latter does. N o r m a l science must encounter them i n order to pass into the transformative revolutionary phase w h i c h accelerates progress i n the discipline. Whether a problem is seen as a counter-instance to the theory or as merely a puzzle depends upon from w h i c h matr ix the problem is being considered. What proponents of the normal matr ix see as a puzzle can be seen, from the point of v iew of the prospective matr ix as an anomaly. What Ptolemaic astronomers attempted to continue v iewing as puzzles to be solved by their theory, Copernicus v iewed as anomalies. There is no sharp d iv id ing l ine between puzzles and counter-instances. This is what generates the vagueness at times of potential crisis as to whether the currently dominant theory w i l l continue to hold sway, or give way to the emergence of a new research-structure. A matrix-shift, once it occurs, accommodates, or assimilates, discoveries w h i c h had proven to be destructive to the o ld order. Through it, scientists become able~at least after a sufficient period of new-matrix-development has been a l l o w e d - to account for a wider range of phenomena than before, or to account for previously k n o w n phenomena w i t h greater precision. It is crucial to understand that no scientific theory can be declared inva l id i n the absence of an available alternative. To reject a matr ix without simultaneously substituting another is either to reject science altogether, or at least to make the inexplicable move of a l lowing it to return to a state of immaturi ty . Revolut ion also often contributes to the growth of science by generating new cognitive specialties, or fields of knowledge. As a result of the crisis, there may be a narrowing of the scope of a particular community 's professional concerns, increasing its extent of specialisation, and l imi t ing its capacity for communicat ion w i t h other scientific groups, not to ment ion heightening its al ienation from the laity. W e might term this a "lexical divergence", leading to a lexical heterogeneity of science i n general. This is fostered by the importance of the value of unanimity w i t h i n the communi ty : to mainta in this, the specialty may have to be subdivided. This illustrates the supreme significance of the approval of the scientist's work wi th in his specialist group. Specialisation and the nar rowing of the range of expertise are the price of increasingly powerful cognitive tools. Clearly, this process works against the unity of knowledge; but the pursuit of such uni ty may place the growth of knowledge at risk. So specialisation permits the sciences collectively to solve the problems of a wider range of natural phenomena than w o u l d be possible w i t h lexical homogeneity. Incommensurabil i ty is thus an important isolating mechanism, permitt ing overall scientific progress. 7 1 CGK, pp. 233, 237, 245; ET, pp. 227, 236. SSR, p. 170; CGK, p. 21; RSS, pp. 7-8. 55 The Distinction Between Science and Non-science Kuhn states that the arts and science cannot readily be distinguished "by the application of the classic dichotomies between, for example, the world of value and the world of fact, the subjective and the objective, or the intuitive and the inductive". "Close analysis must again be enabled to display the obvious: that science and art are very different enterprises or at least have become so during the last century and a half."72 Both science and the arts face persistent technical problems which must be resolved in the pursuit of their endeavours. Both are guided by aesthetic considerations, and are governed by established modes of perception. But too much attention to these similarities obscures their important differences. According to Kuhn, the artist's goal is to produce aesthetic objects, and he resolves technical puzzles to do so. The scientist's goal is the solving of the technical puzzle, and the aesthetic can be one of the tools for doing this. In fact, in order to qualify as a science, a discipline must operate according to the function of puzzle-solving, and this requires, as we have seen, an accepted matrix. In the sciences, despite the importance of aesthetics as a scientific value which helps to guide theory-selection, it is seldom an end in itself, the way it often is in the arts; and if it were an end, it would not be the primary one. Aesthetic considerations must always be subsumed to the unlocking of the puzzle at hand: "only if the scientist's aesthetic turns out to coincide with nature's, does it play a role in the development of science". Otherwise, research-styles and aesthetic responses tend to be more private, or individual, and often end up being eliminated from the finished, public work.73 Another difference is that, while artists generally have a public audience, the only true audience for a scientist is comprised of other scientists, and this audience becomes particularly narrow in the case of specialised sub-disciplines. A scientist generally seeks approval for her work strictly within her own specialised group. Despite the fact that scientific development demonstrates, more clearly than does the development of any other field, a succession of tradition-bound periods, punctuated by non-cumulative breaks, Kuhn nevertheless says that he has borrowed this notion from his observation of other fields, such as literature, music, art and politics. These other fields also show periodisation, from revolutionary breaks in style, task, and institutional structure. They also produce products which are modelled upon one another, rather than being produced in conformity to rules; we see the same sort of process in science, which operates primarily in accordance with set, concrete exemplars. ET, pp. 340-1. CGK, pp. 10-11; ET, pp. 342-3. 56 Thus, i n both the arts and science, the historian can discover stable periods (during w h i c h practice conforms to a tradition) and periods of relatively rapid change ( in w h i c h one tradi t ion gives way to another). In the arts, however, unl ike i n science, the success of one artistic t radit ion does not render a previous t radi t ion inva l id . Artists can, and sometimes do, voluntar i ly undertake dramatic changes i n style. Such changes i n the career of an ind iv idua l scientist are more rare, and are not voluntary, but rather are forced by acute internal difficulties w i t h i n her tradit ion, or by a successful innovat ion produced by someone else. In science, unl ike i n art, to change one's "style" is "to confess that one's earlier products and that of one's masters are wrong" . So art can support, far more than science, a number of simultaneous, incompatible traditions or schools. As a consequence, when controversy arises, it is solved far more rapidly i n science than i n a r t . 7 4 This is part icularly true of science at a mature stage, by w h i c h t ime a dominant research-matrix has emerged. The past products of artistic activity continue to be vi ta l parts of the artistic scene. O l d scientific works, on the other hand, are generally on ly read by historians of science. As scientific orders are succeeded by subsequent ones, new breakthroughs initiate the removal of out-dated books and journals. "Unl ike art, science destroys its past." 7 5 The internal crises of science w i l l be much more intense than those i n art. This is as we w o u l d expect, given the importance i n science of allegiance to some sense of truth, the relative dependence upon a guiding matrix, and the pressure to achieve pragmatic results, i.e., to develop theories w i t h sufficient explanatory power to solve puzzles. Crisis i n science signals a real need for innovation, and directs the attention of scientists towards the area where fruitful innovat ion may arise. It must be remembered, however, that innovat ion is not i n itself a prime value for scientists, and i f pursued for its o w n sake, w o u l d be condemned by a scientific research communi ty w h i c h is necessarily to some degree conservative. Innovation really only is admitted on the heels of crisis. A scientific avant-garde w o u l d threaten the steady work of science w h i c h is what occurs most of the t ime. So innovat ion is an often reluctant response to challenges to the dominant matrix. In contrast, artists do make innovat ion a pr imary value; even though the avant-garde may not immediately find institutional expression, it is sti l l more appreciated than it is i n the sciences. Artist ic work is not required to progress i n the same steady fashion w h i c h is crucial to most scientific research. So it is normal science w h i c h most clearly distinguishes science from other enterprises, since the stable periods of science are more markedly so than are any w h i c h appear i n the arts. To say that science is cumulative and art is not is to mistake the developmental pattern i n both fields, al though the relatively greater cumulative nature of science does indicate the relatively lower value placed upon innovat ion i n the sciences. W e can sum up the differences of science from other fields as follows: ET, pp. 348-9. Ibid., p. 345. 57 1. There is a relative scarcity of competing schools i n the sciences, due to the fact that science has achieved a level of maturity w h i c h other fields have not attained; i t is this w h i c h renders science more capable of regarding itself as ahistorical, and w h i c h strengthens its stable periods to the point where the prospect of change creates greater tension. 2. Science is more insular than other fields. 3. Puzzle-solving is a pr imary goal i n science; aesthetics is secondary. 4. Science has its o w n set of specific values, as does any area o f culture. Despite a l l of this, however, we do not generally identify necessary and sufficient conditions for any given discipline, inc lud ing science. Instead, we recognise a group's activity as, for example, "scientific", partly by its resemblance to other fields w h i c h we already have come to recognise as scientific, and by its difference from the activities o f other, non-scientific, d iscipl inary clusters. Partly, a particular discipline is identified by its posit ion i n the semantic field w h i c h contains a l l these disciplines. The End of Science Despite the facilitation of scientific progression granted by the pattern of mutual ly enhancing alternation between ordinary and extraordinary research, there is nothing to guarantee that the process w i l l continue indefinitely. Speaking of science, K u h n says, There was a beginning to it....There are lots of societies that don't have it. It takes very special conditions to support it. These conditions are now getting harder to find. Of course it could end. 6 He indicates that science could also w i n d d o w n i f scientists, even given adequate resources, failed for whatever reason to make further headway. If we keep i n m i n d Kuhn's rejection of a correspondence-based construal of truth, the door is left more open to a pragmatically oriented understanding of science. K u h n enunciates this himself: I think this way of talking and tiiinking that I am engaged in opens up a range of possibilities that can be investigated. But it, like any scientific construct, has to be evaluated simply for its utility-for what you can do with i t . 7 7 Were the usefulness of the special activities w h i c h constitute science to prove exhausted at some point, then, the particular picture of knowledge to w h i c h we have become accustomed i n this culture w o u l d presumably give way to Horgan, pp. 44, 46. Ibid. 58 something very different. This, i n itself, w o u l d be a dramatic i l lustrat ion of Kuhn's point: a radical shift i n a priori structure, al though of a more general order than the sort seen w i t h i n science itself. 59 3 Allies or Adversaries: the Troubled Meeting of Rorty and Kuhn Nature, Mind, and Correspondence Theories The most obvious area of agreement l ink ing Rorty and K u h n lies i n their rejection of the sort of "traditional" model typified by correspondence theories of knowledge. I have referred to this i n point one of the list, included i n m y introduction, of similarities w h i c h Rorty sees between himself and K u h n . 1 Neither Rorty nor K u h n wants to promote the v iew where in knowledge is gained through the mind's achieving an accurate representation of a mind-independent, external reality. Both w o u l d recognise that such representation w o u l d probably require the discovery of a pure observation-language, i.e., one free of significant perspectival influence, the existence of w h i c h both deny. Rorty observes that the tradi t ional v iew entailed that procedures of justification, i n order to be truly rational, must lead to the truth, to correspondence to reality. This picture is a straightforward metaphysical-epistemological one, antithetical to both Rorty's post-modern outlook of shifting, criterionless discourses and Kuhn's basically Kant ian v i e w where in the w o r l d is, i n part, constituted by an a priori template. What they are rejecting, i n particular, is the m i n d as blank slate-Locke's tabula r a s a - w h i c h receives ideas as sensations from a w o r l d outside the m i n d . But despite this commonali ty, they repudiate the correspondence mode l for strikingly different reasons. Rorty cannot entertain any framework w h i c h makes use of a not ion of the mind , so long as he advocates the abolishment of this not ion altogether. Since he takes himself to be removing al l elements w h i c h might serve as a foundation for epistemology, he is also hostile to the not ion of there being anything at a l l w h i c h might viably count as knowledge. K u h n , on the other hand, shows no interest whatsoever i n throwing out either the not ion of m i n d or of knowledge. I believe that K u h n w o u l d see it as a pla in v iola t ion of common sense to argue against m i n d and mental experience. This is reflected i n his assertion that we have no elements more basic than sense-data reports, such as "green there". This is not to say that we need impute to K u h n a metaphysical belief i n m i n d as mental substance, or sense-data as mental entities. K u h n - a g a i n , l ike Rorty, al though for different reasons-eschews the making of metaphysical claims i n epistemic talk. For Kuhn , metaphysical claims w o u l d reflect an expectation of our having knowledge of a mind-independent w o r l d . Since, for h im, our knowledge is 1 See pp. 1-2. 60 to some degree a world-consti tuting affair, metaphysical claims are totally out of place. There is no theory-independent way to determine what is really there. A w o r l d not understood through a particular mental template is an unintell igible one. But K u h n does not want to abandon entirely the not ion of a w o r l d "out there"; this is evidenced by his insistence that the "given"-however nebulous it may be wi thout the application of a template- is an essential element of experience, and thus of knowledge (the template itself being the other essential element). H e understands Nature as "giving" the stimulus, w h i c h is determined into a sensation by means of being perceived through a human template of some sort. This leaves discovery as part of what is involved i n the acquisit ion of knowledge. Therefore, for K u h n , the question cannot merely be one of language 2 , as it w o u l d be for Rorty, i f the latter were even to a l low for something as robust as knowledge. This relates back to point two i n m y introductory list o f putative similarities between Rorty and Kuhn ; Kuhn's retention of discovery as an essential part of knowledge-acquisit ion clearly speaks of a commitment to the not ion of observation, i n contrast to Rorty's rejection of this notion. Rorty's re-conception of certainty as a matter of conversation between persons, rather than a matter of interaction w i th non-human reality w o u l d be far too weak a thesis for K u h n . Without some sense of certainty conferred by their research-matrix, scientists w o u l d be unl ikely to develop the commitment to that matr ix w h i c h is necessary for normal scientific research to proceed. The matr ix helps them to formulate their understanding o f Nature. Dur ing a revolution, that sense of security is shaken, but it w i l l inevitably (for as long as science lasts) be restored by convergence of the communi ty around a new dominant matr ix i n w h i c h scientists' sense of certainty w i l l come to be invested. Scientists must have some convict ion where the most basic of their professional beliefs are concerned. But even though these procedures of what we might cal l social justification (wi th in the scientific community) are employed to delineate what is to count as knowledge, they are not invoked i n a climate devoid of input from the natural w o r l d . Further, to say that social justification is a matter of conversation implies that it is less structured-less grounded i n a conceptual and methodological f ramework- than the way i n w h i c h K u h n believes it to operate i n the matrix-based system of modern science. Even where conversation is understood to be guided by a set of conventions, those conventions are more easily subject to change (i.e., w i t h less resistance and dependence upon a cri t ical need for change, and w i t h considerably less sense of crisis) than i n matrix-changes i n science. The scientific matr ix is relatively more robust than are conversational conventions. This point renders Rorty's belief as expressed i n point five of m y introductory list of putative similarities inaccurate. In contrast, Ror ty rejects the metaphysical on account of his ult imate rejection of any w o r l d at a l l beyond the various and specific l inguistic patterns w h i c h make up the plural i ty of discourses; he takes the stance that the "given" is nothing but empty and misleading baggage, and that a l l attempts to make sense 2 See p. 47. 61 of any sort of w o r l d "out there" have proven utterly fruitless. Consequently, and on the basis of Rorty's rejection of the input of any special mental processes, we arrive at the posit ion where the w o r l d is nothing at a l l more than a linguistic creation. W h i l e for Kuhn , the w o r l d as we k n o w it is the product of a process w h i c h includes both discovery and creation, Rorty's w o r l d (not something w h i c h can really be said to be known) is strictly a creation. K u h n has remarked on the recent populari ty of conceptions of this sort, p lacing them under the rubric of proposals advocating a mind-dependent wor ld , although, for obvious reasons, the lexicon of Rorty's philosophy w o u l d have to be adjusted somewhat to accommodate the sort of v iew w h i c h K u h n is dr iv ing at: a human- or language-dependent wor ld , perhaps. The question of scepticism w h i c h Rorty finds so t roubl ing as a legacy of correspondence theories of knowledge, is not an irritant i n his o w n programme, because mind , wor ld , and a l l considerations of knowledge fall away, anyway. For Kuhn , the situation is somewhat different: we w o u l d not expect scepticism to appear as a problem haunting any programme w h i c h defined knowledge as being part ial ly constituted by a priori mental components. This Kant ian sort of picture is not conducive to wor ry about whether or not we can know the w o r l d as it real ly is, independent of our minds. Because the known w o r l d is not mind-independent, we do not have to be concerned about h o w we could possibly bridge a gap between it and our minds. One might advance to the question of by what or whom this Ror t ian w o r l d is created. Clearly, it cannot be a creation of the mind , for the m i n d is barred from the Ror t ian picture. W e seem to be left w i t h some sort o f materialistic neuro-psychological framework of understanding w h i c h Rorty insists that we refrain from v iewing i n ontological terms. W e are instructed to treat this as a metaphor for our discourse on what remains of experience i n the Rort ian v iew. But one does not need to have ontological sentiments i n order to have a w o r k i n g reference to mind , or to the ideas and raw feels connected w i t h i t . Even i f we do not want to posit m i n d as an ontological entity, it w o u l d nevertheless probably strike K u h n as counter-intuitive to avoid mentalistic talk completely, and resort to stimulus-oriented talk, w h i c h is set at a remove from experience, and w h i c h is the product of elaborate theory; we do not have direct access to s t imul i . Kuhn's belief i n the primit ive nature of the phenomenal ("green there") supports this. He prefers to talk i n terms of experience (and observations), rather than of the theoretical elements w h i c h s t imul i are. Rorty's preferred discourse-centr ing on stimulated C-fibres, for instance, and i n w h i c h phenomenal experience is denied al together-operates at a strictly theoretical level, one w h i c h is at a remove from experience itself. Consistent w i t h his retention of the mind , then, and i n clear opposit ion to Rorty, K u h n preserves the not ion of "something l ike a mental representation" 3 , w h i c h we understand from our vantage-point inside a communi ty to mediate our interactions, and those of the rest of our community, w i t h the environment. K u h n 3 See p. 48. 