RIVISTA INTERNAZIONALE DI FILOSOFIA E PSICOLOGIA ISSN 2039-4667; E-ISSN 2239-2629 DOI: 10.4453/rifp.2014.0004 Vol. 5 (2014), n. 1, pp. 43-57
Studi
Epistemological Contextualism and Cultures of Knowledge Wolfgang Detel
Ricevuto il 22 dicembre 2012; accettato il 7 luglio 2013
█ Abstract My first aim in this article is to describe the origin of the notion of culture of knowledge due to the re- search activities carried out at the University of Frankfurt by the Research Group Culture of Knowledge and Social Change since 1999. In this context we examined the relation between knowledge and society and proposed the no- tion of culture of knowledge as a key-concept to emphasize that knowledge does always appear in a specific histori- cal form, and can be investigated only as a social practice. This key-concept turned out to be a helpful heuristic construct and had therefore a widespread diffusion as a general and extremely flexible category which can shed light on mutual relations between knowledge and culture in several historical contexts. Secondly, I will compare the approach taken in the culture of knowledge with other contextualist approaches of contemporary research areas like History of Science and Theory of Culture to show that it is particularly helpful in terms of connecting different fields. Finally, I will drive attention to a basic question concerning the approach of the culture of knowledge which remains still unanswered. The Frankfurter Research Group considered it as a model of knowledge alternative to all those positions in History of Science and Theory of Culture that assume the universal, trans-historical, and trans-cultural validity of fundamental forms of knowledge and cognitive abilities. I will discuss whether and to what extent these positions are really alternative, and for this purpose I will focus on a central point: whether the thesis of a radical historicity of knowledge undermines the very idea of universal rationality and to what extent. KEYWORDS: Cultures of Knowledge; Contextualism; History of Science; Epistemology; Models of Knowledge.
█ Riassunto Il contestualismo epistemologico e le culture epistemiche – In questo articolo intendo illustrare la na- scita della nozione di cultura epistemica nell’ambito delle indagini svolte presso l’Università di Francoforte dall’unità di ricerca Cultura epistemica e mutamento sociale a partire dal 1999. In questo contesto si è indagato il rapporto tra sapere e società e proposto la nozione di cultura epistemica come concetto-chiave per sottolineare come il sapere compaia sempre in una forma storica specifica, e come questo stesso possa essere indagato solo co- me pratica sociale. Questo concetto-chiave si è rivelato un concetto euristicamente fruttuoso e perciò nel frattem- po ha conosciuto ampia diffusione come categoria generale ed estremamente flessibile per gettar luce sui reciproci rapporti tra sapere e cultura in diversi contesti storici. Inoltre, intendo comparare l’approccio delle culture epistemi- che con altri approcci propri della ricerca contemporanea in arre come la storia della scienza e la teoria della cultura per mostrarne l’utilità nel relazionare tra loro campi diversi. Infine, voglio portare l’attenzione su un problema di fondo, ancora aperto, che riguarda l’approccio delle culture epistemiche. L’unità di ricerca francofortese ha inteso questo approccio come modello di conoscenza alternativo a tutte quelle posizioni nella storia della scienza e nella teoria della cultura che assumono la validità universale, trans-culturale e trans-storica, delle forme fondamentali del sapere e delle capacità cognitive. Intendo discutere se e fino a che punto queste posizioni siano realmente alter- native. Per questo mi concentrerò su un elemento centrale: se e fino a che punto con la tesi della radicale storicità del sapere venga intaccata l’idea stessa di una razionalità universale. PAROLE CHIAVE: Culture epistemiche; Contestualismo; Storia della scienza; Epistemologia; Modelli della conoscenza.
W. Detel - Institut für Philosophie - Universität Bremen - FB 9 - Enrique-Schmidt Straße 7 - 28359 - Bremen (D) () E-mail: [email protected]
Creative Commons - Attribuzione - 4.0 Internazionale
44 Detel
█ Epistemological contextualism tion that is context-sensitive, but also, for in- and skepticism stance, epistemic goals, articulated doubts and social expectations of meeting objections.4 In MOST PEOPLE AND MANY EPISTEMOLOGISTS particular, epistemic reasons R for belief p can are convinced that have defeators D such that D-and-R is not an (1) they know a lot, epistemic reason for p any more; however, de- (2) knowledge is infallible, and feators are context-sensitive.5 (3) knowledge claims are fallible. From this point of view epistemological con- These three thoughts create a serious para- textualism is often taken to be, by getting past dox that seems to imply scepticism.1 One way traditional epistemology, the best way of escap- of expressing this consequence is to maintain ing scepticism. In particular, epistemological that knowledge is an absolute concept. That is contextualists claim to be able to explain why to say, there is, in our world, no mental state most people and many epistemologists stick to that is absolute knowledge excluding all possi- thoughts (1)–(3). The explanation offered is ble evidence that it might not be knowledge. At that people neglect the context of knowledge most a mental state can be, it seems, called claims and therefore falsely think that proposi- knowledge if it is, as a matter of fact, an ap- tions expressed by knowledge claims uttered in a proximation to knowledge. But knowledge certain context are the same as propositions ex- does not come in grades. Therefore, an approx- pressed by the same knowledge claims uttered in imation to knowledge is not knowledge at all. a different context. Thus, in our world knowledge cannot be had. And this is exactly what the skeptic urges.2 █ Epistemological contextualism as relevant The most plausible diagnosis of this situa- alternatives account tion is that thoughts (1)–(3) come down to conceiving of knowledge as natural kind (as A popular form of epistemological contex- epistemological realists do) and trying to find a tualism is the relevant alternatives account ac- foundation for knowledge other than cording to which a claim to knowledge that p is knowledge itself (as epistemological founda- correlated with possible alternatives being in- tionalists do). That is, the paradox follows from compatible with p. Knowledge requires that we epistemological realism and epistemological exclude all these alternatives. But mobilize your foundationalism that are, in turn, the most imagination, and you will find possible alterna- fundamental assumptions guiding traditional tives that you cannot