62 also speaks of the perspective from outside a communi ty , from w h i c h we see it adapting to its environment. I w o u l d expect that mental representations w o u l d be involved i n this perspective, as we l l , since observation is involved. Rorty describes W i l l a r d van Orman Quine as rebuking such writers as K u h n "for want ing to drop the not ion of observation altogether". 4 W h i l e Rorty disagrees w i t h Quine's belief i n the wor th of retaining this notion, it does not appear that he takes any issue w i t h Quine on the point of the latter's reading of K u h n as want ing to drop it (assuming that Rorty has himself read Quine correctly). But i f we can conclude from this that Rorty, l ike Quine, sees K u h n as want ing to drop the not ion of observation, then it appears that Rorty has not only overlooked an important feature of Kuhn's Kant ian project, but also has rendered K u h n far more distant from the tradit ional epistemological picture than he, i n fact, is. Overcoming Epistemology Rorty re-casts the terms "knowledge" and "objective" ( if they are to be retained at all) as applicable to areas of discourse where unforced agreement is relatively frequent and easy to obtain; the terms "matter of opin ion" and "subjective", i n contrast, are to be applied to areas where unforced agreement is relatively infrequent and achieved w i t h greater difficulty. Therefore, it is not surprising that one of Rorty's few criticisms of K u h n is that he should have discarded the epistemological project altogether, rather than seeking, as he d id , an alternative epistemology to the tradit ional variety. But Kuhn's commitment to a programme w h i c h reserves due place for epistemic considerations~as opposed to Rorty's, w h i c h emphatically does not~reflects a sentiment that many of us w o u l d find it too ja r r ing to abandon: that we want to be able to believe that we can know what the w o r l d is l ike, even i f our beliefs i n this regard are inescapably subject to change. Hence, for most of us~and apparently Kuhn , as wel l~the more tradit ional picture of selves inhabi t ing a w o r l d about w h i c h they justifiably struggle to learn something-such learning culminat ing i n knowledge of that w o r l d - h a s greater viabi l i ty than does the picture w h i c h Rorty paints. This more tradit ional v iew is also part icularly important where the normal operation of science is concerned, since it forms part of the force of matrix-commitment w h i c h inspires and fuels research w i t h i n a given conceptual framework. Rorty complains that " K u h n grants too m u c h to the epistemological tradition", this being revealed i n the latter's suggestion "that the philosophy of science has quite a distinct mission from the hermeneutic activities of the historian of science". 5 K u h n distinguishes the historical enterprise from that w i t h w h i c h the philosophy of science typical ly concerns itself, i n that "[ujnl ike history...[the phi losophy of science] is comparatively little concerned w i t h the 4 Rorty, PMN, p. 225. 5 Ibid. p. 340n20. 63 temporal development of theory, emphasizing instead the theory as a static structure, an example of sound knowledge at some particular, though unspecified, t ime and place....Philosophy's business is w i t h rational reconstruction, and it need preserve only those elements of its subject essential to science as sound knowledge". 6 This is to say that philosophy is generally concerned w i t h what counts as knowledge, and not w i t h how what counts as knowledge has changed, or even that it has changed. For Rorty, this has the unfortunate consequence of preserving "intact the myth that there is something called 'the nature of sound knowledge' for philosophers to describe, an activity quite distinct from what counts as justification wi th in the various discipl inary matrices constituting the culture of the day". 7 But for Kuhn , al though philosophy w o u l d do w e l l to take some lessons from history, a historically enlightened philosophy of science w o u l d sti l l occupy itself w i t h the search for an account of what is to be accepted as sound knowledge. Kuhn's observation that criteria of knowledge have transformed, sometimes dramatically, over t ime does not jeopardise the val idi ty of an epistemic orientation per se. This consideration highlights the fact that, whi le both Rorty and K u h n regard sensitivity to historical circumstance as important 8 , their respective understandings of the overall context i n w h i c h one is to express this historical sensitivity are radical ly divergent. For Kuhn , this awareness ought to inform our approach to an understanding of what it is to have knowledge; for Rorty, such a project is misguided and counter-productive. W e are also brought again to the inaccuracy inherent i n Rorty's belief as expressed i n m y introductory point five. The Need for an A Priori Framework Experience and the knowledge-acquisit ion w h i c h may fol low from it reflect, for Kuhn , a process w h i c h is part discovery (taking account of the "given" from Nature) and part creativity (taking account of a priori matrices). The matr ix is the movable template through w h i c h whatever is "given" is rendered intell igible experience. The learning of a matr ix w h i c h determines sensations i n this way is not just a matter of learning to assume postures-or make linguistic utterances~in response to s t imul i . It is learning how to see or hear or feel s t imul i as particular sensations. As K u h n has advised, we do not just discover that something is (and later interpret what it is), but rather discover what something is from the start. The matrix, then, plays a crucial role i n the generation of facts, as we l l as of beliefs about what sensations one has experienced. One of Kuhn's points w i t h respect to the role of the matr ix relates precisely to this situation. The matr ix determines, and thus makes possible, the 6 Kuhn, ET, p. 14. 7 Rorty, PMN, p. 340n 20. 8 See point eight in my introductory list. 64 observations w h i c h scientists make. One of the facets of Kuhn ian incommensurabil i ty is that alternative matrices w i l l generate alternative sets of observations, not a l l of w h i c h w i l l be commensurable. Observations w i l l be a part of what is different between alternative discourses. Different matrices tell us different things about Nature. Evidence plays a considerable role i n the acceptance or rejection of a theory. 9 K u h n suggests that, i f we subtract accuracy-of-fit-to-Nature from the set of scientific values, "the enterprise that results may not resemble science at a l l " . 1 0 So K u h n is clearly not advocating that we dispense w i t h observation-talk. 1 1 Ror ty m a y be surmising that K u h n is so advocating, on account of the latter's rejection of a single, overarching, neutral observation-language. Rorty's observation that we have so far failed to identify causal mechanisms l ink ing raw feels to neural-stimulation events is echoed in Kuhn's remarks on the insufficiency of our understanding of how the m i n d works i n perception. St i l l , K u h n - u n l i k e Rorty- takes the general posit ion that there is a route from stimulus to sensation, and that the presence of such a pathway (acted upon as it is by the influences of genetic predisposition, education, and socialisation) is what allows us to make sense of the idea that different matrices can generate different facts or beliefs. Rorty's dismissal of a generalised framework of thought and experience leads h i m to emphasise particulars~"from the sound of a w o r d through the color of a leaf to the feel of a piece of sk in"~which can "dramatize and crystallize a human being's sense of self-identity". 1 2 But the Kuhn ian point w h i c h speaks against the abil i ty to do this is that there can be no particulars without the presence of a general framework by means of w h i c h particulars are picked out i n the first place. A given Gestalt is set by a matrix, and particulars are fragments of a specific Gestalt, w h i c h must be set before particulars can be picked out of it. If the Gestalt shifts, the character of identifiable particulars may also change, and possibly radical ly enough that what were distinct particulars before are no longer, whi le new ones emerge. Rorty appeals to imaginat ion as a force w h i c h generates a particular discourse, and habit as one w h i c h sustains i t for a t ime as the status quo. In a Kuhn ian scientific wor ld , imaginat ion and habit no doubt have a role of some magnitude or other, but the story runs much deeper than that. A l l the values w h i c h I have mentioned i n Chapter Two-prob lem-so lv ing abilities, scope, etc.~ come into play here much more strongly than do these relatively superficial elements w h i c h Rorty employs to describe the bir th and success of discourses. Rorty asserts that the justificatory process cannot operate outside of reference to already-accepted beliefs and language-practices. So it w o u l d seem that the structure of the "web of beliefs" works to justify what further beliefs can 9 See quote, p. 49. 1 0 Kuhn, ET, p. 331. 1 1 See introductory point two. 1 2 See p. 13. 65 be incorporated into that web, whi le retaining coherence. But n o w it looks as though we are approaching being able to abstract a set of criteria of intel l igibi l i ty from the body of the discourse, or at least to identify similarity-relations, upon w h i c h it seems to me that a "web" w o u l d have to depend. Specific sets of similarity-relations are what are required to set discourses off from each other. Indeed, Rorty wants to be able to judge "particular present situations and options as similar to or different from particular past actions or events". Further, he describes his version of pragmatism as geared towards replacing the not ion of true beliefs w i t h "successful rules for action" (my emphasis) . 1 3 But i f this is what we end up wi th , then Rorty's system is beginning to look more l ike Kuhn's, making Ror ty Kuhn ian i n a w a y w h i c h he w o u l d not want to be, and w h i c h he w o u l d not even recognise as Kuhnian . It is difficult, though, to p in d o w n precisely to what degree Rorty is Kuhnian i n this way, since he both wants to mainta in something like an evaluative framework (albeit an unhelpfully vague one) through w h i c h assertions i n a discourse can be justified, and yet disparages anything w h i c h might resemble criteria (presumably inc luding the less-abstract similarity-relations), favouring a sort of l ight ly structured (if structured at all) poetic, self-creative urge. Rorty underlines this by describing himself as "auxil iary to the poet rather than to the physicis t" . 1 4 A l l of this renders any similari ty between Ror ty and K u h n on the issue of an accepted conceptual and practical f ramework 1 5 considerably less l ikely. Fo l lowing this poetically oriented line, Rorty states that [t]he craftsman typically knows what job he needs to do before picking or mventing tools with which to do it. By contrast, someone like Galileo, Yeats, or Hegel (a "poet" in my wide sense of the term-the sense of "one who makes things new") is typically unable to make clear exactly what it is that he wants to do before developing the language in which he succeeds in doing it. His new vocabulary makes possible, for the first time, a formulation of its own purpose. It is a tool for doing something which could not have been envisaged prior to the development of a particular set of descriptions, those which it itself helps to provide. But I shall, for the moment, ignore this disanalogy. I want simply to remark that the contrast between the jigsaw-puzzle and the "tool" models of alternative vocabularies reflects the contrast between...the wi l l to truth and the wil l to self-overcoming. Both are expressions of the attempt to represent or express something that was already there and the attempt to make something that had never been dreamed of before. 1 6 But it is difficult to see what w o u l d motivate someone towards the construction of a new vocabulary i f he had no idea what he was t rying to do. Since new Kuhn ian matrices arise out of the soil of the o ld , some guidance i n the construction of a new order w i l l be provided by the o ld framework. Also, the previously mentioned (Chapter Two) general values w h i c h guide theory-choice contribute to the Rorty ORT, p. 65. Rorty, CIS, p. 8. See again introductory point five. Rorty, CIS, pp. 12-13. 66 scientist's sense of what needs to be done; foremost among these is the a im of problem-solving. W h i l e some of the problems up for discussion w i l l disappear subsequent to a matrix-shift, and others not yet encountered w i l l develop, many problems w i l l subsist through the change. Leaving aside Yeats and Hegel, we can see that it may be somewhat premature to expect somebody such as Gali leo to have a full picture of what normal scientific research under his new regime w o u l d look l ike, at the t ime that he proposed it. St i l l , K u h n states that "[t]he m a n w h o premises a paradigm when arguing i n its defense can nonetheless provide a clear exhibit of what scientific practice w i l l be l ike for those w h o adopt the new v iew of nature". 1 7 Further research under the new order w o u l d serve to deepen, refine, and extend its "vocabulary". One can at least say w i t h some plausibi l i ty that Gali leo promoted a heliocentric system to answer concerns of calendar-reform, and to deal w i th the build-up of anomalies under the geocentric Ptolemaic system. Both of these demands provided specific motivations w h i c h w o u l d have inst i l led i n Gal i leo some idea of what his task was. This is not to say that imaginative input and "flashes of intui t ion" are alien to the bir th of a new matrix; K u h n has made us aware of their importance. But it is to say that there is greater structure informing even imaginat ion and intui t ion, where scientific projects are the focus, than we see i n the construction of poetry. If we must w i d e n the meaning of the term "poet" to the extent w h i c h Rorty seems to be advocating, we do so at the expense of valuable clarity. Rorty's poetically oriented account of discourse-alteration offers no explanation of why motives change i n the particular way that they do. Rorty's contention that "[t]he application of such honorifics as 'objective' and 'cognitive' is never anything more than the presence of, or the hope for, agreement among inquirers" 1 8 rings rather ho l low w h e n we begin to enquire w h y those areas of the culture w h i c h enjoy a greater degree of agreement-notably science, as K u h n has explained-have gotten to that point. The Kuhn ian argument is precisely that we have come to be able to expect to achieve relatively greater agreement w i th in the sciences (at least, dur ing the periods of normal stability w h i c h are by far the most prevalent periods) because of the presence of a commonly accepted a priori matr ical structure. So the advantage is conferred not only by the existence of a matr ical structure at a l l , but also by the fact that, i n a mature science, there w i l l be only one such structure operative at any one time, increasing scientists' sense of security i n their o w n work . Scientists work towards the facilitation of agreement by first striving to determine an appropriate structure for research, and then by seeking to prevent the decay of that structure at times when it comes under stress (as i n the recognition of an accumulat ion of anomalies). Hence, it is not merely a matter of chance, as it appears that it w o u l d have to be i n Rorty's estimation, that agreement is achieved i n specific areas of culture; rather, it is quite by design. Rorty asserts that we can explain the observations w h i c h scientists make by reference to their "psychologies and sensibilities"; we can explain their See p. 39. Rorty, PMN, p. 335. 67 propensities to react w i t h certain sentences to certain s t imul i by reference to their upbringing. Scientists have been "programmed" so as to respond to certain retinal patterns w i t h statements such as "[T]here goes a neutrino". 9 There are two points w h i c h need be made here. One is that, for Kuhn , there is clearly more to the question of "programming" of scientists than what we might cal l mere "psychologies and sensibilities", al though factors such as this w o u l d be expected to have their influence. What K u h n is interested i n elucidating, however, is that this "programming" is deliberately built in--and i n a quite sophisticated m a n n e r -to the t ra ining of prospective scientists, i n the form of the conceptual matr ix w h i c h w i l l define their future scientific research, and most especially i n the form of the exemplars w h i c h are offered to the student as concrete cases w h i c h illustrate the application of matr ical concepts. The stronger the presence of a collectively and deliberately designed framework, the less room for influence by indiv idual psychological idiosyncrasies. The second point is the crucial one, that this "programming" of w h i c h Ror ty speaks is, again, significantly reminiscent o f Kuhn ian matrices, and we are brought back round to the point that Rorty has not effectively escaped the structural tone w h i c h K u h n has maintained throughout his o w n thesis, whether Rorty has detected it or not. O n the issue of the presence of matrical elements, one might take Rorty's and Kuhn's respective positions to be similar, i n that we can substitute Ror t ian discourses for Kuhn ian matrices, or vice versa. Rorty, himself, views his discussion of discourses to be a generalised account of the more specific Kuhnian "discourses". Support for this v i ew might be assumed to be found i n both Rorty's and Kuhn's denial of criteria, or rules, as being important i n how we approach the w o r l d . But Rorty's discourses are more superficial than Kuhn's matrices. W e see this i n the miscibi l i ty of Rorty's discourses w i t h each other, as i n the sentence "If we had just stuck i n an electrode i n the right place i n the cortex, he w o u l d never have decided he was Napo leon" . 2 0 Here, we have a neuropsychologica l vocabulary side-by-side w i th a behavioural-psychological one. Kuhnian matrices, on the other hand, could not possibly co-exist i n the same communi ty at the. same time, for they represent a conception of Nature, and alternative conceptions of Nature cannot receive simultaneous commitment from a homogeneous group. It is clear from their relation to a sense of the way i n w h i c h the w o r l d is, that Kuhn ian matrices must run deeper than Rort ian vocabularies. As such, they w i l l be more significant i n guiding action. This leaves us, i n a Ror t ian wor ld , w i t h a severely weakened framework for belief, and hence for action. This weakening is reflected i n the fact that Rorty's dismissal of criteria for judgment is far more radical~and more comple te- than is Kuhn's . K u h n diminishes the importance of criteria really on ly dur ing times of matr ical stability, when the normal investigative process occurs through the construction of similarity-relations between things and events i n the w o r l d . Learning a part icular way of carving up the wor ld , so to speak, is achieved through particular instances 1 9 Rorty, ORT, p. 56. 2 0 Seep. 11. 68 of ostension more than through the learning of general rules. It is possible to abstract rules from the accepted beliefs and procedures of the matrix, although this is generally done after the fact of the construction of the matrix. The need for abstracted rules becomes much more evident during times of matrix-instability. This is clearly a far more subtle position than is Rorty's, where the acceptance or rejection of criteria seems to be an all-or-nothing affair, and no allowance is made for any sort of framework guiding the construction of even similarity-relations. Again, this underscores the inaccuracy of Rorty's evaluation of the similarity of Kuhn's programme to his own, as expressed in my introductory point five. Rorty's allowance-albeit a begrudging one~for the remote possibility of "innatist" views of a genetically "wired-in" language of thought reveals further potential for an embarrassing slide into Kuhnian territory. Possibly such a language of thought counts as part of that to which Kuhn is referring in his talk of the genetic elements which people share, which includes the common ground provided for through biological determination of lexical structure. This relates to the perceptual integrity of which Kuhn speaks, which is born of the selection process, driven by survival-pressure.21 Rorty's dismissal of the importance of our determining which elements of human behaviour would be directed by an innate genetic matrix, and which not, is further illustration of a regrettable lack of subtlety in his programme. Distinguishing that which is innate from that which is not can only help to advance our self-knowledge, and to yield improved discriminating-power with respect to what it is possible for us to alter in our thought and behaviour, and what remains more fixed. If we duly recognise the significant function of some sort of matrix through which we necessarily experience the world, then clarification as to which elements of that matrix are more fluid and which more solid proves to be of epistemic importance. This would be a matter of concern to Kuhn. Rorty's denial of the importance of the distinction between fixed and movable elements has led him into some unfortunate confusion. He warns us that allowance for any fixed language of thought should not slide into the premise that that language is immune to correction on the basis of experience. But we are not accustomed to viewing genetically fixed elements as subject to correction on the basis of experience, anyway, at least not directly. What will be experientially corrigible are the transformable elements of non-innate matrices, such as Kuhn describes for science: his movable Archimedean platforms. If Rorty is allowing for these, this again brings him closer to Kuhnian structure. Moreover, Rorty's reference to Fodor is evidence of his own recognition of the identifiability of at least some sort of a priori structure. Rorty seems to be attempting to further his argument against the tendency to formulate methodologies of any sort by objecting that the view that science operates on the basis of our knowing in advance in what sort of discourse an explanation of concept-formation shall be formulated "takes us around a rather See p. 46. 69 t iny circle". Certainly, K u h n has recognised that the way i n w h i c h science must operate is inherently circular: it cannot function wi thout first delineating what is to count as acceptable scientific questions and answers. It sets out its o w n a priori parameters before proceeding according to them; this is precisely what it is to have a matrix. But K u h n reminds us that not a l l circularities are vicious. Rationality and Solidarity One of the chasms on account of w h i c h it becomes evident that Ror ty and K u h n are ta lking past each other results from their strikingly different conceptions of rationality. This contrast seriously undermines the ways i n w h i c h Rorty sees himself and K u h n to be similar that I have noted i n m y introductory points three and especially four. This is yet another term w h i c h Ror ty has undertaken to define non-standardly. We are offered a frustratingly loose description of what it is to be rational, i n the form of a) the recognition of being part of one's o w n community, w h i c h derives from the common dictates of the communi ty and its language, and b) victory or persuasiveness i n argument. The latter rendering is something we w o u l d consider to be consistent w i t h Kuhn , since he does acknowledge the importance of persuasiveness i n theory-choice. But because K u h n defines "rationality" i n the standard manner, he does not come to equate it w i t h persuasiveness, as does Rorty, and is therefore able to cite both rational and non-rational influences on the persuasiveness of attempts to promote a new theory. Again , his characterisation achieves greater subtlety than does Rorty's. K u h n also maintains a strong commitment to tradit ional reasoning, w h i c h both must constitute the rational elements of the consideration of new matrix-candidates, and must figure prominent ly i n the internal logical workings of an accepted matrix dur ing periods of normal stability. So for Kuhn , the social practice cannot entirely define what is to count as rational, because rationality, al though to some degree defined specifically by the internal logic of a given matr ix (roughly ak in to a specialized case of Rorty's social practice), is also to some degree an element w h i c h enters into the choice of a particular social practice, or, i n more precise Kuhn ian terms, theory-choice (although not i n a form sufficiently complete to generate an algori thm). The scientist's recourse to persuasion does not suggest a lack of very good reasons for choosing a theory . 2 3 K u h n comments that "[t]o suppose...that we possess criteria of rationality w h i c h are independent of our understanding of the essentials of the scientific process is to open the door to cloud-cuckoo l a n d " . 2 4 Presumably, the same goes for suppositions that we do not possess standard criteria of rationality at a l l . Adherence to logical rules is a presupposition of normal scientific discourse. Rorty, ORT, p. 57. Kuhn, CGK, p. 261. Ibid., p. 264. 70 Rorty asks, "Is the sort of community which is exemplified by scientific inquirers and by democratic political institutions a means to an end, or is the formation of such communities the only goal we need?" He considers debates about Kuhn's "irrationalism" to boil down to this question.25 It seems that Kuhn's "irrationalism" relates to this issue on account of the presence of standard rationality pervading the traditional adjustment of means to ends. Refusing to see something in terms of this adjustment would have the beneficial effect of obviating standard rationality. But I believe that Kuhn would see scientific communities not as ends in and of themselves, but rather as instruments functioning to further science as an epistemic organ. The community exists for a quite specific purpose: to promote growth of the body of scientific knowledge. The maturation of a science to the point where one dominant community has replaced a plurality of competing schools furthers the realisation of this end, since the unified matrix of the mature scientific community need no longer expend energy in disputes over the fundamentals of the research-matrix. It is not here just a matter of a single scientific school demonstrating greater solidarity than a diverse collection of such schools simply for the sake of increased comradeship, which we may imagine to be the case in other sorts of endeavours, very different from the scientific one. So the scientific community described by Kuhn is clearly a means to an end. In fact, to some degree, it functions as an expression of the matrix. This latter point seems not so far-fetched, if we consider that matrical shifts, in re-organising and re-defining what is to count as scientific, may result in established members of the community being edged-out, on account of what might be seen as an obstinate tendency to remain faithful to a dying matrix, while a group of researchers perhaps new to the field, having proposed the new matrix, gains ascendancy in the field as that matrix becomes dominant. In this respect, the matrix, as the tool exercised by the dominant majority of scientists within the community (dominant since the ascendancy of this new matrix), can exert an effect upon the constitution of that community, even though the primary influence is probably in the other direction-i.e., wherein a specific community of practitioners is united by their disciplinary matrix, which then becomes an expression of the community. Rorty invokes Dewey as having argued (in speaking of morals) that we only know what we want after we have seen the results of our attempts to get what we once thought we wanted, and that the use of new means changes ends 2 6 Rorty sees the case of "post-positivistic philosophy of science" as analogous. It is probable that Kuhn would agree that Dewey's picture of morality corresponds plausibly with his own view of science. The fact that we are moved continually to adjust our ends-sometimes upon the introduction of previously unknown means-is perfectly consistent with the conception of science put forward by Kuhn. There are two processes going on here, in the statements which Rorty is attributing to Dewey. Firstly, a matrix may be required to shift as it becomes evident that its Rorty, ORT, p. 43. Ibid., p. 68. 71 problem-solving abil i ty is beginning to w a n significantly--it is fai l ing to y ie ld what was desired, and we see that the results of our attempts to get what we once thought we wanted are less than satisfactory. A shift ushers i n a new set of goals, at least to some degree, such that what we want has changed. The new matr ix is expected to provide what is desired more effectively than the o ld one proved able to do, but it is clear that what is desired may exhibit a quite different form after the shift from before. The new regime may br ing pressure to bear upon the scientific communi ty to hasten the development of new instrumentation w i t h w h i c h to gather evidence by means of w h i c h the new matr ix may be refined and extended. But alternatively, and this is the second aspect of what Ror ty is attributing to Dewey~new instrumentation may be devised w h i c h w i l l so revolutionise the phenomena available for discovery that the ends of research w i l l be altered to keep pace w i t h i t . The development of the electron microscope provides an example of this, and is cited by K u h n as an instance where an occurrence i n one scientific field has the effect of instigating a revolut ion i n another. In this way, the introduction of novel means can transform the ends w h i c h are sought. But this Dewey-Kuhn correspondence does not provide evidence i n favour of Rorty's characterisation of science as more l ike a paradigm case of mudd l ing through "than l ike a series of choices between alternative theories on the basis of observational results" 2 7 Kuhn's account of what must go on i n science is considerably more exact and structured than Rorty's v i ew of discourses. For Kuhn , alternative theories become necessary, often because observational results begin to reveal a progressive build-up of anomalies. Even dur ing a crisis-period, it does not seem appropriate to characterise what is going on as "muddl ing though", given that the threads of the out-going matr ix still provide some residual structure for conceptualisation, and that, at any rate, the w ay out of the state of relative chaos w i l l be aided by the constancy of the various scientific values w h i c h I listed i n Chapter Two . The further point might be made here that, even i f there is some adjustment to be made i n what we expect a theory to do for us, the adjustment of m e a n s - w h i c h really amounts to the adjustment of a methodology-is not equivalent to a lack of such methodology. Rorty also cites Dewey as saying that intellectual progress usually occurs through the sheer abandonment of questions: rather than solving them, we get over them. Kuhn's posit ion is again divergent: the very constant existence of some set of questions or other is vi ta l i n demarcating the matr ix of scientific research-the matr ix stipulates admissible questions. W e can make no sense of the crucial mot ivat ion of problem-solving without acknowledging the need to solve questions. W e might say that this very need is the engine w h i c h drives research. Even though revolutionary shifts i n matrix lead to the "overcoming" of some questions, other questions w i l l nevertheless survive the shift, and moreover, new questions w i l l be generated as a result of it. Ibid, p. 69. 72 Rorty states that what he calls the "objectivist t radi t ion" seeks an ahistorical human nature; it is this nature, as shared common ground, w h i c h is supposed to b ind humani ty together. The tradit ional not ion o f rationali ty figures prominent ly i n the general account of what it is to be human. But this "metaphor of inquiry and human activity generally" is one of convergence and unification; it is, for Rorty, inaccurate, because human activity demonstrates proliferative tendencies, and a movement towards diversification. W e need "to throw out the 28 last residues of the not ion of ' transcultural ra t ional i ty ". In this connection, he notes w i t h approval that, i n K u h n , the phi losophy of science becomes increasingly historicist and decreasingly logical . To h im, this is beneficial, since our understanding of w h y the new science was superior to Aristotle's work, and of the relations between the new science and mathematics, common sense, theology, and moral i ty is aided by our turning outwards, towards the social context of justification, instead of by our turning inwards, towards the relations between inner representations. Typical ly, the turn inwards has rel ied upon the employment of reason as a quintessentially human modali ty, one w h i c h unifies and defines a l l of humanity. In contrast, social contexts of justification as Rorty envisions them tend towards plural ism. St i l l , Rorty argues for the preservation of the intellectual, social, and pol i t ical "habits" w h i c h were nurtured by the Enlightenment, al though he does not want to see these justified by tradit ional conceptions of rationality and truth, but rather by "a conception of rationality as criterionless mudd l ing through, and by a pragmatist conception of truth". The sense of rationality w h i c h Rorty advises us to discard is that w h i c h directs us to be methodical , to have criteria for success la id d o w n i n advance. In Ror t ian terms, this is the stronger sense of rationality, of w h i c h science is generally taken to be the paradigm: we tradit ionally expect to have clear criteria for the success of scientific theories, by means of w h i c h their abil i ty to predict is augmented. Strong rationality is associated w i t h objective truth, correspondence to reality, method, and criteria. Rorty's re-envisioning of rationality leads us i n a rather different direct ion: towards mora l virtues, such as tolerance, respect for others' opinions, a will ingness to listen, and reliance on persuasion i n place of force. To be rational, then, is to be "sane" and "reasonable". Rationali ty is not the exercise of a faculty called "reason", a faculty w h i c h stands i n some determinate relat ion to reality. W e define the term more i n relat ion to being "c iv i l ized" than to being "methodical": "We should avoid the idea that there is some special virtue i n knowing i n advance what criteria y o u are going to satisfy, i n having standards by w h i c h to measure progress". Holis t ic pragmatism wants to ho ld on to the materialistic wor ld-v iew that typical ly forms the background of contemporary l iberal self-consciousness, whi le refraining from advancing the c la im that this v i ew has been established through a method. The formulat ion o f methodological principles is considered to be generally a waste of t ime, anyway, since one often ends up only w i t h "a string of platitudes, hooked up to look l ike an Ibid., p. 32. 73 a l go r i t hm" . In a v o i d i n g cr i ter ia, w e a v o i d d o g m a t i s m , defens iveness, a n d r ighteous i nd i gna t i on . 9 T h e r e are a coup le o f e lements here w h i c h I f i nd p rob lemat i c . F irst ly, the type o f p ic ture w h i c h R o r t y here paints, a n d w h i c h he imp l ie s is i n concer t w i t h Kuhn ' s v i s ion , is qu i te s tr ik ing ly ant i thet ica l to w h a t K u h n is a t tempt ing to s h o w . 3 0 W h i l e for K u h n , cr i ter ia, o r rules, are not the n o r m a l means b y w h i c h prospect ive scientists are t r a i ned in to the i r a d o p t e d sc ient i f ic t rad i t ion , it is poss ib le to abstract such cr i ter ia f r o m the mat r i ca l f r a m e w o r k after the fact o f the l ay ing -d o w n o f that f r amework . So, w h i l e cr i ter ia d o no t part ic ipate i n the b u i l d i n g o r c o m m u n i c a t i o n o f the matr ix , t hey can be m a d e sense of, i f w e w i s h to focus a t tent ion o n the ske leton o f the m a t r i x - a n e e d w h i c h w i l l b e c o m e m o r e apparen t w h e n the mat r i x appears i n d a n g e r o f c r u m b l i n g , i.e., at t imes o f crisis. A t a n y rate, the m a n i p u l a t i o n o f s imi lar i ty -re lat ions b e t w e e n th ings a n d events i n the w o r l d w h i c h occurs as the mat r i x is re f i ned a n d ex tended d u r i n g per iods o f n o r m a l sc ience is not a process w h i c h is c l ean o f ra t iona l j u d g m e n t s . So Rorty 's s tatement that "Kuhn ' s de fenders [ a m o n g w h i c h he counts h imsel f ] . . . typ ica l ly d r a w the l i ne b e t w e e n the ra t i ona l a n d the non - ra t i ona l soc io log ica l l y ( in te rms o f a d i s t inc t ion b e t w e e n per suas ion a n d force) ra ther t h a n m e t h o d o l o g i c a l l y ( in terms o f the d i s t inc t ion b e t w e e n possess ion a n d l ack o f exp l ic i t c r i t e r i a ) " 3 1 reveals a n i rony : K u h n h imse l f w o u l d not espouse the pos i t i on w h i c h his supposed defenders take. Second ly , R o r t y appears to be d i rec t i ng his p o l e m i c against a sort o f sc ient ist ic r o m a n t i c i s m w h i c h regards sc ient i f ic re search s o m e w h a t as a s h o w o f v i r tue i n the batt le o f know ledge a n d en l i gh tenment versus i gnorance a n d the chaos o f i r rat ional i ty. W h i l e K u h n c o u l d cer ta in ly be expec ted s im i l a r l y to t u r n a w a y f r o m this p ic ture o f sc ience, it is c lear, nevertheless, that he want s to m a i n t a i n a def in i te d i s t inc t ion b e t w e e n the att i tude a n d m e t h o d o l o g y o f sc ience, a n d the tone o f o the r cu l tu ra l e n d e a v o u r s . 3 2 So w o r k i n g w i t h a n a priori ma t r i c a l f ounda t i on , w h i l e not necessar i ly " v i r tuous " , is at least usefu l . E v e n a n u n d e r s t a n d i n g o f one's research as con t r i bu t i ng to a n e luc ida t i on o f N a t u r e does not necessar i ly suggest the sort o f m o r a l r ighteousness i n sc ience that R o r t y is h o l d i n g u p for scorn. T h e e m p l o y m e n t o f ra t iona l m e t h o d s does not enta i l , o r e v e n a lways reflect, the scienti f ic t radi t ion ' s h a v i n g m a d e " the na tura l scientist in to a n e w sort o f pr iest", e v e n i f w e v i e w her ro le as p r o v i d i n g " a l i nk b e t w e e n the h u m a n a n d the n o n - h u m a n " . 3 3 C o n s i d e r i n g the issue o f w h e t h e r h u m a n act iv i ty s h o u l d p rope r l y be character i sed as conve rg i ng a n d un i fy ing , o r pro l i fe ra t ing a n d d ivers i fy ing, f r o m Kuhn ' s po in t o f v iew, the ques t i on is not perhaps so s imple . In the evo lu t i on o f a scientif ic d i sc ip l ine into the m o r e m a tu re stage typ i f ied b y w o r k b e i n g d o n e u n d e r 2 9 Ibid., pp. 28, 36-7, 62, 65. 3 0 See again my introductory point five. 3 1 Rorty, ORT, p. 48. 3 2 See pp. 56-8, and the next section of this chapter; this issue also relates back to point seven of my introduction. 3 3 Rorty, ORT, p. 37. 74 a single dominant matrix, clearly we see a convergent pattern. But there is a corresponding proliferation, over t ime, as the matrices of scientific disciplines become more detailed, and as a result, more specialised and narrow i n their focus; K u h n has described how this tendency leads to single disciplines split t ing up to produce new sub-disciplines, w h i c h eventually become quite separate scientific communit ies i n their o w n right, complete w i t h their o w n distinct defining matrices. Rorty may be making the mistake of equating the tradit ional search for an ahistorical scientific conceptual and methodological framework w i t h the acknowledgement of the presence of such a framework per se. But this w o u l d provide a rather crude analysis, as an examination of Kuhn's writ ings indicates. The movable a priori w h i c h is a fundamental feature of Kuhn's Kant ian type of programme is clearly both sensitive to historical fluctuations and responsive to the need for a framework of some sort through w h i c h investigation is facilitated. Indeed, the use of a given framework, though it is stable only for periods of variable longevity, is, according to Kuhn , the n o r m . 3 4 Returning to the not ion of a transcultural rationality, we might identify a central feature of such a rationality as the process of induct ion. As we ought to expect, one of Rorty's complaints against K u h n is that K u h n "occasionally makes too large concessions to the tradit ion, part icularly when he suggests that there is a serious and unresolved problem about w h y the scientific enterprise has been doing so nicely lately". 3 He quotes K u h n as l ink ing this to the problem of induct ion: Even those who have followed me this far wi l l want to know how a value-based enterprise of the sort I have described can develop as a science does, repeatedly producing powerful new techniques for prediction and control. To that question, unfortunately, I have no answer at al l , but that is only another way of saying that I make no claim to have solved the problem of induction. If science did progress by virtue of some shared and binding algorithm of choice, I would be equally at a loss to explain its success. The lacuna is one I feel acutely, but its presence does not differentiate my position from the tradition. 3 6 Rorty argues that the desire to provide such an explanation as K u h n seeks "is one more result of hypostatizing the Platonic focus imaginarius-truth as disjoined from agreement~and a l lowing the gap between oneself and that uncondit ional ideal to make one feel that one does not yet understand the conditions of one's existence". 3 7 But the point, for Kuhn , is not to understand the conditions of one's existence, but to seek to understand the conditions of knowledge, and specifically, of scientific knowledge. The desire to explain the success of science, wh ich , at its root, entails the desire to solve the problem of induct ion, does not entail a See again point five in the Introduction. Rorty, PMN, p. 340 Kuhn, ET, pp. 332-3, quoted in Rorty, PMN, p. 340. Rorty, PMN, p. 340. 