exclude. However, pro- epistemology. If we stick to traditional episte- ponents of the relevant alternatives account mology, we generate the paradox and are stress that some of these alternatives are irrele- forced to surrender to the skeptic.3 vant, while others may be relevant. Most im- One way of dealing with this skeptical threat portantly, which alternatives are relevant or ir- is to insist that knowledge is context-sensitive, i.e. relevant, respectively, is context-dependent.6 to urge that it depends on the context which For instance, in everyday life, whether a standards of justification are accepted and which warehouse can be called empty depends on premises are taken for granted and thus, wheth- whether there are still goods around or not. In er true beliefs count as knowledge or not. This this context, the possible alternative that there familiar kind of epistemological contextualism re- are still molecules left in the house is clearly not lies on the traditional JTB-account of knowledge relevant. In other contexts, however, this alter- (treating knowledge as justified true belief). Ac- native may be relevant, for instance if we look cording to epistemological contextualism at the warehouse as a gigantic vacuum- knowledge is some sort of true belief, but it is chamber. Therefore, it can be maintained that justification that makes knowledge context- in the everyday context, the warehouse is emp- sensitive. More precisely, it is not only justifica- ty in the full sense (in which empty is taken as
Epistemological Contextualism and Cultures of Knowledge 45 absolute concept) iff there are no goods around important ingredient of epistemological con- anymore. It follows that from the fact that A is textualism is a contextualist semantics for an absolute concept it cannot be derived that in knowledge sentences.8 our imperfect world there is nothing that is A in the full sense. In particular, then, the fact █ Standards of justification that knowledge is an absolute concept does not entail skeptical conclusions. Thus, according to In its standard version epistemological con- the relevant alternatives account, knowledge is a textualism does not reflect on the relations be- belief such that the possible relevant alterna- tween different epistemic contexts nor between tives can be excluded. In this sense, knowledge different standards of justification correlated can often be had.7 with different epistemic contexts. This is cer- However, there are two different sorts of tainly a theoretical deficit. There are some epis- possible alternatives to knowledge claims, one temological contextualists, though, who seem being determined by subjective factors concern- to have realized this problem. Shiffer, for ex- ing the knower, the other being defined by as- ample, distinguishes between two standards of cription factors concerning the speaker who as- justification, Tough (in science) and Easy (in cribes knowledge to somebody else. Epistemo- everyday life), respectively.9 David Annis, being logical contextualism, set out as relevant alter- a convinced epistemological contextualist, con- natives account, claims that not only subjective cedes, in addition, that there has been a refine- factors, but also ascription factors are context- ment in the historical development of methods sensitive. If S believes, and claims to know, that of discovery and testing, for instance in science, there is a barn in front of her, and if speaker S* and thus, that there are not only different considers whether she should ascribe know- standards of justification, but also differently ledge to S, in respect of this claim, then it is rel- refined standards of justification.10 evant for this knowledge claim whether S and Nevertheless Annis sticks to a relativist po- S* are in barn facades county or in a landscape sition. For he thinks that where there are only a few, if any, barn facades. (a) if contexts C and C* with correlated jus- It is in the former but not in the latter context tification standards S and S*, respectively, are that for the belief to be knowledge the possible differently refined (such that C* and S* are alternative of there being just a barn facade be more refined, and so epistemically more ade- eliminated. But what counts epistemologically quate, than C and S), and is whether S* knows what the context is. If S* (b) if person P is justified to claim to know knows that they are in barn facade county, then that p in C and under S, but not in C* and un- S* will demand that S excludes the possibility of der S*, then there being just a barn facade in order to be (c) P has knowledge of p in C and under S, able to ascribe knowledge to S, whatever S even if P does not have knowledge of p in C* thinks the context to be. and under S*. Some epistemological contextualists sup- The thought (a)–(c) articulates, as Annis porting the relevant alternatives account em- emphasizes, precisely that justification and phasize that ascription factors have an impact knowledge are context-dependent. But he does for the truth conditions or the meaning of not realize that argument (a)–(c) defines a knowledge sentences. So if ascription factors stronger version of epistemological contextual- are context-sensitive, this will go for the mean- ism, that is, that we can distinguish ings of knowledge sentences too. Let AS* and (EC 1) Weak epistemological contextualism: AS** be two different contextual ascription fac- (1) Knowledge depends on different contexts. tors, then the sentences S knows that p relative (2) The contexts mentioned in (1) are not dif- to AS* and S knows that p relative to AS** have ferently refined. different meanings. It turns out, then, that one (EC 2) Strong epistemological contextualism:
46 Detel
(1) Knowledge depends on different contexts. (i) the distinction between better and worse (2) The contexts mentioned in (1) are differ- knowledge requires the distinction between ently refined. better and worse epistemic contexts; However, (E1) is, to say the least, much more (ii) if a belief can be called knowledge in a plausible than (E2). worse epistemic context, but not in a better ep- Sometimes contextualists conceal the prob- istemic context, then, from the perspective of lem with (E2) by confusing knowledge and at- the better epistemic context, the belief could tributions of knowledge. For example, Cohen truly have been called knowledge in the worse wants to show that social standards are im- context, but this belief was not knowledge; portant for determining whether a person has (iii) if there is a worse and a better epistemic knowledge; but a little later he puts his claim by context available at the same time, if a belief saying that attributions of knowledge are con- can be called knowledge in the worse, but not text-sensitive.11 Of course P knows that p if P* in the better epistemic context, then the belief truly attributes the knowledge that p to P, but if in the worse epistemic context could not even S* attributes the knowledge that p to P, then it have been truly called knowledge in the worse does not follow that P knows that p. Further- context. more, we must, as we have seen, distinguish be- From this point of view epistemological tween the context the potential knower S is in contextualists can, and should, distinguish be- and the context the speaker S* who considers to tween three kinds of knowledge: ascribe knowledge to S is in. And it must also be Suppose person P believes truly that p, then decided whether one of these contexts puts its (a) P a-knows that p iff there is an available inhabitants into a better epistemic position epistemic context (including standards of justi- than the other. If so, claims to knowledge in the fication) C (i.e. a context the community S be- better context override claims to knowledge in longs to is familiar with) such that P is, relative the worse context – which is incompatible with to C, justified in believing that p. strong epistemological contextualism. (b) P b-knows that p iff P a-knows that p One of the most thoughtful epistemological and there is no available context C* better than contextualists, David Lewis, seems to agree C such that P is, relative to C*, not justified in with this conclusion by saying: believing that p. (c) P c-knows that p iff P b-knows that p and Better knowledge is more stable knowledge: there is no better context C** whatsoever such it stands more chance of surviving a shift that P is, relative to C**, not justified in believ- of attention in which we begin to attend to ing that p. some of the possibilities formerly ignored. Obviously, c-knowledge cannot be achieved. If, in our new shifted context, we ask what But a-knowledge and b-knowledge can be had, knowledge we may truly ascribe to our ear- and even if c-knowledge is beyond our reach, lier selves, we may find that only the better this does not exclude that much of what we a- knowledge of our earlier selves still de- know or b-know is in fact c-knowledge. One of serves the name. And yet, if our former ig- the failures of skepticism is not to see this. It fol- norings were proper at the time, even the lows that worse knowledge of our earlier selves could (a) saying that we know a lot is to say that truly have been called knowledge in the we a-know or b-know a lot (implying that we former context.12 have, so far, reasons to believe that this knowledge is in fact c-knowledge); █ Mild epistemological contextualism (b) saying that we are all good fallibilists is to say that c-knowledge cannot be known to The remark by David Lewis just quoted is have been reached. important. One should add, though, that There is nothing inconsistent in entertain-
Epistemological Contextualism and Cultures of Knowledge 47 ing both claims (a) and (b). If people feel un- can be fruitful and illuminating. However, happy about splitting the notion of knowledge, there are two sorts of social epistemology: Ver- no problem. We should not argue about termi- itistic and non-veritistic social epistemology. nology. We can likewise say that c-knowledge Non-veritistic social epistemology stresses is the notion of knowledge, while a-knowledge that empirical, historical and sociological stud- and b-knowledge are two important forms of ies of scientific practices show that science and context-sensitive justified defeasible ascriptions of scientific results do not depend exclusively on knowledge. In any case, it follows, obviously, the objective external world which, according that a-knowledge and b-knowledge come in to traditional epistemology and philosophy of grades, while c-knowledge remains an all-or- science, is supposed to make scientific beliefs nothing affair. More importantly, distinguish- true or false, but rather also (even mainly or ex- ing between better and worse epistemological clusively) on social arrangements resulting contexts and between three kinds of justifica- from negotiations between scientists taking tion provides the conceptual resources to in- place in the course of scientific practices. The troduce a notion of ideal or perfect knowledge crucial idea is that science, scientific (roughly c-knowledge). It is only against the knowledge, and scientific practices are socially background of such a notion of perfect determined. knowledge that we can talk about better or Non-veritistic social epistemology proposes worse contexts and about defeasibility condi- three fundamental theses:13 tions of context-sensitive justified knowledge (i) Contingency thesis: Natural sciences ascriptions. In this sense it is, in talking about could have developed in alternative ways. The contexts of knowledge, not only possible, but way natural sciences have in fact developed is even necessary to bring a notion of ideal not inevitable, but was to a certain extent con- knowledge into play. tingent.14 If epistemological contextualists accept that (ii) Nominalist thesis: At most some indi- (i) talking about contexts that are epistemo- vidual things, but not facts or universals, can be logically better or worse is possible and some- taken to exist independently of humans and times even necessary, their social actions. (ii) knowledge comes in three kinds (a- (iii) Stability thesis: Explanations for the knowledge, b-knowledge, and c-knowledge, stability of scientific beliefs must rely at least in according to (a)–(c) above), and part on social factors that are external to the (iii) to talk about context-sensitive episte- internal criteria of theory choice maintained in mological contexts we need a conceptual grasp modern natural sciences and articulated in ra- of what ideal perfect knowledge would look tionalist philosophies of science. like, The earliest prominent version of non- then epistemological contextualism turns veritistic social epistemology is the Edinburgh into a mild form, viz. mild epistemological con- school of the sociology of science (sometimes textualism. called the strong programme of the sociology of science). The proponents of this school urge █ Non-veritistic social epistemology that all sciences including mathematics and natural sciences are socially determined. More- Another brand of epistemological contextu- over, it is not only the historical development alism is social epistemology which looks not of science, its rise and success, that is influenced only at contexts of knowledge, but in particular by social forces; rather, it is also the content of also at contexts of science. Social epistemology accepted scientific beliefs that is determined by proceeds from the assumption that a sociologi- social factors or social interests being involved cal analysis of science and scientific knowledge in scientific practices. (that is, an analysis of social contexts of science) The four methodological principles guiding