75 disjoining of truth from agreement. Indeed, it is the desire to explain how such agreement might reliably be reached at a l l . It appears, i n this particular instance, that Ror ty is impor t ing a correspondence-view into the question, imput ing to K u h n a certain degree of slippage back into that territory from w h i c h he has moved away. But K u h n need not be, and I th ink it is clear is not, influenced by correspondence-theory models, i n his concerns about the problem of i n d u c t i o n . 3 8 Rorty argues that Kuhn's wor ry about t rying to explain h o w science as a value-based enterprise could evolve powerful techniques for predict ion and control reflects Kuhn's desire to solve "the problem of fact and value". He contends that [w]hat we need, rather than a solution to the 'problem of induction', is the ability to think about science in such a way that its being a "value-based enterprise" occasions no surprise. A l l that hinders us from doing so is the ingrained notion that "values" are "inner" whereas "facts" are "outer". 3 9 But I do not think that it is Kuhn's purpose to lean very heavily on any inner-outer dist inction on this issue. His concern here is to explain h o w it is that the level of agreement w h i c h science characteristically a t ta ins-and w h i c h permits its success-is achieved at a l l ; it is to this end that he explores the vi ta l importance of the matrical structure of normal science. Values manifest ul t imately as facts, through their influence on theory-choice, given that particular sets of facts are generated by particular theories. Induction is one of the means through w h i c h this process happens. So exploring the problem of induct ion can offer a means by w h i c h we are enabled to see that it is not surprising that science operates according to certain values. Science and Cultural Holism In some moods, Ror ty speaks favourably of the prospect of the term "science", as w e l l as the distinctions between science and other aspects of the culture, gradually fading away. He claims that there is nothing to be gained from differentially va lu ing explanations as either "scientific" or "unscientific". There is just, rather incidentally, a greater degree of agreement among people i n , say, scientific activities than among those involved i n many other activities. My rejection of traditional notions of rationality can be summed up by saying that the only sense in which science is exemplary is that it is a model of human solidarity. We should think of the institutions and practices which make up various scientific communities as providing suggestions about the way in which the rest of culture might Rorty, in fact, does, in general, recognise that Kuhn's programme is not one which is compatible with correspondence-theorisation. See introductory point one. Rorty, PMN, p. 341. 76 organize itself. When we say that our legislatures are "unrepresentative" or "dominated by special interests", or that the art world is dominated by "fashion", we are contrasting these areas of culture with areas which seem to be in better order. The natural sciences strike us as being such areas. But on this view, we shall not explain this better order by thinking of the scientists as having a "method" which the rest of us would do well to imitate, not as benefiting from the desirable hardness of their subjects compared with the undesirable softness of other subjects. If we say that sociology or literary criticism "is not a science", we shall mean merely that the amount of agreement among sociologists or literary critics on what counts as significant work, work which needs following up, is less than among, say, microbiologists. 4 0 But it would appear that Rorty has failed to see that this stance is anything but Kuhnian. Much of Kuhn's point is that this state of being in better order is hardly accidental: it is achieved through the presence of the matrix which guides thinking and research. Pragmatic order does not appear ex nihilo, but rather is actively and deliberately sought. It is because of the convergently matrical nature of the mature natural sciences that they do enjoy the relatively heightened degree of agreement that Rorty cites as advantageous. Rorty claims "open-mindedness" to be an optimal posture for the sciences to assume, an attitude which would, for instance, permit the Royal Society "to reinvent phlogiston if that happened to be what the next scientific revolution demanded".41 But this sort of move is proscribed by Kuhn's account of the history of science as demonstrating a unidirectional and irreversible progression of successive matrices. The most that we might expect to happen in the direction of re-invention of theories or their elements, if we rely on historical precedent to guide our expectations, would be the appearance of a theory which, if circumstances demanded it (and if someone thought of it), was somewhat more similar in some way(s) to the old phlogiston-theory than to the newer oxygen-theory. Kuhn mentions a situation similar to this in the history of science: "in some important respects, though by no means in all, Einstein's general theory of relativity is closer to Aristotle's [theory] than either of them is to Newton's".42 But a wholesale re-introduction of a previous matrix is not the pattern which science has demonstrated, according to Kuhn, and most likely for good reason. We would not expect a previously discarded matrix, be it phlogiston-chemistry, Aristotelian mechanics, or whatever, to be re-introduced, if only on the grounds that it has already proven itself inadequate to the tasks demanded of it, which was precisely why it came to be replaced on the heels of a period of scientific crisis. Rorty sees Kuhn as having taken us further towards dismantling the grip which rationality had on human endeavour, by promoting an understanding of human activity which replaces a static and universal human nature with a historically situated and variable one, as well as by bringing our attention to the myriad of non-rational influences on human behaviour. Thus, Kuhn's "pragmatist Rorty, ORT, pp. 39-40. Ibid, p. 59. Kuhn, SSR, pp. 206-7. 77 friends (such as [Rorty]) routinely congratulate him on having softened the distinction between science and nonsciehce". This leads pragmatists of Rorty's stripe to try to enlist Kuhn in their campaign to drop the subjective-objective distinction altogether, substituting the idea of "unforced agreement" for that of "objectivity".43 In fact, we ought to break down "all the old philosophical oppositions between mind and world, appearance and reality, subject and object, truth and pleasure". He presumes that the resulting "rhetoric of culture" would be more Kuhnian, in the sense that it would mention particular concrete achievements-paradigms-more, and "method" less. There would be less talk about rigor and more about originality. The image of the great scientist would not be of somebody who got it right but of somebody who made it new. The new rhetoric would draw more on the vocabulary of Romantic poetry and socialist politics, and less on that of Greek metaphysics, religious morality, or Enlightenment scientism. A scientist would rely on a sense of solidarity with the rest of her profession, rather than a picture of herself as battling through the veils of illusion, guided by the light of reason. 4 4 Rationality will thus no longer be seen to be an intellectual virtue, leaving scientists with no general virtues other than reliance on persuasion (rather than force), respect for the opinions of their colleagues, and curiosity with respect to new data and ideas. On this view there is no reason to praise scientists for being more "objective" or "logical" or "methodical" or "devoted to truth" than other people. But there is plenty of reason to praise the institutions they have developed and within which they work, and to use these as models for the rest of culture. For these institutions give concreteness and detail to the idea of "unforced agreement". Reference to such institutions fleshes out the idea of "a free and open encountef'-the sort of encounter in which truth cannot fail to win. On this view, to say that truth wi l l win such an encounter is not to make a metaphysical claim about the connection between human reason and the nature of things. It is merely to say that the best way to find out what to believe is to listen to as many suggestions and arguments as you can. 4 5 But Rorty is too eager to read Kuhn as an anti-rationalist, and to enlist him in the crusade to abolish all of the traditional dichotomies which Rorty lists. I would expect that Kuhn would favour some sense of an overall human commonality, based on his acknowledgment of general genetic similarity, for instance, as well, perhaps, as an overall similar environment. Moreover, even historical situation and variability in human nature do not of themselves speak against the employment of reason, traditionally conceived. Kuhn's recognition of the impossibility of an algorithm for theory-choice, combined with his allegiance to the standard conception of rationality, is indicative of a position where Rorty, ORT, p. 38. Ibid., p. 44. Ibid., p. 39. 78 t radit ional rationality can be employed i n variable ways, depending upon the particular Archimedean platform considered scientifically appropriate at the t i m e . 4 6 The presence of non-rational influences on theory-choice do not function to chase away a l l rational influences. What K u h n outlines is a situation i n w h i c h both rational and non-rational factors affect the bir th of a new matrical candidate as w e l l as its eventual success i n achieving dominance over competing candidates. This point qualifies the similari ty w h i c h Rorty sees between himself and K u h n , that I have noted i n m y Introductory point three. None of this is intended by K u h n to "[soften] the dist inct ion between science and nonscience". 4 7 In fact, as I illustrated i n Chapter Two, K u h n has certainly made efforts to explain i n some detail h o w scientific disciplines are distinct from non-scientific ones. Some of these differences relate quite directly to the evolut ion of science towards ever greater efficiency and effectiveness i n the problem-solving w h i c h is its chief a im . Furthermore, it w o u l d be quite peculiar indeed i f K u h n were to be interested i n softening this distinction, given that his programme is specifically designed to explicate the workings of science; he does not offer his thesis as an elucidat ion of general human cul tural activities. Rorty's interpretation of K u h n as espousing "pragmatist h o l i s m " 4 8 is therefore less than accurate. Dissolving the distinctions between subject and object, m i n d and wor ld , appearance and reality, and truth and pleasure are acts w h i c h I w o u l d expect K u h n to find unhelpful, as far as scientific activity is concerned. Science, as K u h n describes it, needs to be able to dist inguish between what is to count as truth and reality, and what is deemed pleasure and appearance. After a l l , that is what a matr ix is for. Rorty's tendency to v iew such distinctions as necessarily intertwined w i t h fidelity to correspondence theories ignores the fact that one can make perfectly good use of these distinctions whi ls t adhering to a Kant ian type of framework, as K u h n does: repudiat ing correspondence theories of knowledge, and refraining from ontologising any part of these distinctions. To speak of observation, and of the subject's role i n constituting objects (recalling the Kant ian v iew o f objects conforming to our knowledge o f them, rather than the other w a y round) requires that some sense be made of a subject-object dist inction. W h i l e Rorty sees his new "rhetoric of culture" as "more Kuhnian" than the tradit ional v iew, his description of it reveals it to be decidedly un-Kuhnian. He equates "particular concrete achievements" w i t h paradigms, indicat ing that the emphasis on these signals a de-emphasis on "method". But Kuhnian "paradigms" (which K u h n later broke d o w n into the two categories of conceptual discipl inary matrices and concrete exemplars) are the very vehicle of methodology. To be sure, K u h n was concerned to stress that typical accounts of science fail to acknowledge the importance of a recognition of concrete particulars; these are See again point four in the Introduction. See p. 19 and point seven in my Introduction. Rorty, ORT, p. 65. 79 embodied within the training of future scientists, in the form of exemplars. But these exemplars encompass both the concrete results achieved by science and the methodology which was employed in the generation of those results. Moreover, concrete achievements are important in demonstrating the character and worthiness of a particular conceptual matrix. Concrete achievements are only seen as achievements through their being sanctioned by the matrix of which they are representative. After a matrix changes, what were seen to be valuable scientific achievements at one time may indeed not even be considered scientific at all, which certainly entails their being rejected as achievements. The conceptual commitments and associated methodology which are prescribed by the matrix of a particular scientific discipline in a particular period are inextricably tied to the concrete achievements of that discipline in that period; the exemplars provide a voice through which conceptual commitments and their associated methodology may speak effectively. So one cannot celebrate the former and turn away from the latter. Adherence to a matrix is, moreover, a sign of valuation of rigor in research. Originality is admittedly necessary for the emergence of new matrical candidates, something which becomes crucial at times of serious impending decay of an old matrix. But, during periods of normal science, an undue orientation towards originality and novelty would be a hindrance, rather than an aid, to scientific functioning. This is why Kuhn refers to the balance which must be kept between conservatism and originality within the scientific community as a whole. Such a balance will serve to maintain the stability of an old matrix until such time as the build-up of anomalies, and possibly various other pressures, will truly indicate the wisdom of its abandonment; thus, it ensures that matrix-change will not occur so easily as to impede the stability necessary for rigor in ordinary scientific research. On the other hand it will always ensure that new ideas are ready to move into the emerging vacuum created by a crumbling old order. So rigor~as well as reason-is important in promoting the necessary sense whereby science can see itself to be seeking truth and knowledge. Solidarity within the scientific community only exists for this purpose. Thus, it is inherently part of the scientific attitude to seek to "[get] it right", as opposed to "[making] it new". A thrust towards novelty would indeed be anathema to most scientists immersed in problems defined by the current matrix, at least without there being an undeniable state of crisis. Curiosity will play a role in both periods of normal science (propelling the refinement and extension of the extant matrix in regular problem-solving), and in periods of revolutionary science (fostering the genesis of new matrical Gestalteri). Without the notion of "battling through the veils of illusion" (although perhaps conceived with less dramatic flourish), the Kuhnian scientist's commitment to her matrix will falter. Some sense of what is to count as truth must be present. Further, solidarity cannot forge a commitment to any particular matrix. We would not be able to make any sense out of an assertion that quantum physics, for example, exemplifies more (or less) solidarity within the community of physicists than does Newtonian physics. 80 Kuhn's statements regarding persuasion as an instrument for theory-change are not meant to advocate the opposit ion to tradit ional rationality w h i c h Rorty is promoting. Reason and intel lectual thought w i l l be significant virtues i n this process. Kuhn's point is that reasoning w i l l not be the only tactic used by proponents of a new matr ix i n their promot ion of it; nor w i l l it be the only factor helping to establish the routine acceptance of a fledgling m a t r i x . 4 9 This relates to the absence of a n algorithm for theory-choice. To a certain extent, praising the "institutions" w h i c h science has developed is tantamount to praising scientists for being more "objective", " logical", "methodical" , o r "devoted to t r u t h " than other people. After a l l , a m o n g the institutions of science are just those methodologies a n d practices w h i c h direct research towards the goals of objectivity and the search for t ruth (aided by logical thought), even i f these are not to be taken i n some absolutist ontological sense. Rorty's b i d to praise scientific institutions, w h i l e turning entirely away from objectivity, logic, method, and truth is therefore confusing: it is unclear just what he envisions scientific institutions to be. Even i f these institutions "give concreteness a n d detai l to the idea o f 'unforced agreement' ", this should not be meant to say that they do not contribute a significantly robust structure to the ways i n w h i c h such agreement can be obtained. It is this structure w h i c h is exemplif ied i n the methodology w h i c h marks the matrix . Final ly , for a K u h n i a n scientist, advocating that "the best w a y to f ind out what to believe is to listen to as m a n y suggestions a n d arguments as y o u can" is to offer a hopelessly vague prescription: there must, i n the end, be some w a y of m a k i n g decisions o n scientific matters w h i c h is directed towards achieving the particular goals w h i c h characterise the scientific endeavour. In n o r m a l science, it is the matrix w h i c h offers guidance as to what beliefs are to be considered plausible; i n revolutionary science, the struggle to f ind a new scheme of guidance makes recourse to the scientific values detailed i n Chapter T w o . Matrices are crucial ly employed to place l imits o n precisely w h i c h suggestions and arguments w i l l be deemed worthy of being listened t o . 5 0 Clearly, scientific models, specific as they are to the very specialised projects o f science, are not adaptable to the activities o f the culture at l a r g e . 5 1 Quite explicitly, then, K u h n does not share Rorty's brand of cultural "hol ism". Pragmatism I t u r n n o w to the question of pragmatism, a type of doctrine to w h i c h Rorty makes frequent reference. I have not encountered i n Rorty m u c h detai l w i t h respect to his concept of pragmatism; it remains a rather fuzzy notion, w h i c h m a y See again my introductory points three and four. 5 0 See again point five in the Introduction. 5 1 See again point seven in the Introduction. 