48 Detel the Edinburgh school of the sociology of sci- of an extremely complex process involving ence are: mainly social causes. In this way, social episte- (i) Causality, i.e. exploration conditions that mology is not only a form of anti-rationalism, causally bring about claims to some sort of but also a form of anti-realism and relativism. knowledge. Another version of a non-veritistic social (ii) Impartiality, i.e. examination of success- epistemology is the actor-network theory.17 In ful as well as unsuccessful knowledge claims. this version it is claimed that scientific (iii) Symmetry, i.e. use of the same types of knowledge flows from established relations be- explanations for successful and unsuccessful tween objects, animals and humans engaged in knowledge claims alike. scientific practices. An actor is supposed to be (iv) Reflexivity, i.e. application of the theory everything that in some causal way affects the to itself.15 production of scientific theories: not only sci- From their sociological studies the Edin- entists and their actions, but also, for instance, burgh sociologists of science conclude that background assumptions, methodologies, there is no definite unique set of rational meth- techniques, social rules and institutions, rou- ods that guide scientific practices and can be tines, experiments, measurements and the ap- referred to in order to explain how scientific propriate instruments as well as scientific texts. results and beliefs are established. Another re- There are many kinds of interactions between sult is that a sharp distinction between the con- actors; in particular, some actors can transform text of discovery and the context of justifica- other actors. A network is a set of actors such tion is not helpful in sociological investigations that there are relations and transformations be- of scientific practices. In this way, the Edin- tween the actors that are stable, in this way de- burgh school rejects fundamental assumptions termining the place and functions of the actors held by the traditional philosophy of science. In within the network. Once a network has been terms of epistemological contextualism these established it implies a closure that prevents results seem to show that, not the distinction other actors or relations from entering the net- between the context of discovery and the con- work, thereby opening the possibility of accu- text of justification, but different standards of mulating scientific knowledge that is the result rationality are parts of epistemic contexts. of transformations within the network. Estab- On this view the sociological analysis of sci- lishing a scientific belief or theory or fact comes entific practices must be a causal explanation of down, from the point of view of the actor- belief formation: it is supposed to show in de- network theory, to placing these actors in a sta- tail how specific scientific beliefs are estab- ble network. In this sense, scientific beliefs, lished as a result of causal processes proceeding knowledge, and theories are taken to be con- from social conditions, social interests, and ne- structed by transformations taking place in es- gotiations between scientists.16 Therefore, so- tablished networks. Obviously, the actor- cial epistemology is, in the view of the Edin- network theory takes the actor-network to be burgh school, itself a kind of natural science. It the epistemic context of scientific knowledge.18 can also be formulated as a methodological A well-known general objection against claim mandating a naturalistic approach to sci- non-veritistic approaches to social epistemolo- entific practises. As such, social epistemology is gy is that the leading studies use empirical, so- best understood as being part of a naturalized ciological and historical evidence to justify fun- epistemology concentrating on the investiga- damental and far-reaching claims that are sup- tion of social causes of belief formation and is posed to debunk the epistemic authority and committed to the claim that scientific universal rationality of the sciences. Obviously, knowledge cannot simply be seen as an ade- in doing so they rely themselves on exactly quate representation of the external objective those scientific methods they try to debunk. world, but must rather be taken to be the result For instance, the claim that «there are no con-
Epistemological Contextualism and Cultures of Knowledge 49 text-free or super-cultural norms of rationali- epistemic practices and cultures of knowledge ty»19 can only be defended by carefully examin- seems to involve a relativism of rationality and ing the historical and cultural evidence as well as justification.25 In any case, veritistic social epis- by applying the hypothetical-deductive method. temology presents a fairly rich view about epis- temic contexts including communicational █ Veritistic social epistemology acts, institutional structures, journals, speech practices, ways of access to, and dissemination Veritistic social epistemology aims at identi- of, information, information technologies, legal fying and evaluating social interactions, prac- systems, division of cognitive labour, and social tices and forces that influence – in a positive or consensus practices. negative way – the pursuit of truth and the production of true justified beliefs.20 On this █ Theories of cultures of knowledge view, science and education place a positive val- ue (“veritistic value”) on having true beliefs ra- Another way of addressing the social aspects ther than false beliefs. Social interactions, prac- of science in is to develop theories of cultures of tices and forces that influence – in a positive or knowledge. The starting point of such theories is negative way – the pursuit of truth include the idea that knowledge is often intrinsically tied communicational acts (like discussions and cor- to epistemic cultures, and epistemic contexts can respondence between scientists), institutional be treated as a sort of culture. structures that frame such acts (like universities, This approach relies on a notion of culture journals publishing peer reviewed articles, or and is committed to outline this much debated conferences), speech practices (of reporting and notion. Basically we can distinguish universal, arguing), ways of access to different sorts of in- epistemic, semantic and social conceptions of formation (libraries, internet resources), market culture. In the universal sense, culture is the mechanisms regulating the flow of speech, in- whole set of human social activities; in the epis- formation technologies, legal systems, and