81 indeed be the way w h i c h Ror ty w o u l d prefer it to be, given the tenor of his philosophy. But what it serves to do is indicate that the motivat ion for discourse is not to reach something called "truth" i n the style of tradit ional epistemology, but to adjust means to ends i n an opt imal ly practical manner, al though against the background of a sentiment of tolerance. If we w i s h to make use of the term "truth" at a l l , we re-describe it i n terms of what it is good for us to believe, transforming it into an expression of commendat ion. A t some future time, i f a better proposal for what it is good to believe were to appear, our judgment of commendat ion w o u l d shift to the new proposal, a long w i t h our assessment of what is true. This w i l l be a collective affair, t ruth attaching to proposals upon w h i c h general cultural agreement has been bestowed. Rorty's pragmatic outlook is seated i n a continuation of discourse only unt i l its participants are "satisfied". He envisions this as a settled state of cultural agreement. However, the details of what is to count as agreement are also missing, there being no indicat ion as to whether it is to be based upon a sort of majority rule, consensus, or some other arrangement. To make even this decision of what sort of agreement is to be sought subject to the discursive approach w h i c h Rorty describes w o u l d be to enter into an unfruitful infinite regress. Kuhn's understanding of what satisfaction might entail bears a far harder edge: satisfaction, i f we w i s h to cal l it that, is achieved through the smooth operation of a research-matrix. Hence, the researchers i n Kuhn's programme are work ing w i t h a robust framework, one w h i c h w o u l d be intolerant of the k ind of infinite regress o f propositions i n discourse w h i c h Ror ty believes to be not at a l l problematic. Researchers work ing w i t h i n a matr ix must set some basic assumptions as inviolable, and as the grounding for a l l others; these constitute the most foundational elements of the matrix. One need consider the chain of justification only as far back as these fundamental elements, at least dur ing periods of normalcy. It is the presence of this network of fundamental elements wh ich sets off normal from revolutionary science; i n the latter, much energy is devoted to the re-establishment of some such framework for agreement w h i c h w i l l render future research coherent. K u h n does consider his model of science to be largely motivated by pragmatic considerations, given the importance of problem-solving, w h i c h is aided by such values as aesthetics, simplici ty, etc. W h e n Rorty speaks of pragmatism, however, his posit ion is rather more enigmatic, since he refuses to a l low for anything w h i c h w o u l d provide a solid ground for the evaluation of what is to count as pragmatic. Surely scientific judgments based upon pragmatic concern w i l l achieve sufficient complexity to require some logical orientation or other w h i c h w i l l help to sort out the more effective solutions to questions or problems from those of lesser effectiveness. Rorty's "criterionless mudd l ing through" fails adequately to acknowledge the importance of the learning process to such pragmatic endeavours. This learning process is to be found important ly i n Kuhnian scientific exemplars. Pragmatism ought to be aided by a certain level of efficiency, w h i c h is, i n turn, fostered by the learning of what works and what does not, and much of this learning involves the similarity-relations w h i c h form the 82 bulk of Kuhn's matrices. Rorty does want to acknowledge learned responses to s t imul i ; our assessment that an animal wr i th ing around before us is i n pain and may require help reflects our learned response of a judgment of "pain" upon the stimulus of our observation of this behaviour. But that responses should be learned at al l~for instance, as part of a coherent framework of understanding i n science-is more consistent w i t h a model w h i c h incorporates the presence of matrices and exemplars w h i c h guide the learning process w h i c h leads to understanding and action than w i t h the "fuzzier" articulations of Rorty's philosophy w h i c h he sometimes advances. W h i l e Rorty recognises K u h n as having encouraged a pragmatist posit ion, he is dismayed at what he sees as the tetter's having d rawn away from it i n desir ing an explanation for w h y science works . Pragmatists of the Rort ian school do not search for an explanation of the success of science. This reveals some of the exceedingly puzz l ing features of Rorty's version of pragmatism. The suggestion that seeking an explanation of w h y science works betrays a drawing-away from pragmatic concern seems to me frankly unintel l igible. Kuhn's description of science as strongly directed towards the development of problem-solving abil i ty indicates a clearly pragmatic motivat ion, a point w h i c h K u h n himself has emphasised. Thus, examining the workings of science is a paradigm example of what it is to be practically motivated. It may be instructive i n this connection to note that, i n some moods, at least, Rorty appears to be more hostile to pragmatism than promoting of it, as w h e n he asks what is so special about prediction and control? Why should we think that explanations offered for this purpose are the "best" explanations? 2 A n d , Nietzschean history of culture, and Davidsonian philosophy of language, see language as we now see evolution, as new forms of life constantly killing off old forms-not to accomplish a higher purpose, but bl indly. 5 3 This attitude betrays a stance w h i c h has moved very far from the concerns w h i c h K u h n addresses. Certainly, predict ion and control figure prominent ly i n the motivat ion and operation of science, by Kuhn's account, or any standard account, for that matter. Moreover, the removal of predict ion and control from pragmatic concern makes it difficult to see what might be left. Also , discovering a Gal i lean vocabulary to be practically superior to an Aristotel ian one w o u l d hardly be an example, for Kuhn , of a blind replacement. Ror ty describes the pragmatic v iew as one i n w h i c h theory follows after, rather than being pre-supposed by, concrete achievement . 5 4 As I have noted Rorty ORT, p. 58. Rorty, CIS, p. 19. 83 earlier, the particulars which are concrete achievements, in a Kuhnian system, can only exist by virtue of the general framework which contributes to the determination of those particulars in the first place. All observations are theory-dependent, which is at the root of why we are unable to isolate a neutral-observation language. It does not help to contend that concrete achievements are simply self-creating acts, for these, too, must be counted as observations in relation to any theory which incorporates them. If theory is to follow after the identification of particulars, then we have moved back in the direction of the correspondence theories which both Rorty and Kuhn have rejected. Rortian pragmatists recommend that we worry only about the choice between two hypotheses, rather than about whether or not there is something which makes either true. This is supposed to rid us of questions of objectivity of value, rationality of science, and viability of language-games. We replace these with pragmatic questions about whether or not we ought to keep our present values, theories, and practices, or replace them with others. Pragmatists interpret the goal of enquiry as an appropriate mixture of unforced agreement and tolerant disagreement; what is appropriate is determined by trial-and-error. For Kuhn, the question of truth is integral to the acceptance, by the relevant scientific community, of any particular matrix. The Kantian orientation of Kuhn's programme removes the notion of truth from the sphere of correspondence-theories. But that scientists envision their work as a search for truth is still an essential motivating factor, as I have mentioned in Chapter Two. Rorty seems to be missing the subtlety that, even if we reject the scientific realists' search for non-perspectival truth-attained upon our matching our theories with the way in which the world is, in some ontological sense-we need not reject entirely all notions of truth. Neither would Kuhn be prepared to give up the notions of the objectivity of value, the rationality of science, or the viability of language-games. Devoid of a grounding in correspondence-theorising, objectivity of value comes to look more like intersubjectivity of value: agreed-upon standards of thought and methodology, which are expected to be informed also, however, by the basic presence of the common "given". I have already discussed in this chapter the importance of reasoning in Kuhn's account of science. If "viability of language-games" is to mean how well various vocabularies function in facilitating communication, then this surely must be an aspect with which Kuhn is deeply concerned, given all that he has written on the question of the incommensurability of the vocabularies of different matrices, as well as the vital role played by the relatively static vocabularies of normal science. Even just in general terms, we would expect the fulfilment of pragmatic goals to be aided by effective communication. Rorty's neglect of Kantian-type epistemological frameworks is also evident in the following passage: Rorty, ORT, p. 79. 84 When the notion of "description of the world" is moved from the level of criterion-governed sentences within language-games to language-games as wholes, games which we do not choose between by reference to criteria, the idea that the world decides which descriptions are true can no longer be given a clear sense. 5 5 But "the w o r l d " does not have to be the source of criteria for there to be criteria (or better, similarity-relations) at a l l . In Kuhn's description of science, patterns of similarity-relations originate w i t h the scientific community; they are not read off of Nature, as a scientific realist or empiricist w o u l d believe. Also, one can move the analysis to whole language-games (or whole matrices) from single sentences (or isolated observations), as Rorty observes that K u h n has done, wi thout jett isoning the not ion that c r i t e r i a - in the form of scientific values, for e x a m p l e -have something to do w i t h theory-choice. Rorty seems to be unfortunately impeded by the not ion that "[t]he temptation to look for criteria is a species of the more general temptation to th ink of the wor ld , or the human self, as possessing an intrinsic nature, an essence". 5 6 As we have seen w i t h Kuhn's programme, this is not the only spirit i n w h i c h we can conduct the search for paradigms for belief and action. Also, i n Kuhn's framework, the input of Nature as the "given" means that Nature does play some role i n the viabi l i ty of any theory, even i f it does not play a decisive, or the sole, role. I w i s h to consider one last passage i n Rorty on the subject of pragmatism. This concerns the question of progress. Rorty says that To say that we think we're heading in the right direction is just to say, with Kuhn, that we can, by hindsight, tell the story of the past as a story of progress. To say that we still have a long way to go, that our present views should not be cast in bronze, is too platitudinous to require support by positing limit-concepts. 5 7 But this is crucial ly under-specified, i n respect of to what we might be heading i n the right direction, or to what we might still have a long way to go. One wonders at what we are supposed to be looking, i n this hindsight. Rorty has resisted spell ing out anything w h i c h might furnish us w i t h a sense of how one determines a direction for the cu l tu re -or for a specific cultural activity such as science. There are no goals; there is nothing for w h i c h to a im save the rather empty prescription of maximis ing unforced agreement and fostering a certain amount of tolerance for disagreement. But even this goal is not clearly defined. To make any sense, the c la im that we are heading i n the right direction, i.e., making progress, must be matched against the possibility, i n principle, of our not being able to tel l the story of the past as a story of progress: we ought to be able to identify what it w o u l d be for us not to progress. For Kuhn , progress is to be read, i n the main , off of scientific puzzle-solving ability, w h i c h increases w i t h successive matrices, not by Rorty, CIS, p. 5. Ibid., p. 6. Rorty, ORT, p. 27. 85 chance, but by a clear and conscious direct ion of purpose: "Newton's mechanics improves on Aristotle's and Einstein's improves on Newton's as instruments for puzzle-solving" . 5 8 Judgment of progress involves our looking to our means to evaluate the effectiveness w i t h w h i c h they have achieved our ends, a process, w h i c h w i l l be facilitated, i n part, by rational thought, t radi t ionally conceived. So, K u h n w o u l d not concur w i th Rorty that the not ion that Newton's vocabulary "[got]...at the truth about the heavens...is not an explanation of anything", but "just an empty compl iment -one tradit ionally paid to writers whose novel jargon we have found useful" . 5 9 Such emptiness is antithetical to pragmatism. The discussion of this section should indicate that Rorty's belief, as expressed i n point six of m y introductory list, wou ld , to be accurate, require such extensive qualification that we must regard it as fundamentally f lawed; Rorty's and Kuhn's respective orientations towards pragmatism indeed provide little, i f any, basis for compatabil i ty between them. Truth and Relativism Both Ror ty and K u h n have strived to communicate to their readers the importance of keeping an eye to historical conditions and influences i n our attempts to understand human endeavour. This has been an important factor o n the basis of w h i c h Rorty sees his commonal i ty w i t h K u h n . 6 0 It is as a result of their taking historical elements as seriously as they do that both have had to deal w i t h charges of relativism from their critics. Such cri t icism also focuses on their a l lowing a communi ty to set standards for what is to count as true on the basis of considerations w h i c h are pragmatic i n nature, rather than metaphysical (and hence more l ike ly to be universal) . But each denies that he is a re la t ivis t . 6 1 In Rorty's defence of his philosophy as non-relativistic, he appeals to a sort of in i t ia l , or provisional, "ethnocentrism". A communi ty starts w i t h the beliefs w h i c h it regards as acceptable: "[ejither we attach a special privilege to our o w n community, or we pretend an impossible tolerance for every other group" . 6 2 But the communi ty must, at some point, subject those beliefs to scrutiny through discourse w i t h other communities. As a result, the community 's web of beliefs may end up being re-woven: "[b]eliefs suggested by another culture must be tested by t rying to weave them together w i t h beliefs we already have" . 6 3 Rorty argues that it is this ethnocentrism-the community 's in i t ia l adherence to its o w n beliefs, coupled w i t h its in i t ia l rejection of the beliefs of Kuhn SSR, p. 206. Rorty, CIS, p. 8. See again point eight in the Introduction. See introductory point nine. Rorty, ORT, p. 29. Ibid., p. 38. 86 other commiinities--which serves as evidence against his position being relativist. The only way I can make sense of this is to interpret Rorty as taking relativism to be a position wherein anything goes, so to speak, i.e., where any belief is as good as any other. But that is not the way in which relativism is typically defined. A relativistic position would be one which describes beliefs of what is true or false, or right or wrong, as relative to a particular community. But this is just the sort of picture with which Rorty presents us. Each culture would espouse its own ethnocentric point of view --shifts in its views notwithstanding~and there is no place for a God's-eye view, as it were, from which the belief-webs of different communities might be evaluated. Therefore, from the point of view of any given community, its own beliefs are true. Although ethnocentrism motivates us to employ our own present beliefs in our decisions of how to apply the term "true", Rorty advises us that we must take care to refrain from defining "true" in terms of those beliefs64, for this is a definition which would only apply to our own culture, and it would be a mistake to universalise beliefs which are only typical of a particular group. I take Rorty to be offering an account of truth which relies on definitions which are qualified (in that any definition will be group-specific) as well as provisional (in that any definition will be subject to change over time, with exposure to new ideas). There is a certain tension, clearly, between the concepts of ethnocentrism and of qualified, provisional truth. There will have to be some sense as to how strongly a culture ought to maintain its ethnocentricity before it comes seriously to entertain the potential truth of the beliefs of other cultures, upon which a re-weaving of the web of beliefs must be considered. (Incidentally, this is a parallel question to that of Kuhn's balance between maintaining allegiance to a disciplinary matrix under stress, and seriously considering up-and-coming matrical candidates.) Rorty states that, even if one drops the idea that a common ground for appeal in the question of truth is provided by evidence, this does not mean that one culture's web of beliefs is as good as another's.65 This seems a rather hollow statement. From within a given culture, by ethnocentrism, its own web of beliefs will be seen as superior to all others. But outside of the cocoon of mere intra-societal agreement, the lack of any cross-cultural evidence to which appeal may be made removes all means of weighing beliefs from different webs against each other. Seen from this broader vantage-point, then, one web of beliefs must be considered to be as good as any other. Again, it is difficult to see how this is anything but a relativist position, since, if values are not to be universal, then they must be relative to something, i.e., the particular culture which holds them. Kuhn's approval of the authority of the accepted and time-proven matrix to sanction certain beliefs as true might be seen as at least superficially similar to Rorty's relativism. That which is to be understood as true is relative to whatever the scientific community sets out as admissible thought and practice. But, as I Ibid., p. 50. Ibid., p. 67. 87 have noted i n Chapter T w o and above, K u h n adheres to a thesis whereby some sort of scientific progress is achieved, evidenced by the increasing refinement, sophistication, and explanatory power of theories, as w e l l as by the unidirect ional and irreversible evolut ion of one matrix into the next. "One scientific theory is not as good as another for do ing what scientists normal ly do." Later theories are progressively better as tools for the practice of normal science; earlier theories are later abandoned as false . 6 7 Progress is not a feature of a relativistic programme; nor is it an intelligible feature of Rorty's. The question then is whether or not K u h n is correct i n seeing science as progressive i n this way . At any rate, as I have argued i n the previous chapter, Kuhn's description of the operation of different matrices is considerably more structured and complex than is Rorty's portrayal of a plural i ty of discourses. Rorty complains that critics of pragmatism charge that pragmatists "have defined 'true' as 'satisfies the standards of our community ' ". He objects that "we pragmatists do not ho ld this relativist v i e w " . 6 8 The most that I can make out of this is that Rorty believes his posit ion and Kuhn's to escape the shadow of relat ivism through what he estimates to be their mutual avoidance of standards, or criteria, against wh ich a relativist conception of truth w o u l d be defined. Rorty attempts to further this by asserting that "true" should be neither analysed nor defined, anyway. So i f we have neither standards against w h i c h to define truth nor a concept of t ruth to begin wi th , then we cannot be accused of occupying the relativist posit ion w h i c h claims that t ruth is relative to whatever standards a society ho lds . 6 9 Despite the fact that it appears that Rorty has misread K u h n i n suggesting that the latter does not indeed v iew truth as being measured against whatever the communi ty has decided upon as a set of standards (viz., a matr ix) , there is st i l l the obvious problem that Ror ty s imply cannot avoid the relativism of his o w n posit ion i n this way. One might viably re-cast the definit ion of relativism to suit Rorty's programme rather admirably, as defining whatever it is good to believe as that discourse to w h i c h a given culture has habituated itself. In this connection, Rorty's protest that pragmatists "do not infer from 'there is no way to step outside communities to a neutral standpoint' that 'there is no rational way to justify l iberal communit ies over totalitarian communities ' " 7 0 sheds no light on how pragmatists might achieve such justification, given Rorty's disdain for any standards at a l l , let alone neutral ones. Rorty seems unprepared to admit that even his programme requires standards of some sort, i n order to get off the ground. He insists that "[w]hat [pragmatists] i n fact infer is that there is no w a y to beat totalitarians i n argument by appealing to shared common premises". W e are to "forget about being responsible to what is 'out there' ", recognising instead "that human communities can only justify their existence by comparisons w i t h See especially p. 51. Kuhn, CGK, p. 264. Rorty, ORT, p. 42. Ibid, pp. 25, 50. Ibid, p. 42. 88 other actual and possible human c o m m u n i t i e s " / 1 But comparisons can only be made by means of specified parameters. Moreover, it is difficult to envision on what persuasive power one could possibly draw, i n discussions w i t h totalitarians, i f one is not to appeal to common human sentiments of some sort, such as empathy or justice. It is on the grounding of such sentiments that we enter into comparisons between other communit ies and our o w n . This need for standards of some sort i n order to render Rorty's philosophy at least min ima l ly intelligible again thrusts h i m back into Kuhn ian territory, even though he recognises neither K u h n nor himself as resorting to such standards. Again , this underscores the problematic nature of Rorty's belief as expressed i n my introductory point five. Rorty may be leaning on the notions of solidarity and tolerance, as means by w h i c h to justify l iberal modes over totalitarian ones. In that case, Rorty's programme w o u l d indeed not be fully relativistic. But then he must clearly state this as a common, universal value, and his allowance for ethnocentrism w o u l d be correspondingly significantly d iminished. If the idea of ethnocentrism is just that a culture begins from the beliefs w h i c h it already holds, and strives to achieve the ideal of greater tolerance from there, we appear st i l l to have a universalist type of thesis, and not one w h i c h takes ethnocentrism very seriously. W h i l e both Rorty and K u h n take truth and reference to be relative to a conceptual scheme (although one may question the val id i ty of Rorty's use of the concepts of reference here, given what he has said elsewhere), there is a basic difference i n the way each of them conceives of a conceptual scheme. For Rorty, it is s imply "what we believe now~the collect ion of views w h i c h make up our present-day cul ture" . 