ways temic sense, the culture of a social group is the of disseminating information.21 group’s entire knowledge;26 in the semantic Another social aspect of science is the divi- sense, culture is a system or net of meanings;27 sion of cognitive labour.22 For instance, it will and in the social sense, the culture of a social optimize the scientific pursuit of adequate and group is the group’s system of social rules.28 If successful theories if different members of a we want to reflect on cultures of knowledge, scientific community pursue different strate- then it seems obvious that the universal notion gies of solving a scientific problem, rather than of culture is too broad, while the epistemic and all work on the most promising strategy. Sci- semantic notions of culture are in danger of ence can, on this view, only make progress if it producing a circular definition of cultures of relies on a social consensus practice that in- knowledge. Therefore, cultures of knowledge cludes an exploitation of reliable scientific au- should be treated as a brand of cultures in the thorities, formal and empirical methodologies social sense. However, this sort of culture must and background assumptions.23 be worked out a bit more precisely: Basically, there is nothing wrong with ver- Cultures in the social sense itistic approaches to social epistemology. This (1) consist basically in certain human prac- is not to say, though, that there are no open tices some of which are practices of rule- problems with this view. For example, one of the following. topics that need to be addressed is in which way (2) These practices are usually correlated the pursuit, not of true beliefs, but of justified with background assumptions guiding the beliefs can be socially determined. In this con- methods and goals of the practices. text, the problem of reliable testimony must be (3) In addition – and very importantly – the discussed.24 In addition, the historical change of practices and background assumptions are
50 Detel handed down to following generations in a cumu- Finally, as a first move we can introduce the lative way by teaching and learning. notion of cultures of knowledge simply by identi- (4) Part of this historical process are power fying cultures of knowledge with epistemic cul- relations between teacher and student; in par- tures. In the framework of theories of cultures ticular, in social cultures regulative power is of knowledge, it is the epistemic culture that is prominent, viz. the power to get other persons supposed to be the epistemological context for to follow certain rules.29 states of knowledge. Alternatively, we can define cultures in the To investigate cultures of knowledge comes social sense as a set of people such that these down to analyzing the epistemic practices of a people, and only these people, exercise certain group of people and exploring the social reali- practices that satisfy conditions (1)–(4). Social zation of these practices. Examining cultures of cultures in this sense are not determined by local knowledge is an attempt, not to examine simp- or ethnic or national contexts, but rather by his- ly the social conditions in which theories did torical chains that can be reconstructed in terms rise and fall, as traditional social epistemology of theories of social learning.30 In particular, has it, but to look at the complex of epistemic since there are presumably no sets of practices practices, theories and ideals of knowledge, as such that all and only the people of a certain na- well as their social conditions and realizations. tion exercise these practices, there are probably The point is not to treat definitions of no national cultures in the social sense. knowledge as mere ideological epiphenomena, This notion of a social culture can easily be but to relate them to epistemic practices and specified to a notion of epistemic cultures: the social environment.33 Epistemic cultures (5) consist basically in certain human prac- █ Veritistic versus non-veritistic social tices (some of which are practices of rule- epistemology: A brief comparison following) aiming at producing, selecting, and disseminating (scientific) knowledge (true jus- In the last two decades, there has been an tified beliefs). ongoing controversial debate between propo- (6) Knowledge is taken to be a mental state, nents of veritistic and non-veritistic versions of i.e. a brain-state displaying correctness condi- social epistemology. Indeed, it is hard to imag- tions und thus, providing representations and ine how a consensus between the parties could operating in the space of reasons.31 be achieved. Actually, one can look at the two (7) These practices are usually correlated approaches as belonging to very different scien- with background assumptions determining the tific projects. The representational and sensi- methods and goals of producing knowledge tive mind is, as the modern theory of mind has and outlining a notion of what perfect it, realized in physical states and things. There- knowledge would look like. fore it is certainly possible to explore mental (8) In addition – and very importantly – the states and minded animals exclusively in terms practices and background assumptions indicat- of their physical realizers.34 This possibility is ed in (6) are handed down to following genera- exploited by non-veritistic approaches to social tions in a cumulative way by teaching and learn- epistemology. These approaches are looking ing the production of knowledge in education exclusively at causal relations between thinkers institutions like schools and universities.32 and elements of epistemological contexts. (9) Part of this historical process are power Therefore, they don´t care whether thinkers relations between teachers and students in have true or false beliefs, and they don´t care schools and universities; in particular, in epis- either whether thinkers develop successful or temic cultures regulative power is prominent, unsuccessful theories. For taking into account viz. the power to get other persons to follow the the difference between true and false beliefs or rules of producing knowledge. between successful and unsuccessful theories
Epistemological Contextualism and Cultures of Knowledge 51 comes down to treat thinkers, their beliefs, ena. So it may be a waste of time and energy for their theories and their epistemic activities as each party to continue debating with each other belonging to the space of reasons. instead of pursuing their specific objectives. As already mentioned above, one problem is that the non-veritistic approach to social █ Ingredients of epistemic contexts, epistemology cannot be applied to itself. An- extended disquotational truth, other problem is that this approach talks, and superjustifiable belief among other things, about negotiations and discussions between scientists as having impact Epistemological contextualists speculate a on what beliefs and theories will be selected lot about context-dependent knowledge, but do and accepted; these processes take place, not tell us what exactly they take epistemic con- though, in the space of reasons which indicates texts to be. From our discussion of different an inconsistency. A third problem is that non- brands of epistemological contextualism we veritistic approaches don’t have the resources can conclude, though, that epistemological for distinguishing between beliefs coming contextualists believe that the main ingredients about by, say manipulation or political repres- of epistemic contexts are a language L, specific sion and beliefs coming about in a free non- scientific presuppositions SC, epistemological repressive intellectual environment; that is to assumptions EA about what knowledge and say, these approaches cannot explore the norma- science in general are, a logic LO, a mathemat- tive dimension of belief formation (ways of how ics M, testing rules TR, and educational institu- belief formation should, and should not, come tions EI. So epistemic context C, taken as an ep- about). It must be conceded, however, that this istemic culture that can be, and is usually, physical stance toward belief formation can re- taught, learned, handed down and institution- sult in illuminating insights. At the same time, it alized, amounts basically to the structure C = must be stressed that this stance does not take
52 Detel been acceptable for S*, then by S*‘s lights, LO the same time be preserved in all translations and M would have been acceptable for S*.36 into other languages. Since languages are parts Viewed from C*, (c1)–(c5) are increasingly of contexts and often vary with contexts, epis- weaker cases of knowledge ascription. The up- temological contextualists need an extended shot of these distinctions is that if we make disquotational notion of truth (in short EDN- states of knowledge in a context dependent of truth). true knowledge ascriptions being located in a EDN-truth: p is EDN-true in language L iff different and better context (which is, I think, (a) p is the case, and (b) for any sentence q in very often the case), then we get a fine-grained language L* such that q is an adequate transla- structure of knowledge states consisting of c- tion of p, q is the case. knowledge, c1-knowledge (=b-knowledge), c2 This notion of truth is not verificationist, – c5 – kinds of knowledge, and a-knowledge. but it does have a sort of epistemological back- Mild epistemological contextualism implies ground insofar as it relies on an empirically not only a distinction between different kinds confirmed interpretation theory that deter- of context-dependent knowledge, but also mines which translations are adequate. EDN- (a) a notion of better or worse epistemic truth is a mild form of a contextualized notion contexts, and of truth that should be part of mild epistemo- (b) a notion of ideal perfect knowledge.37 logical contextualism. Point (a) requires, of course, to adopt some To turn now to the notion of perfect sort of universal standard of some sort of uni- knowledge (c-knowledge) that is also part of a versal rationality from which we are able to mild epistemological contextualism, the key to evaluate which context may be better or worse this notion is the idea, expressed for instance by than some other context. Form a contextualist Lewis, that perfect knowledge is knowledge point of view this may be a problematic as- that is acceptable in a context and survives new sumption. I cannot discuss this question here, shifts of context – only that it is the shift to all but I want at least to indicate that Davidson´s better contexts that perfect knowledge must be conception of rationality may be helpful here, supposed to be able to survive. We can restate since it makes room for contextual differences this idea as follows: of a basically universal rationality. Perfect knowledge: Let person S be in con- Point (b), however, needs some elaboration. text C and have belief b; then b is perfect (ide- Epistemological contextualists should be able al) knowledge iff to talk about the truth of beliefs independently (a) b is EDN-true, of their justification and context. Knowledge (b) b is justified according to
Epistemological Contextualism and Cultures of Knowledge 53
█ Veritistic approaches to social Here is a brief version of McDowell’s prob- epistemology and the problem of relativism lem: (a) The following assumptions seem to be Veritistic social epistemology proceeds compelling: from the assumption that mental states display (i) The idea that our thinking is representa- correctness conditions and are therefore cor- tional, i.e. correct-or-incorrect, implies that our rect-or-incorrect (in specific cases true-or- thinking is answerable to the empirical world, false). Furthermore, in a certain sense the JTB- viz. to empirical experience. account of knowledge is, as we have just seen, (ii) Minimal empiricism: Empirical experi- still part of the picture. At the same time, this ence constitutes a tribunal for justifying our approach is a sort of epistemological contextu- representational thinking about the world. alism. That is to say, while having correctness (iii) Justification operates in the space of conditions is taken to be an objective property reasons. of mental states and the JTB-account of (iv) Empirical experiences including percep- knowledge is taken to be a definition of tions are natural phenomena belonging to the knowledge that holds universally the proce- realm of nature. dures of determining and verifying whether a (b) Assumptions (i)–(iv) in (a) are incon- given claim to knowledge is correct or not is sistent. For if (iii) and (iv) are true, then (i) and seen as heavily context-dependent, i.e. relative (ii) cannot be correct. That is to say, the truth to specific epistemic contexts. This epistemo- of (iii) und (iv) makes it impossible to see that logical relativism39 relies, among other things, minimal empiricism is acceptable and that our on the following four premises: thinking is answerable to the empirical world.40 (a) Epistemic contexts are not better or The solution of this problem McDowell worse, but are epistemologically on a par. himself suggests relies basically on the old idea (b) Perceptive reaction might be in some that our perceptions (being the core of empiri- cases universal among human beings, but are cal experience) are language – laden and can in physiological states outside of the space of rea- this way belong to the space of reasons. So it is, sons and can therefore not belong to reasons according to McDowell, not perception, but justifying claims to