7 2 For K u h n , it is what makes those beliefs possible i n the first place, namely the a priori matr ical structure. W e might consider this latter to compose a special category, or order, of beliefs. Alternative Discourses/Matrices The same lack of clari ty w h i c h bedevils Rorty's v i ew of single discourses plagues his account of a culture's re-weaving of its web of beliefs. He contends that, i n hol ism, [o]ne wi l l do, in short, just what the "new fuzzies" in philosophy of science say scientists do when some relatively large-scale proposal to change the way nature (or part of nature) is pictured is up for discussion. One wi l l muddle through, hoping that some reweaving wi l l happen on both sides, and that some consensus may thus emerge. 7 3 Ibid. ; Rorty, PMN, p. 276. Rorty, ORT, p. 67. 89 It should be clear here that, al though Rorty apparently views K u h n as one of these "new fuzzies", the Kuhn ian account of matrix-transformation does not match this Ror t ian account of discourse-re-weaving. A n y sense of "muddl ing through" w i l l be displaced, i n any endeavour recognisable as science, by the thrust to retrieve some sort of matr ical structure. Also , adherents to an o ld matr ix generally resist re- • weaving, some for longer than others, i n the attempt to stretch that matr ix to accommodate anomalies or answer new demands. A t any rate, i n Kuhn ian terms, as I shall explain i n the next section, the question of some mutual re-weaving of matrices, i n the interest of coming together o n some median territory is moot, on account of the incommensurabil i ty of alternative matrices, or scientific discourses. Rorty re-iterates his rejection of epistemology, this t ime, i n the context of inter-discourse dialogue: The advice to see if it might not pay to reweave your web of belief in the interests of a better ability to solve your problems is not the advice to formulate epistemic principles. The one piece of advice would only entail the other if experience had shown that having a conscious epistemological view were always an efficient instrument for readjusting old beliefs to new. But for K u h n , this is precisely what experience has shown. Science progresses through the refining and extension of matrices, fol lowed by revolutionary shifts to new ones. One of the major reasons for the adoption of a new matr ix is scientists' commitment to a belief i n its epistemic benefit. In the same vein, Rorty asserts that the notions of criteria and choice (including that of 'arbitrary' choice) are no longer in point when it comes to changes from one language game to another. Europe did not decide to accept the idiom of Romantic poetry, or of socialist politics, or of Galilean mechanics. That sort of shift was no more an act of w i l l than it was a result of argument. Rather, Europe gradually lost the habit of using certain words and gradually acquired the habit of using others. 7 4 There is a certain degree of s imilar i ty to K u h n here. As I have described i n Chapter Two, K u h n cites the inaccuracy commit ted i n describing theory-choice i n terms of conscious deliberation and decision, or choice. He finds the term "conversion" more accurate to peg the way i n w h i c h the shift to a new theory occurs, what we might loosely term a Gestak-shift. But he has also stressed that this does not point to the intrinsic irrationali ty of the process, arguing that one can sti l l pick out good reasons w h i c h take part i n the conversion from one theory to its r iva l . So argumentation is not al ien to the process. 7 5 Moreover, as I have mentioned several times, the Kuhn ian account of science, i n respect of both the matrices of normal science and the matrical shift w h i c h occurs i n times of Rorty, CIS, p. 6. See pp. 38; this also relates back to the inadequacy of Rorty's belief as expressed in my introductory point four. 90 revolution, does not completely eschew criteria (the d iminished role he allows for them notwithstanding), and certainly does not discount the importance of a framework of similarity-relations. There is also, as I have previously noted, more at stake than a mere shift i n linguistic habit. If habituation to a discourse were as important a feature of its usage as it seems to be for Rorty, at least i n some moods, then we must surely a l low for a d iminished focus on the pragmatic appeal of that discourse. A gradual change i n habit of the sort w h i c h it w o u l d appear is being advocated i n the above quote-nei ther w i l l e d nor argued f o r - w o u l d speak against there being anything l ike an assessment o f the pragmatic advantage o f a particular discourse. Rorty also states that revolutionary achievements i n the arts, sciences, and mora l and pol i t ical thought typical ly occur when somebody realises that two or more of our vocabularies are interfering w i t h one another; a new vocabulary is invented to replace b o t h . 7 6 This adds another dimension, i n the form of a clash w i t h a second vocabulary, beyond the consideration of the pragmatic effectiveness of a single discourse. It is un-Kuhnian-as Ror ty w o u l d himself agree - in its omission of the expectation of some assessment of the goodness-of-fit of a theory w i t h Nature, on account of the input of Nature into observations, v i a the "given". It also departs from K u h n i n that, for the latter, there can be no question of revolut ion being fomented by a clash between two discourses (matrices), for only one discourse (matrix) can prevail i n the pre-revolutionary period to guide research at any one t ime. At least that is what is true for science, w h i c h raises the further point that the account of science, for K u h n , must be distinguished from that of the arts, and of mora l and poli t ical thought . 7 7 In science, crisis is typical ly instigated w h e n the extant theory's goodness-of-fit to Nature begins to weaken. The Question of Incommensurability Rorty defines the term "commensurable" as "able to be brought under a set of rules w h i c h w i l l tel l us how rat ional agreement can be reached on what w o u l d settle the issue on every point where statements seem to conflict". W e are able, as a consequence of these rules, "to construct an ideal situation" for the reaching of agreement. A l l disagreements w i l l then be seen as " 'noncognitive', or merely verbal, or else merely temporary". This structure "allows the interlocutors to agree to differ-being satisfied of each other's rationality the whi le" . A l l of this illustrates "[t]he dominat ing not ion of epistemology...that to be rational, to be fully human, to do what we ought, we need to be able to find agreement w i t h other human beings." He asserts that "[t]o construct an epistemology is to find the m a x i m u m amount of common ground w i t h others", w i t h "[t]he assumption that an epistemology can be constructed" being "the assumption that such common Rorty, CIS, p. 12. See again point seven in the Introduction. 91 ground exists". C o m m o n ground has been sought outside of us i n such places as the Platonic Forms, or the realm of Being; w i t h i n us, i n the m i n d ; and i n language, where the search has been for a universal scheme for a l l possible content. But we are warned to '[n]ote that this sense of 'commensurable' is not the same as 'assigning the same meaning to terms' ". This latter sense, "which is the one often used i n discussing Kuhn-does not seem to [Rorty] a useful one, given the fragility of the not ion of'sameness of meaning ' ". So he asserts that "[t]o say that parties to a controversy 'use terms i n different ways' seems to [him] an unenlightening way of describing the fact that they cannot find a way of agreeing on what w o u l d settle the issue". 8 It is essential to notice here at the outset that Rorty's definit ion of the term "commensurable" covers what he has elsewhere described as tradit ional epistemology, bolstered by equally tradit ional rational thought. This definit ion of the term is profoundly peculiar, since, as we have seen w i t h other of Rorty's terms, it is quite non-standard. It is unclear w h y he deemed such a re-definition necessary. The sense of "commensurable" w h i c h Rorty wants us to discount is not merely "the one often used i n discussing Kuhn" , but is significantly close to the one w h i c h K u h n himself employs. Incommensurable elements are not inter-translatable~a reflection of their bearing different meanings-because they belong to separate vocabularies w h i c h carve up the wor ld , so to speak, i n different ways. This understanding of commensurabi l i ty encompasses, but goes further than, the linguistic one (assigning the same meaning to terms) w h i c h Rorty has rejected i n favour of his o w n re-definition. K u h n basically describes these vocabularies as exhibi t ing a fundamental difference i n taxonomic categories (ways of "carving up the wor ld") , rendering the situation deeper than a mere linguistic account of it w o u l d a l low; a vocabulary is ak in to a Kuhn ian type of conceptual scheme, and must be i n place before a description of the w o r l d can begin. Rorty states that K u h n has questioned the need for commensura t ion . 8 0 But what this tells us, i n Ror t ian terms, is that K u h n has questioned the need for rules directing the attainment of rational agreement, the need for an epistemology, and the need for maximis ing common ground w i t h others. As we have seen, it is not at a l l part of Kuhn's programme to renounce epistemology, rational thought, or rules (or more accurately, similarity-relations) for direct ing thought and action. The maximisat ion of common ground among people is the only element here w h i c h seems at a l l correctly included, since K u h n is on ly interested i n the attainment o f agreement among members of relevant scientific communities, and not among people i n a more general context. Further, agreement between the proponents of Aristotel ian mechanics and those of Gal i lean mechanics is not something w h i c h K u h n w o u l d seek to maximise, since he is pr imar i ly interested i n the ways i n w h i c h they diverge. Kuhn's exposition o f science is largely directed towards elucidating just how little common ground there can be between proponents of Rorty, PMN, p. 316 and footnote, 317. See pp. 34-6. Rorty, PMN, p. 317. 92 alternative matrices. In fact, it is the tradit ional cumulative picture of science w h i c h has over-estimated the amount of common ground among past scientific theories. In Rorty's view, epistemology has tradit ionally v iewed knowledge as having a logos, w h i c h can be given by a method of commensuration; this is needed for "genuine cognition". O n the other hand, the pragmatic approach to knowledge w h i c h we see exemplified i n the epistemological behaviourism w h i c h Rorty advocates divides commensurable discourses from those w h i c h are not. This is, for h i m , equivalent to d iv id ing normal from abnormal discourse, w h i c h "generalizes Kuhn's dis t inct ion between normal and revolutionary science". It parallels also the tradit ional dist inction between the search for "objective knowledge" and other, less privileged areas of human activity. N o r m a l science solves problems against the background of a consensus about what counts as a good explanation of phenomena and what constitutes a problem-solution, whi le revolutionary science yields the introduct ion of a new paradigm of explanation and a new set of problems. 8 1 Rorty goes on to say that normal science is as close as real life gets to the epistemologist's notion of rationality. Abnormal discourse is born when someone w h o is ignorant of these conventions, or w h o sets them aside, joins i n the discourse. The product of abnormal discourse can be anything from nonsense to intellectual revolution. There is no discipline w h i c h describes it: for Rorty, it is unpredictable and creat ive. 8 2 But, the practice of revolutionary science does not admit participants w h o are ignorant of scientific conventions; nor does it admit those w h o are w i l l i n g s imply to set them aside. Proponents of new matrices have already been trained i n the dominant matrix; they w o u l d not be admitted to the process otherwise. Scientists do not invite non-scientists to propose new templates for their work . Accepted conventions are not s imply "set aside". Some of them w i l l remain intact, i f they have not been th rown into question by the dissolut ion of the o ld matrix. But even i f they do not survive the change, the o ld conventions are not "set aside" so easily: they begin seriously to crumble, and are reluctantly abandoned, under circumstances of great tu rmoi l . The discipline w h i c h existed i n the t ime of normalcy (e.g., inorganic chemistry) w i l l s t i l l be recognisable as a discipline after the shift, even i f a new related discipline has been cleaved off of it as a result. There is no question of there being no discipline at a l l dur ing abnormal periods. Al though a certain amount of unpredictabil i ty and creativity can be expected i n the bir th of a new research-matrix, the not ion that the product of a period of abnormali ty could possibly be "nonsense" places what Rorty is describing at a considerable remove from what K u h n is interested i n elucidating. If a scientific revolution failed to produce a new order, then it w o u l d devolve into either the sort of pattern characteristic of pre-paradigm science ( in w h i c h case, Rorty, PMN, pp. 11, 319-20. Ibid., p. 320. 93 there w o u l d be a plural i ty of parallel orders) or a situation where there was no order at a l l . The first case w o u l d signal a transformation from a mature state to one of immaturi ty, w h i c h I w o u l d expect K u h n to v i ew as an extremely unl ikely outcome, and one unprecedented i n the history of science. The second case w o u l d signal the end of science. Neither of these is a productive opt ion w h e n one is concerned to describe the special process w h i c h is science (as K u h n is), both historically and i n terms of what must be the parameters of the endeavour i f it is to continue to be called "science". For Rorty, we w i l l be epistemological i n circumstances where we k n o w what is going on, and want to codify it, i n order to extend, strengthen, ground, or teach it. Thus, we do get epistemological commensuration, but only where we already have agreed-upon practices of enquiry (or more generally, discourse). W h e n a practice has continued for long enough, its conventions are relatively easy to isolate, w h i c h is what makes epistemological commensurat ion possible . 8 3 Rorty compares this stance towards epistemology to "Nelson Goodman's pragmatist attitude toward logic", where we discover our rules of inductive and deductive inference by discovering what inferences we habitually accept 8 4 : " A rule is amended i f it yields an inference we are unwi l l ing to accept; an inference is rejected i f it violates a rule we are unwi l l i ng to amend" . 8 5 This particular version of the pragmatic approach to knowledge reflected i n epistemological behaviourism diverges radical ly from the prescription for the treatment of standard epistemology w h i c h Rorty has given elsewhere. Hence, it appears to be a contradiction internal to Rorty's philosophy. Generally, he offers no opt ion where in we might "be epistemological", and hence be engaged i n "divid[ing] commensurable discourses from those w h i c h are not". The usual admonishment is to dispense w i t h standard epistemology~and therefore w i t h commensurable discourses-entirely. Rorty's al lowance for epistemological commensurat ion where a long-standing practice has made for well-engrained conventions is at odds w i t h his favourable (albeit inaccurate) assessment of K u h n as having questioned the need for commensurat ion at a l l , whether "commensuration" be defined on Rorty's terms, or standardly. Addi t ional ly , the "rules of inductive and deductive inference" w h i c h Ror ty is here a l lowing, even i f they are only isolable after the fact of the laying-down of the order, had no place at a l l i n the Rort ian cultural programme w h i c h I have described so far. The codification of a discourse/matrix, " i n order to extend, strengthen, ground, or teach it" sounds suspiciously s imilar to the matrix-oriented picture of science w h i c h we inherit from K u h n . Again , Rorty seems to be ambiguous i n what he is advocating, and appears to be, at times, s l iding into a Kuhn ian picture w h i c h he w o u l d not recognise as Kuhnian . The rather skewed picture of Kuhn's programme w h i c h we receive from Rorty is further evidenced i n the fol lowing passage: 8 3 Ibid., p. 321. 8 4 Ibid. 85 Nelson Goodman, Fact, Fiction and Forecast, quoted in Rorty, PMN, p. 321. 94 Since the Enlightenment, and in particular since Kant, the physical sciences had been viewed as a paradigm of knowledge, to which the rest of culture had to measure up. Kuhn's lessons from the history of science suggested that controversy within the physical sciences was rather more like ordinary conversation (on the blameworthiness of an action, the qualifications of an officeseeker, the value of a poem, the desirability of legislation) than the Enlightenment had suggested. In particular, Kuhn questioned whether philosophy of science could construct an algorithm for choice among scientific theories. Doubt on this point made his readers doubly doubtful on the question of whether epistemology could, starting from science, work its way outward to the rest of culture by discovering the common ground of as much of human discourse as could be thought of as "cognitive" or "rational". 8 6 As we have seen, particularly i n Chapter Two, Kuhn's lessons from the history of science clearly do not suggest a greater s imilari ty between the physical sciences and ordinary conversation than the Enlightenment had suggested. Kuhn , i n fact, does not offer cri t icism of Enlightenment th inking in general. His cr i t ic ism is pr imar i ly directed at the tradit ional , ahistorical picture of science as more or less smoothly cumulative. That he has denied the possibility of the construction of an algori thm for theory-choice is not tantamount to his denying a role for reason i n matters scientific, and does not lead inevitably to a merging of the scientific endeavour w i t h other cultural activities or ways of th inking. It seems safe to say that K u h n w o u l d share w i t h "his readers" a rejection of such an algori thm, but I w o u l d want to be more careful about suggesting that he w o u l d share w i t h them a rejection of the not ion that epistemology could work its way out from science to the rest of culture through the discovery of a common human rational grounding. Caut ion is required here not because K u h n does see epistemology's role i n this light, but because his remarks on the genetic commonali t ies among human beings imply loyalty to the not ion of at least some degree of fundamental common cognitive ground, upon w h i c h socio-psychological influences act. This particular point is at least part of the Enlightenment attitude w h i c h Rorty rejects. Also , whi le it is true that K u h n is not advocating an epistemology w h i c h operates through science, to influence the rest o f culture, neither is he advocating a meld ing of the projects of science w i t h those of the rest of culture, as Ror ty often does. K u h n is clear on the divergence of structure found between the sciences and the non-sciences; his concern is to explore science as the distinct discipl ine w h i c h it has become. 8 7 Although Rorty is i n agreement w i t h K u h n on the lack of a neutral observation-language, he complains that K u h n has, on the basis of this issue, ventured too far into the territory of ideal ism. This relates to Kuhn's v i ew that the lack of such a language results from the situation where in the proponents of different theories "see different things" or "live i n different worlds". He refers to remarks of Kuhn's on this score as "incidental". Rorty assesses that what K u h n Rorty, PMN, p. 322. See again introductory point seven. 95 wished to oppose was the tradit ional c la im that "what changes w i t h a paradigm is only the scientist's interpretation of observations that themselves are fixed once and for a l l by the nature of the environment and of the perceptual apparatus". Rorty contends that "this c la im is innocuous i f it means merely that the results of looking can always be phrased i n terms acceptable to both sides ('the f luid looked darker', 'the needle veered to the right', or, i n a pinch, 'red here now!')"; it w o u l d have been enough for K u h n to show that phrasing things i n terms acceptable to both sides is of no help i n r inding an algori thm for theory-choice. 8 8 The alleged foray into ideal ism probably relates to the point that different theory-dependent experiences signal divergent representations. For K u h n , such different representations are the consequence of a difference i n the a priori template through w h i c h a l l experiences are had. This is the basis of his remarks concerning the way i n w h i c h in some sense, the proponents of different theories live i n different worlds . As we have seen, this is not meant to be an ontological proposit ion. Since observations (or more broadly, experiences) are theory-dependent, different theories generate different experiences, at least i n those areas where the theories are crucial ly divergent. W e might expect Rorty to be clearer on this point, given his report that K u h n espouses "post-positivistic philosophy of science", followed by the explanation that the positivists held language and wor ld , and theory and evidence, apart . 8 9 Moreover, he states that "there is no question of taking an object out of its o ld context and examining it, a l l by itself, to see what new context might suit i t . . . . [A] belief is what it is only by virtue of its posit ion i n a web". W e must drop the tradit ional opposit ion between context and thing contextual ised. 9 0 It is curious that Rorty refers to Kuhn's remarks as "incidental", since an understanding of this issue is at the core of Kuhn's programme: the theory-dependence of observations, and the resulting impossibil i ty of a neutral observation-language, are the underpinning of his argument for the existence of scientific revolutions. A revolutionary history of science, as opposed to a relatively smooth cumulative one, is such precisely because of this fundamental divergence i n experience (i.e., i n the observations w h i c h are made, and i n what thoughts it is scientifically intelligible to have). The c la im that it is only the scientist's interpretation of fixed observations w h i c h changes w i t h a change i n matr ix is not, for K u h n , "innocuous". This c la im s imply cannot be salvaged, because it does not go deep enough. Interpretation of results is something w h i c h goes on within a given matrix, because it is something w h i c h is practiced on a given set of results. On ly w i t h i n a matr ix can there be any such thing as a given set of results. Different matrices generate different sets of results, and so the way i n w h i c h different propositions are made using different theories is not a result of interpretational differences. Again , this strikes at the heart of what K u h n is t rying to accomplish i n introducing the not ion of a history of science sensitive to the presence of revolutionary shifts. Rorty's attempt to Rorty, PMN, p. 324; and Kuhn, SSR, p. 120, as quoted in Rorty, PMN, p. 324. Rorty, ORT, pp. 64-5. Ibid, p. 98. 96 promote solidarity by suggesting that one can always phrase "the results of looking. . . in terms acceptable to both sides" is not possible on the Kuhn ian view, and this, as we shall see shortly, is the basis of Kuhn's description of alternative matrices as incommensurable. In light of a l l of this, it w o u l d not only not have been enough for K u h n t o show that phrasing things i n terms acceptable to both sides is of no help i n locating an algori thm for theory-choice, it wou ld , i n fact, be impossible for h i m to show this, since, i f things could always be phrased i n terms acceptable to both sides, this w o u l d indicate a persistent common template of experience, existing at a level overarching alternative matrices. K u h n denies the presence of such a common template (which is not to say, however, that he denies all commonali t ies between proponents of alternative matrices). Incidentally, Rorty's examples of candidates for such mutual ly acceptable terms-perhaps most obviously "red here now!"~bear an ironic s imilar i ty to the phenomenal entities w h i c h Rorty has been so diligent elsewhere i n attempting to purge from our understanding of things. It is an important point i n K u h n that such phenomenal sorts of reports are pre-theoretical; a scientist w i l l make an observation of a needle veering to the right, for instance, i n a way w h i c h goes beyond this pre-theoretical level : laden w i t h a lot of background information concerning the apparatus and its function i n his matrically-oriented research. So i n science, observations w i l l not occur i n this sort of broadly acceptable form, but w i l l diverge according to matrix-specificity. Rorty complains that it is unfortunate that K u h n used "romantic" notions such as that of scientists being presented w i t h a new wor ld , since this has contributed to the fears of his critics that he is promoting a created, rather than a discovered, w o r l d . For Rorty, "nothing deep turns on the choice between" discovery and creation of the wor ld -"be tween the imagery of making and of finding". He sees K u h n as having invited trouble unnecessarily by using the not ion of a new w o r l d : he could have avoided this by st icking "to the classic not ion of 'better describing what was already there' ". This is not meant to slide back into a metaphysical v iew: we stick to the classic not ion because we speak of history from our o w n Whiggish standpoint, and we need to keep something constant throughout the story. "The forces of nature and the small bits of matter, as conceived by current physical theory, are good choices for this ro l e" . 9 1 I have already explained why, for Kuhn , the notions of discovery and creation need be dis t inguished. 9 2 They are radical ly different notions, and it is difficult to see how it could be that nothing deep turns upon the choice between them. In c i t ing the matrix-shift embodied i n Kuhn's example of Aristotle's seeing the constrained fall of a stone where Gali leo saw a pendulum, Rorty states that "we need make no more of the gestalt-switch i n question than the fact that people became able to respond to sensory stimulations by remarks about pendulums, Rorty, PMN, p. 344. See pp. 60-2; this also relates to my introductory point two. 97 without having to make an intervening inference". Similar ly , he notes that we have to be clear that the shift, for example, from Ptolemaic to Copernican astronomy was not brought about by "rational argument" . 9 4 This is roughly i n l ine w i t h Kuhn's view, although altered somewhat to reflect Rort ian eliminative material ism. What we might approximate as Kuhn ian GestaZt-switches are not brought about directly by inference, or rational argument. However, for K u h n , reason does play some role i n the process. Rorty's admonishment that "we need make no more o f this switch than that it is not fully inferential downplays the dramatic significance that a scientific revolut ion can have. Also, that Rorty refers to these changes as GestaZt-switches betrays an understanding, or at least ought to, that things are experienced differently after the shift from before. Returning to Rorty's previously cited argument for solidarity's promoting the seeking of terms acceptable to both sides of a dispute, a GestaZt-switch precludes any possibility of there being a phrasing of things w h i c h is acceptable to both sides of a matrix-dispute. Us ing the familiar example, one cannot see a d rawing simultaneously as a duck and a rabbit. Ror t ian ethnocentrism prevents us from justifying our beliefs to everybody; justification extends only to those whose beliefs overlap our o w n to some appropriate extent. However, we are warned that this is not a theoretical problem about "untranslatability", but s imply a practical problem about the l imitations of argument: it is not that we l ive i n different worlds from the Nazis, say, or the Amazonians, but that conversions of point of v iew w i l l not occur as a result of inferences from previously shared premises. 9 5 Rorty's point here is that inference from previously shared premises is not open to us, but that this nonetheless does not mean that we go so far i n our divergence as to live in different worlds . Rorty makes reference to Quine's and Dona ld Davidson's hypothetical examples of an anthropologist s tudying jungle tribes, and attempting to communicate w i t h them. He states that when the natives' and our behavior in response to certain situations is pretty much the same, we think of both of us as recognizing the plain facts of how things are~the noncontroversial objects of common sense. But when these patterns of behavior differ wildly, we shall say that we have different Weltanschauungen, or cultures, or theories, or that "we carve up the world differently". But it would create fewer philosophical problems just to say that when these patterns differ, communication becomes harder and translation less helpful. Translation mav become so awkwardly periphrastic, indeed, that it w i l l save time simply to go bilingual. 6 It may be that Ror ty is wor r ied about the ramifications of Kuhn's thesis o f incommensurabil i ty among different scientific matrices (roughly extendible to different general discourses), given his remark that "[t]he strong point of Kuhn's Rorty, PMN, p. 325. Ibid, p. 332. Rorty, ORT, p. 3 In. Ibid, p. 104. 98 critics was that incommensurabi l i ty seemed to entail indiscussability". Inability to discuss things w o u l d be anathema to anyone whose thesis is buil t around the promotion of solidarity and tolerance through discourse. H e goes on to note that "ft]he strong point of [Kuhn's] defenders was that.. .nobody could answer Kuhn's challenge by explaining how commensurat ion was possible". But according to Rorty, n o w these "Kuhnian wars" seem to be d rawing to a close, since "both sides are coming to agree that untranslatability does not entail unlearnabili ty, and that learnabili ty is a l l that is required to make discussability possible" . 9 7 I th ink that K u h n w o u l d agree both that untranslatability does not entail unlearnabili ty, and that learnabili ty makes discussability possible. This does involve a process of becoming "bil ingual", and al though the proponents of entrenched theories w i l l at first attempt strenuously to achieve this by means of translation, they w i l l , i n the end, f ind that, i n order to have come to understand what the proponents of new theories are ta lking about, they have instead achieved something more l ike a GestaZt-shift; i n this, they come to realise that they have ceased to attempt translation, and have instead begun just to speak the new language. 9 8 But Rorty's conclusion that both Kuhn's critics and his defenders have come to realise "that untranslatability does not entail unlearnability, and that learnabili ty is a l l that is required to make discussability possible" seems to be acknowledging the presence of untranslatability. Otherwise, it is unclear w h y Rorty w o u l d entertain the argument that untranslatability does not, after a l l , preclude discussability. If untranslatability does not even enter into the question, then it is superfluous to invoke it i n any sort of argument w h i c h one deems relevant. If his contention is that, even i f alternative discourses were untranslatable--which they are not, according to him--there could sti l l be discussion among their proponents, then this should be spelled out more clearly, to head off any possible appearance o f ambiguity, especially given other areas o f Rorty's philosophy w h i c h are more clearly plagued by ambiguity. It looks, however, as though Rorty is not completely enough convinced of his o w n thesis of translatability, since he makes the c la im "that we are not ta lking about the same thing i f we say very different things about i t " . 9 9 Presumably this is meant to apply to what is the case for alternative discourses. I th ink that this is arguably reminiscent of the Kuhnian point concerning the way i n w h i c h alternative matrices carve up the w o r l d i n different ways, so to speak, generating ( in some sense) different worlds . If we are not ta lking about the same things, then translation must surely have become hopelessly elusive, and we have a situation of incommensurabil i ty. Translation w o u l d only be possible i f we were applying different terms to the same things. This need not bear an ontological stamp: we might a l l carve up the w o r l d conceptually i n the same way (using a common 9 7 ibid. 9 8 See pp. 36-7. 9 9 Rorty, ORT, p. 105. 99 taxonomy), but s imply do so employing different sets of terms (or lexicons). But then this leads us to a language-world separation w h i c h Rorty disallows. Referring back to the passage quoted above, I th ink it is s imilar ly unclear, i f translation between alternative discourses is considered to be possible (as Rorty considers it to be), why it w o u l d be "less helpful" i n cases where patterns of communicat ion differ significantly. Translation ought to be either possible or not, and i f it is no t -as K u h n argues, i n respect o f alternative scientific matr ices- then the only remaining option is to resort to a Gestdt-shift i n understanding. This is not necessary i f translation, however difficult, is s t i l l at least t ruly possible. (I a m assuming here that Rorty's exhortation, i n situations such as this, to "go bi l ingual" involves not translation, but a sort of GestaZt-shift.) Moreover, surely what we are concerned w i t h here is not a question of saving time, but of determining what is possible. Hermeneutics Before taking a look at the v iew of hermeneutics w h i c h Rorty espouses, and subsequently at how that measures up to what I take to be Kuhn's understanding of the role of hermeneutics i n his o w n programme, I w i s h to take a general look at the portrayal of hermeneutics w h i c h we find i n the work of W i l h e l m Dilthey, to serve as a background for Rorty and K u h n on this issue. It is not m y intention to evaluate this at a l l , but merely to present i t~albeit i n fairly rough and truncated form-as a basis from w h i c h to consider Rorty and K u h n . Dilthey's exploration of hermeneutics was directed towards reaching an understanding of wri t ten works based upon an imaginative consideration of the creative process of the author. The intent of an interpreter of a text was thus to include the purpose and meaning of the author i n the consideration of the work itself. Earl ier interpretive methodologies had conceived of the process i n logical-rhetorical terms; there were attempts to reduce li terary interpretation to a system of rules. Fr iedr ich Schleiermacher, whose work on the interpretation of b ibl ica l texts inspired much of Dilthey's work i n hermeneutics, introduced new concepts of receptivity and creativity into the process. The interpreter must become receptive to the mentali ty of the author, and to whatever were the features of the environment i n w h i c h that ind iv idual was wr i t ing . For this, the interpreter must invoke her o w n creative capacity i n generating the necessary empathetic sense for the development of this receptivity. She must empathetically identify w i t h the author. The idea here is that the interpreter moves her o w n being i n some way into the relevant historical setting of the author, by momentar i ly emphasising some mental processes, whi le a l lowing others to fade away. In this way, she is expected effectively to reproduce the al ien life of her author w i t h i n herself. Di l they termed 100 this process "transference". 1 0 0 But we cannot understand what we have not experienced; we can understand of the psychic states of others on ly as much as we know i n ourselves. Thus, the greater our o w n inner richness, the better-equipped we w i l l be to come to an understanding of others. This is the foundation of the understanding required for interpretation. Since there is no immediate communicat ion between individuals , such as the author and his interpreter, the latter must rely upon inference. The interpreter attempts to judge the relation of a work to the mentali ty of its author. This is achieved through the workings of the hermeneutic circle, w h i c h , i n general terms, entails a circular movement of interpretation between a whole and its parts: to quote H .P . Rickman, [The hermeneutic circle] arises in the understanding of complex wholes and their parts, because a whole can only be comprehended in terms of its parts while the latter acquire their proper meaning within the whole. Words and sentences are the most obvious example. We understand "hand me my clubs" by grasping the meaning of the individual words; but we can only select the appropriate meaning of "club" or discard the use of "hand" as a noun when we have an idea of what the whole sentence means. In practice we solve this problem by a to-and-fro, or shuttlecock, movement, though in simple cases we are hardly aware of i t . 1 0 1 So the interpreter strives to understand the author's whole work from the combinat ion of its ind iv idual words and phrases, as w e l l as to understand the indiv idual words and phrases based upon a consideration of the w o r k as a whole . For this, language offers a basic interpretive tool , one through w h i c h the inspired creativity of the interpreter can act. The creativity of the interpreter and that of the author meet as one, such that the interpreter is enabled to re-create and re-live that w h i c h is past. But notwithstanding the relative fluidity of the creative aspects of the interpretive process, a developed historical consciousness must ground the personal inspiration of the interpreter. The hermeneutic "procedure of circular exegesis...promises to provide a means of averting what has been called the commonest error of the intellectual historian: to wri te about things he does not really understand-things he has not ' internalized' and thought through for h i m s e l f ' . 1 0 2 One may object that an adequate degree of empathy cannot be generated between two individuals who may be quite significantly separated i n t ime and circumstance. Certainly, the lack of immediate communicat ion between individuals i n such situations, and the resulting reliance on inference, renders the possibility for the required degree of empathetic understanding somewhat doubtful. The hermeneutic v iew of Schleiermacher takes the personalities of both author and interpreter to be adequately similar, i n having been formed w i t h i n a Dilthey, SW, p. 454. Rickman's introduction to Dilthey, Swr, p. 10. Hughes, quoted in Ermarth, p. 10. 101 common human nature. O n this v iew, differences among individuals are not really radical ones of a qualitative sort; they merely reflect differences of degree of or relative emphasis on mental traits w h i c h are common to us a l l as human beings. It w i l l be interesting to note i n relation to Rorty that, for Dil they, hermeneutics is commit ted to the not ion of truth. Hence, it has an epistemic function: to promote knowledge of ourselves and of humanity, through an examinat ion of personal and cultural creativity. W e may extend this attitude towards truth to a consideration of the sciences, as H . P . R ickman has done: A n outstanding example [of this circularity] is the fact-already noted as crucially important for the human studies-that the thinking of individuals can only be understood by reference to the world of mind or the cultural sphere while comprehension of the latter involves knowing about the mental processes of individuals. This is, of course, logically unsatisfactory; but the scientist, concerned mainly with relationships between the general and the particular rather than the whole and its parts, is involved in an analogous circle. He can only have a general conception of what dogs are like by gaining knowledge of individual dogs, but could never recognize an animal as suitable for his study on dogs unless he had a general idea of what a dog i s . 1 0 3 I tu rn now to look at Rorty's plea for hermeneutics, and to examine h o w it relates to the above-described picture. Since Ror ty sees the desire to develop a theory of knowledge as a desire for constraint, his advocating that we give up the desire for confrontation and constraint means that we rel inquish the desire for an epistemology. He describes hermeneutics as an expression of this hope. He is careful to point out that it is not intended to be a successor to epistemology, a new discipline, or a programme for research . 1 0 4 Epistemology sees the hope of agreement as a token of the existence of common ground which, perhaps unbeknown to the speakers, unites them in a common rationality....For epistemology, to be rational is to find the proper set of terms into which all the contributions should be translated if agreement is to become possible. 1 0 5 Thus, epistemology unites people through "mutual interests i n achieving a common end". O n the other hand, For hermeneutics, to be rational is to refrain from epistemology-from thinking that there is a special set of terms in which all contributions to the conversation should be put~and to be willing to pick up the jargon of the interlocutor rather than translating it into one's Hermeneutics unites people through civi l i ty : there is no c o m m o n goal, no c o m m o n ground. The hermeneutical v iew sees the relations between various Rickman's introduction to Dilthey, Swr, p. 11. Rorty, PMN, p. 315. Ibid., p. 318. 102 discourses as those of strands i n a possible conversation, and presupposes no discipl inary matr ix w h i c h unites the speakers. The hope of agreement is never lost, so long as the conversation lasts. 0 7 W h i l e we get epistemic commensurat ion only where we already have agreed-upon practices of discourse-where we k n o w what is going on, and want to codify it, i n order to extend, strengthen, ground, or teach i t~we must be hermeneutic where we do not understand what is happening, and where we admit that, rather than being blatantly Whiggish about it. So this difference between epistemology and hermeneutics boils d o w n to a question of familiari ty. Ror t ian ethnocentrism dictates that we must operate wi th in , at least at the start of discussion, the belief-web of our o w n culture. Ror t ian hermeneutics works consistently w i t h this attitude: "that [it] inevitably takes some norm for granted makes it, so far forth, 'Whiggish ' ". But i t "proceeds nonreductively and i n the hope of p icking up a new angle on things"; thus, it can "transcend its o w n Whiggishness", w h i c h apparently is to say, that its Whiggishness becomes less blatant, and more provisional . It is l ike ly that Rorty w o u l d oppose this mitigated sort of Whiggishness to "the attempt (which has defined tradit ional philosophy) to explicate ' ra t ional i ty and 'objectivity' i n terms of conditions of accurate representation", w h i c h "is a self-deceptive effort to eternalize the normal discourse of the d a y . 