knowledge. our linguistic reports of perceptive states that (c) Linguistic observation reports do belong constitute the empirical tribunal for our think- to reasons justifying claims to knowledge but ing about the world. are, concerning their truth-value, dependent on This is not an entirely new idea, though, specific epistemic contexts. and it does not get us want we need. For we (d) Epistemological foundationalism is need to see that perceptions themselves, not wrong, since there is no neutral (context-, lan- perception reports, can provide an empirical guage- and theory-independent) epistemic tribunal for our thinking about the word. That foundation for all kinds of knowledge. is, we must be able to see that perceptions, not As indicated above, premise (a) is doubtful. only perception reports, may belong to a sort of Claim (d), while being crucial for epistemologi- space of reasons. This intuition has been an in- cal contextualism and epistemological relativ- tellectual taboo in modern philosophy – at least ism, does not seem to be an independent prem- since Kant published the Critique of Pure Rea- ise, since it follows from (b) and (c). On the son. But modern cognitive psychology of per- other hand, (b) and (c), taken together, create ception shows that this intuition is correct. the so called McDowell-problem and must The general psychological picture of the pro- therefore be examined in some more detail. A cessing of distal stimuli leading to conscious percep- solution of the McDowell-problem is required tions consists of a number of different stages: for clarifying the relation between epistemolog- (1) There exists an external world consisting ical relativism. of objects and events having certain properties
54 Detel
(minimal realism). (b) We can distinguish between two kinds (2) Due to their properties, these objects of spaces of reason: and event generate distal stimuli (for instance For every X and Y, photons or sound waves with certain frequen- (i) the logical space of reasons consists of all cies) such that the properties of the objects and elements X and Y such that X logically justifies events can be mapped into the properties of the Y or vice versa; distal stimuli. (ii) the broad space of reasons consists of all (3) In the visual case (which serves in the elements X and Y such that X broadly justifies following steps as example), the distal stimuli Y or vice versa. are transformed into an inverted image on the (c) Perceptions retina (this goes also, for instance, for audi- (i) are representations (that is, are semanti- tion). cally contentful), (4) The retina contains photoreceptor cells (ii) are mostly reliable and correct, and that transform the image (that maps the prop- (iii) do not belong to the realm of nature, erties of the distal stimuli) into electrical signals but to the broad space of reasons. (proximate stimuli) in the nervous system such From (c) (iii) it follows that assumption (iv) that the properties of the distal stimuli can be that is part of generating McDowell’s problem mapped into the properties of the proximate is wrong, which makes the assumptions (i) – stimuli (so called transduction). (iv) that are supposed to generate McDowell’s (5) Through a series of interconnected neu- problem consistent.44 rons that are located throughout the body, elec- More importantly, it follows also that per- trical signals are propagated from the receptors ceptions can broadly justify linguistic observa- cells to the brain (so called transmission). tion reports that can, in turn, be used to verify (6) The brain transforms the proximate or falsify scientific hypotheses. Most epistemo- stimuli (that is, the electrical signals) into percep- logical contextualists are inclined to object that tions specifying the semantic content of the per- even if the picture outlined so far is acceptable, ception (must still be seen as a sort of wonder). linguistic descriptions of perceptions are for- (7) Part of this process is an analysis of prop- mulated in specific languages that are part of erties and a formation of separate perceptive ob- different epistemic contexts. However, if one jects and events relying on psychological group- proceeds from a broadly Davidsonian picture ing mechanisms.41 The resulting perceptions of natural languages (as I do) and is impressed and emotions are representational states display- by Davidson’s arguments against the possibility ing sublinguistic semantic content and thus, cor- of conceptual schemes (as I am), then this ob- rectness conditions.42 jection is wrong. For then it follows that lan- (8) The neural processing of distal stimuli has guages, while being parts of epistemic contexts, been (as is shown by teleosemantic approaches)43 cannot alone by themselves create different ep- developed under evolutionary conditions and has istemic contexts because there are no languages therefore been sufficiently tested. that cannot be translated into other languages. This picture makes room for some im- All in all, then, we are entitled to assume a portant assumptions: perceptual foundation for all other kinds of (a) We can distinguish between two kinds knowledge and so stick to a version of episte- of justification: mological foundationalism. This means, among (i) X rationally justifies Y iff X and Y are other things, that the ideas that our thinking is propositionally contenful mental states or sen- representational, that, furthermore, minimal tences and X is a good reason for Y or X even empiricism is true, and that, finally, we must logically implies Y. distinguish between the spaces of reason and (ii) X broadly justifies Y if X is a reliable test the realm of nature, are consistent. Henceforth, procedure for Y. McDowell’s problem disappears.
Epistemological Contextualism and Cultures of Knowledge 55
Do these conclusions make epistemological 8 This semantic way of clarifying epistemological contextualism and epistemological relativism contextualism is the essential contribution we completely obsolete? These conclusions cer- owe to DeRose. See K. DEROSE, Contextualism tainly do not exclude mild epistemological con- and Knowledge Attributions, in: «Philosophy and textualism. This is because, as we have seen, Phenomenological Research», vol. LII, n. 4, 1992, epistemic contexts do not consist only of ob- pp. 913-929; K. DEROSE, Solving the Skeptical Problem, in: «The Philosophical Review», vol. servation reports and languages, but contain a CIV, n. 1, 1995, pp. 1-52. number of further ingredients that can, and 9 As Schiffer does, see S. SCHIFFER, Contextual So- will, vary historically and culturally from con- lutions to Scepticism, in: «Proceedings of the Aris- text to context. It must be emphasized, though, totelian Society», vol. XCVI, 1997, pp. 317-333, that it is only a restricted epistemological rela- here p. 318f. tivism (restricted by the existence of a percep- 10 The claim that there are contexts that promote tive foundation of knowledge) that is part of knowledge more or less well is also defended in mild epistemological contextualism. This goes social theories of knowledge that call contexts so- for theories of cultures of knowledge too. It cial factors (which include methods, standards, seems to me that the contextualist JTB-account and a lot of more things). The most sophisticated of perfect knowledge outlined in section 9 social theory of knowledge is in A. GOLDMAN, above articulates a possible compromise be- Knowledge in a Social World, Oxford 1999. 11 S. COHEN, Knowledge, Context, and Social tween mild epistemological contextualism and Standards, cit., p. 3. epistemological foundationalism. 12 D. LEWIS, Elusive Knowledge, cit., p. 562f. 13 See I. HACKING, The Social Construction of █ Notes What?, Harvard University Press, Cambridge