1 0 8 According to Rorty, the hermeneutic not ion of knowledge is of culture as a conversation: we play back and forth between guesses about how to characterise particular elements (the parts) and guesses about the point of the whole situation (the whole) . W e cannot avoid the hermeneutic circle: we are unable to understand the parts of a strange culture, practice, theory, language, etc. wi thout having an idea of how the whole th ing works, and vice versa. Thus, hermeneutics is better seen as another way of coping, rather than as another w a y of knowing. "Hermeneutics does not need a new epistemological paradigm... .Hermeneutics, rather, is what we get when we are no longer epis temologica l" . 1 0 9 The typical Ror t ian emphasis, here, on the abandonment of epistemology is not consonant w i th the v iew of hermeneutics w h i c h we see i n the Dil theyan v i ew above. Hughes' reference to "the commonest error of the intellectual historian: to write about things he does not really understand" implies that it is possible to misunderstand historical writ ings, w h i c h further implies that it is possible to get things w r o n g i n history. This presents us w i t h the dis t inct ion between truth and falsity, a dichotomy w h i c h Rorty has rejected. Moreover, Di l they was commit ted to preserving an epistemic role for hermeneutics, such that its dedication to the not ion of truth served to promote knowledge of ourselves and of humanity. It w o u l d appear that R o r t y s account of hermeneutics focuses more on the aspect of the development and maximisat ion of solidari ty among contemporaries, than o n its role i n the interpretation of past historical works . But this latter 107 Ibid. 1 0 8 Ibid., pp. 11, 320-1. 1 0 9 Ibid., pp. 319, 325, 356. 103 function is of considerable importance i n Dilthey's account, and I believe that it is this historical purpose w h i c h K u h n has i n m i n d . For K u h n , as we have seen 1 1 0 , the proper approach of the historian to historical theor ies-such as scientific ones- is dictated by hermeneutics. One must evaluate each theory on the basis of its o w n context. The lack of an overarching context for evaluation precludes the God's-eye v i ew from w h i c h the judgment of truth might be applied to any particular theory, to the exclusion of a l l of the others. Hence, "Kuhn's c la im that no algori thm [i]s possible save a post factum and a Whiggish one (one w h i c h construct[s] an epistemology on the basis of the vocabulary or assumptions of the w i n n i n g side i n a scientific d i spu te ) . " 1 1 1 A l l of this relates to the incommensurabil i ty w h i c h imbues the relationship between alternative discourses, or scientific matrices, since it is this incommensurabi l i ty w h i c h necessitates hermeneutics. Relat ing this more precisely to the Dil theyan account, the way to evaluate Aristotel ian mechanics, for instance, relies upon an attempt empathetically to enter, as much as possible, the context i n w h i c h Aristotle was th inking and wr i t ing . Separate elements i n Aristotel ian theories w o u l d need to be understood i n relation to this entire context, and i n turn, we are aided i n formulating our sense of this context by virtue of particular elements i n Aristotel ian thought. The question w i l l be to what degree we are able to achieve this empathetic "transference"; we ought to expect there to be some l imi t to it, such that we w i l l never quite be able to understand Aristotel ian mechanics i n the way a contemporary of Aristotle w o u l d have. The difficulty of t rying to do so is consistent w i t h the greater divide between matrices that we see i n Kuhn's thesis of their incommensurabil i ty than we see between discourses i n Rorty's thesis of their inter-translatability. St i l l , the hermeneutic system reminds us at least to attempt to do this as far as we are able, rather than "blatantly" interpreting Aristotle strictly according to our o w n contemporary norms. In l ight of this last point, Rorty's c la im that "Aristotle and Gali leo both have to face the t r ibunal of our present beliefs before we shall cal l anything either said 'true' " 1 1 2 is antithetical to the spirit of hermeneutics as presented both by Di l they and his supporters, and by K u h n . This c la im makes it look as though Ror ty is espousing a most blatant sort of Whiggishness, one w h i c h is resistant to "pick[ing] up the jargon of the interlocutor rather than translating it into one's own" . It surely does not resemble the attitude of empathy w h i c h characterises Dilthey's and Kuhn's understanding of hermeneutics. It is difficult to imagine what Rorty could have i n m i n d here, i n l ight of the historical nature of the "discourses" being considered. His point, i n speaking of his ethnocentrism and the periodic re-weaving of beliefs, is that we ought to begin w i t h our o w n set of beliefs, and then re-evaluate them i n light of the beliefs of others, w h i c h we encounter i n discourse w i t h them. Direct discourse w i t h historical figures is obviously impossible. So, h o w do we come to understand historical figures, i f we insist on a conscious 1 1 0 See pp. 50-1. 1 1 1 Rorty, PMN, p. 324. 1 1 2 Rorty, ORT, p. 51. 104 adherence-even i f only in i t ia l ly~to our contemporary framework? Ethnocentrism (even the Ror t ian variety) is antithetical to hermeneutical empathy, as it applies to historical understanding. Rorty equates epistemology w i t h normal discourse (i.e., that w h i c h is able to be brought under a set of rules), and hermeneutics w i t h abnormal discourse (i.e., not ru le -bound) . 1 1 3 He labels hermeneutics as, roughly, a description of our study of the unfamiliar, whi le epistemology is, roughly, a description of our study of the f a m i l i a r . 1 1 4 This is not a dichotomy w h i c h we w o u l d find i n Di l they or Kuhn , since they both understand hermeneutics to be an epistemological tool , one w h i c h is essential for an historically sensitive account of epistemology. This sort of epistemological account embraces the study of the unfamiliar: specifically the s t u d y - i n historical sc ience-of unfamiliar epistemologies. Neither can hermeneutics be s imply opposed to rule-bound discourse, since, i n t rying to develop an empathetic understanding of a historical figure, we attempt to "internalise" the context surrounding that figure-including, i n the case of science, the rules of thought and practice w h i c h the scientist i n question accepted. However, hermeneutics can more safely be identified w i t h what we might cal l abnormal discourse as a generalisation of Kuhn's revolutionary science: K u h n has l ikened the hermeneutic process through w h i c h we attempt to understand previous normal discourses, such as Aristotel ian mechanics, to the pattern seen when the proponents of competing theories struggle to understand each other across a revolutionary divide. Di l they is s imilar to K u h n , i n v iewing this process as somewhat less rule- and logic-bound than others have perhaps envisioned it to be, and more influenced by a f luidly creative energy. But this is not to erase a l l presence of structure, or reason. Ror ty states, however, that, i f we draw the l ine between epistemology and hermeneutics as he has suggested, "then it seems clear that the two do not compete, but rather help each other out." The reason given for this is that "[n]othing is so valuable for the hermeneutical inquirer into an exotic culture as the discovery of an epistemology wri t ten w i t h i n that cu l t u r e . 1 1 5 This is a v i ew of things w h i c h is consonant w i t h Kuhn's, and it indicates the epistemic orientation of hermeneutics. But it is disharmonious w i t h the urgings, w h i c h we often encounter i n Rorty, to abandon the epistemological project altogether. As I have noted before, Rorty is ambiguous on this point. He goes on to say that "[njothing is so valuable for the determination of whether the possessors of [that exotic culture] uttered any interesting truths (by-what else?-the standards of the normal discourse of our own time and place [my emphasis]) than the hermeneutical discovery of h o w to translate them without making them sound l ike foo l s " . 1 1 6 But the point of hermeneutics is to prevent an interpretation of historical figures w h i c h w o u l d make them sound l ike fools 1 1 3 Rorty, PMN, p. 346-7. 1 1 4 Ibid., p. 349. 1 1 5 Ibid., p. 346. 1 1 6 Ibid. 105 precisely by, as much as possible, not, evaluating them using the standards of the normal discourse of our o w n t ime and place. This is w h y it is so valuable for the hermeneutic enquirer to encounter an epistemology i n historical works, since i t w i l l be this past epistemology w h i c h she w i l l a im to internalise (attempting to leave her o w n behind), i n t rying to understand historical theories. Rorty states that "Kuhn's examples of 'revolutionary 1 change i n science were, as he himself has remarked, cases of the sort w h i c h hermeneutics has always taken as its special assignment-cases i n w h i c h a scientist has said things w h i c h sound so si l ly that it is hard to believe that we have understood h i m p rope r ly" . 1 1 7 H e offers this quotation from Kuhn's The Essential Tension: When reading the works of an important drinker, look for the apparent absurdities in the text and ask yourself how a sensible person could have written them. When you find an answer,...when these passages make sense, then you may find that more central passages, ones you previously thought you understood, have changed their meaning. 1 1 8 It should be clear that apparent absurdities i n historical texts appear absurd w h e n one attempts to do just what Rorty is advising us to do: to begin looking at them through the eyes of our o w n circumstances and ways of th inking. The process of "ask[ing] yourself how a sensible person could have wri t ten [such absurdities]" is the very process of making the GestaZt-shift into al ien conceptual terrain; this is the process of hermeneutics, as construed by Dil they and K u h n . Rorty reports that Kuhnians speak of cases i n the history of science i n w h i c h the description of the problem to be solved changes, thus changing the "observation language" used to describe the "evidence". But he warns us that "[t]his is not to say that we cannot, retrospectively, describe the problems and the data of a l l earlier epochs i n a single, up-to-date, commensurat ing vocabulary". Rorty calls this "the abil i ty to commensurate by hindsight~the abi l i ty to say that what Aristotle was looking for was what Newton found". Nevertheless, this "should not mislead us into t rying to describe our favorite ancestors as using 'the hypothetico-deductive observational method' (as Hook sometimes characterizes 'scientific i n q u i r y ' ) " . 1 1 9 In Ror t ian terms, "commensuration by hindsight" w o u l d entail br inging a plural i ty of discourses under a set of rules wh ich w o u l d indicate h o w we might achieve agreement among them. But, i f our description of problems and our observation-language~our discourse, i n short-change, then we ought not to be able retrospectively to describe earlier discourses using a single, commensurat ing vocabulary, for this latter w o u l d be a neutral observation language, the existence of w h i c h both Rorty and K u h n have denied, and w h i c h is frankly contradictory to any change i n observation language as a corollary to a change i n discourse. It seems that we are to understand this "hindsight" as operating to re-contextualise Ibid, p. 323. Kuhn, ET, p. xii, quoted in Rorty, PMN, p. 323. Rorty, ORT, pp. 68-9. 106 a l l previous discourses into the context of our o w n . As I have explained, K u h n has shown the folly of this manoeuver i n a properly historical understanding of science. It is also antithetical to Di l theyan hermeneutics. It is difficult to see h o w we might be able to perform "commensuration by hindsight", since different webs of belief (particularly those of distant historical character) cannot be expected to fit into our contemporary context of belief without considerable warping, given that their natural contexts are different from our o w n . The incommensurabi l i ty of different discourses for w h i c h K u h n has argued i n the context of scientific matrices thus speaks against our re-weaving our web of belief to accommodate beliefs from significantly al ien sources. Hermeneutics, employed i n the study of history, does not make use of such a re-weaving, but rather must find a way to leap across the chasm of incommensurabil i ty w h i c h separates different historical discourses. This chasm is what renders them separate discourses, after a l l . It is peculiar, also, that Rorty states that, i n pre-Kuhnian philosophy of science, rational enquiry involved putt ing everything into a single, familiar context, such that any rational enquirer w o u l d agree that the sentences of this vocabulary were truth-value candidates . 1 2 0 G iven Rorty's rejection of pre-Kuhnian phi losophy of science, we might expect h i m , on the basis of this statement, s imilar ly to reject such a unifying context. But what he describes here as pre-Kuhn ian philosophy of science looks to be performing an identical function to his o w n "commensuration by hindsight". It should be clear from the discussion i n this section that Rorty's belief, as expressed i n point ten of m y introductory list, is inaccurate, given the radical divergence between Dilthey's construal of hermeneutics, w h i c h K u h n arguably embraces, and Rorty's version of hermeneutics. 120 Ibid, p. 95. 107 Conclusion M y intention has been to elucidate the many ways i n w h i c h Richard Rorty finds s imilar i ty between Thomas Kuhn's work and his own , and further, to demonstrate what I see to be a distinct and extensive lack of s imilari ty between them. It is m y contention that Rorty has misunderstood a considerable number of elements i n Kuhn's programme, as w e l l as the intent w i t h w h i c h it was offered. I w i s h to proceed n o w to provide a summary of these problems. Despite their mutual rejection of correspondence theories, Rorty and K u h n part ways i n respect of what they see as their replacement. Ror ty wishes to abolish a l l talk of m i n d and mental representations, whi le K u h n shows no interest i n do ing so. In connection w i t h this, Rorty disdains the not ion of observation, whi le K u h n retains a commitment to it, as the element of "discovery" (that w h i c h is "given" by Nature) w h i c h is an essential component of knowledge-acquisit ion. A significant issue w h i c h drives a wedge between the respective theses of Rorty and K u h n is that of the need (or lack thereof) for a conceptual framework w h i c h guides thought and action. For Kuhn , fol lowing as he does, a generally Kant ian type of programme, such an a priori framework is essential for the intel l igibi l i ty of the endeavours of sc ience-which is the activity that he is concerned to discuss. Rorty, on the other hand, is not always clear about whether he thinks that all traces of a guiding framework should be abandoned, or that there w i l l be some mutual understanding among the participants of a discussion regarding what is to count as acceptable, thus defining their particular culture's web of beliefs. A t any rate, Rorty fails to grasp the crucial importance, for K u h n , of a strong a priori conceptual framework. Hence, he sees K u h n as engaged i n a crusade parallel to his own , of rejecting such a framework. Importantly connected to this is the question of their respective attitudes towards science. The thrust o f Kuhn's work is clearly intended to clarify the way i n w h i c h science works, and has worked throughout its history. He has made specific attempts to delineate ways i n w h i c h science and other cul tural activities can be distinguished from each other. Ror ty appears not to have noticed this, having extrapolated what he sees as Kuhn's programme to cover a l l cul tural endeavours; he takes K u h n to be doing this, as w e l l . For Rorty, this new way of conceiving of cultural activity ( including science) illustrates a programme of "holistic pragmatism", where in the pattern of al l cul tural activities forms a sort of seamless whole, dr iven by pragmatic concern. I have tr ied to show that, not only is Rorty's explanation of h o w he sees pragmatism to work hopelessly vague, but his wr i t ing often betrays an approach w h i c h is arguably antithetical to pragmatic concern. This is at odds w i t h Kuhn's m u c h clearer explication of science as essentially a pragmatic activity. 108 Certainly, Kuhn's attempt to give credence to non-rational factors w h i c h have influenced the course of science throughout its history has been an important l ink, i n Rorty's eyes, between himself and K u h n . However, I believe that Rorty has over-estimated the degree to w h i c h K u h n leans on non-rational factors i n explaining both the historical course of science and its functioning at any given t ime. K u h n wants to mainta in that both rat ional and non-rational factors guide the work of scientific disciplines. Moreover, Rorty confuses this issue by re-defining "rationality" i n terms of solidarity and tolerance among the members of a culture, and of persuasiveness i n discussion (as opposed to what he sees to be the more coercive attitude of tradit ional rationality i n argument). He sees K u h n as basically an al ly i n this new orientation towards rationality. I th ink it is clear, however, that K u h n demonstrates no tendency to understand rationality i n non-traditional terms. Final ly , whi le both Ror ty and K u h n advocate use of the hermeneutic method, Rorty's intent is for people to apply it to attempts to "cope" w i t h their o w n and others' cultures; for K u h n , hermeneutics more specifically provides a methodology for attempting to understand, by reference to their historical context, scientific conceptual and practical schemes w h i c h are now conventionally considered out-of-date and inaccurate. I have argued that Rorty's construal of hermeneutics is at odds w i t h the more standard v i e w of it, common to both W i l h e l m Dil they (who d id extensive work i n the development of hermeneutics) and K u h n . M u c h of Rorty's difficulty here rests on his confusion about (and, again, idiosyncratic definit ion of) incommensurabil i ty between alternative views of the w o r l d . I believe that I have shown that Thomas Kuhn's extensive work i n the history and philosophy of science is not at a l l applicable to the sort of cul tural programme w h i c h Rorty is offering, i n terms of intent, scope, and fundamental content. 109 Bibliography 1. W i l h e l m Dilthey, Selected Writings, ed. and trans. H .P . R ickman (London: Cambridge Universi ty Press, 1976)-referred to i n the text as "Swr". 2. W i l h e l m Dilthey, Selected Works, Vol. 1: Introduction to the Human Sciences, eds. Rudo l f Makkree l and Frithjof R o d i (Princeton: Princeton Universi ty Press, 1989)-referred to i n the text as "SW". 3. Michae l Ermarth, Wilhelm Dilthey: The Critique of Historical Reason (Chicago, London: The Universi ty of Chicago Press, 1978). 4. John Horgan, The End of Science: Facing the Limits of Knowledge in the Twilight of the Scientific Age (Reading, Mass. : Addison-Wesley, 1996), 41-7. 5. Thomas S. Kuhn , International Encyclopedia of Unified Science, Vol. 2, No. 2: The Structure of Scientific Revolutions, Second Edition (Chicago and London: The Universi ty of Chicago Press, 1962)-referred to i n the text as "SSR". 6. T.S. Kuhn , Criticism and the Growth of Knowledge, eds. Imre Lakatos and A l a n Musgrave (London and N e w York : Cambridge Universi ty Press, 1970), 1-24, 91-195, 231-78~referred to i n the text as " C G X 1 ' . 7. Thomas S. Kuhn , The Essential Tension: Selected Studies in Scientific Tradition and Change (Chicago and London: The Universi ty of Chicago Press, 1977), 14, 225-239, 293-352-referred to i n the text as "ET". 8. Thomas S. K u h n , "Rationality and Theory Choice", i n The Journal of Philosophy, Vol. 80, No. 10, Oct. 1983, 563-70-referred to i n the text as "RTC" . 9. Thomas S. Kuhn , "The Road Since Structure", i n PSA 1990, Vol. 2 (East Lansing, Mich igan : Philosophy of Science Association, 1991), 3-13~referred to i n the text as "RSS". 10. Richard Rorty, Philosophy and the Mirror of Nature (Princeton, N e w Jersey: Princeton Universi ty Press, 1979)-referred to i n the text as "PMN". 11. Richard Rorty, Contingency, Irony, and Solidarity (Cambridge: Cambridge Universi ty Press, 1989)-referred to i n the text as "CIS". 12. Richard Rorty, Objectivity, Relativism, and Truth (Cambridge: Cambridge Universi ty Press, 1991)-referred to i n the text as "ORT". 110 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items