(MA) 2000. 1 See David Lewis’ remarks at the beginning of D. 14 For an important historical case see S. SHAPIN, LEWIS, Elusive Knowledge, in: «Australasian S. SHEFFER, Leviathan and the Air Pump. Hobbes, Journal of Philosophy», vol. LXXIV, n. 4, 1996, Boyle, and the Experimental Life, Princeton Uni- pp. 549-567. versity Press, Princeton 1985. 2 See P. UNGER, A Defense of Skepticism, in: «Phil- 15 See D. BLOOR, Knowledge and Social Imagery osophical Review», vol. LXXX, n. 2, 1971, pp. (1976), Chicago University Press, Chicago 1992; 198-219; P. UNGER, The Cone Model of K. KNORR CETINA, The Manufacture of Knowledge, in: «Philosophical Topics», vol. XIV, Knowledge: An Essay on the Constructivist and n. 1, 1986, pp. 125-178; P. UNGER, Philosophical Contexualist Nature of Science, Pergamon, Oxford Relativity, Minnesota University Press, Minneap- 1981; B. LATOUR, S. WOOLGAR, Laboratory Life. olis 1984. The Construction of Scientific Facts, Princeton 3 See M. WILLIAMS, Unnatural Doubts. Epistemo- University Press, Princeton 1986. logical Realism and the Basis of Scepticism, Oxford 16 See M. KUSCH, Knowledge by Agreement. The University Press, Oxford 1991. Programme of Communitarian Epistemology, Ox- 4 See D.B. ANNIS, A Contextualist Theory of Epis- ford University Press, Oxford 2002; D. BLOOR, B. temic Justification, in: «American Philosophical BARNES, J. HENRY, Scientific Knowledge: A Socio- Quarterly», vol. XV, n. 3, 1978, pp. 213-219. logical Analysis, Continuum, Chicago 1996. 5 See S. COHEN, Knowledge, Context, and Social 17 See B. LATOUR, S. WOOLGAR, Laboratory Life, Standards, in: «Synthese», vol. LXXII, n. 1, 1987, cit.; B. LATOUR, Science in Action. How to Follow pp. 3-26. Scientists and Engineers Through Society, Harvard 6 See A. GOLDMAN, Discrimination and Perceptual University Press, Cambridge (MA) 1987. Knowledge, in: «The Journal of Philosophy», vol. 18 The actor-network theory shares a number of LXXIII, n. 20, 1976, pp. 771-791; F. DRETSKE, basic assumptions with social constructivism as The Pragmatic Dimension of Knowledge, in: conceived in the Edinburgh school. Thus, both «Philosophical Studies», vol. XL, n. 3, 1981, pp. approaches entertain a naturalistic account of sci- 363-378. entific practices, do not presuppose a distinction 7 See D. LEWIS, Elusive Knowledge, cit. between true or successful and false or unsuccess-
56 Detel
ful scientific beliefs, and reject the possibility of a Tango: Beyond Reductionism and Non- rational reconstruction of scientific practices and Reductionism. The Epistemology of Testimony, in: their outcomes. This is sometimes declared the J. LACKEY, E. SOSA (eds.), The Epistemology of Tes- end of knowledge (S. FULLER, Social Epistemology, timony, Oxford University Press, Oxford-New Indiana University Press, Bloomington 1988; S. York 2006, pp. 160-191; P. PETTIT, When to Defer FULLER, Philosophy, Rhetoric, and the End of to Majority Testimony — and When Not, in: Knowledge.The Coming of Science and Technology, «Analysis», vol. LXVI, n. 291, 2006, pp. 179-187. Winsconsin University Press, Madison 1993). 25 See H. LONGINO, Science as Social Knowledge. 19 B. BARNES, D. BLOOR, Relativism, Rationalism, Values and Objectivity in Scientific Inquiry, and the Sociology of Knowledge, in: M. HOLLIS, S. Princeton University Press, Princeton 1990; H. LUKES (eds.), Rationality and Relativism, The LONGINO, The Fate of Knowledge, Princeton Uni- MIT Press, Cambridge (MA) 1982, pp. 21-47, versity Press, Princeton 2002. here p. 27. 26 See R. POSNER, Kultur als Zeichensystem. Zur 20 See A. GOLDMAN, Epistemics: The Regulative semiotischen Explikation kulturwissenschaftlicher Theory of Cognition, in: «The Journal of Philoso- Grundbegriffe, in: A. ASSMANN, D. HARTH (eds.), phy», vol. LXXV, n. 10, 1978, pp. 509-523; A. Kultur als Lebenswelt und Monument, Fischer, GOLDMAN, Foundations of Social Epistemics, in: Frankfurt a.M. 1991, pp. 37-74. «Synthese», vol. LXXIII, n. 1, 1987, pp. 109-144; 27 See C. GEERTZ, Thick Description. Toward an A. GOLDMAN, Knowledge in a Social World, cit. Interpretative Theory of Culture, in: C. GEERTZ, 21 A. GOLDMAN, J. COX, Speech, Truth, and the The Interpretation of Cultures. Selected Essays, Free Market for Ideas, in: «Legal Theory», vol. II, Hutchinson, London 1975, pp. 3-30. n. 1, 1996, pp. 1-32; A. GOLDMAN, Social Episte- 28 See L. KÄSER, Fremde Kulturen. Eine Einführung in mology (2001), in: E. ZALTA (ed.), Stanford Ency- die Ethnologie, Erlanger Verlag, Bad Liebenzell 1997. clopedia of Philosophy, updated version 2006 URL 29 See M. FOUCAULT, L’archéologie de savoir, Galli- http:// mard, Paris 1969. http://plato.stanford.edu/archives/sum2010/ent 30 See M. VOGEL, Kultur, Medien, in: S. DIETZ, T. ries/epistemology-social; S. FULLER, Social Epis- SKANDRIES (eds.), Mediale Markierungen. Studien zur temology, cit.; F. SCHMITT, Socializing Epistemolo- Anatomie medienkultureller Praktiken, Transcript, gy: An Introduction through Two Sample Issues, in: Bielefeld 2007, pp. 45-82. F. SCHMITT (ed.), Socializing Epistemology. The 31 Following the work of authors like Sellars, Social Dimensions of Knowledge, Rowman & Lit- McDowell and Brandom the space of reasons is telfield, Lanham (MD) 1994, pp. 1-28. distinguished from the realm of nature to indicate 22 P. KITCHER, The Division of Cognitive Labor, in: that mental states are not just natural states be- «The Journal of Philosophy», vol. LXXXVII, n. longing to the realm of nature (see, e.g., R. BRAN- 1, 1990, pp. 5-22; P. THAGARD, Collaborative DOM, Knowledge and the Social Articulation of the Knowledge, in: «Noûs», vol. XXXI, n. 2, 1997, Space of Reasons, in: «Philosophy and Phenome- pp. 242-261. nological Research», vol. LV, n. 4, 1995, pp. 895- 23 See P. KITCHER, The Advancement of Science. Sci- 908; J. MCDOWELL, Knowledge and the Internal, ence without Legend, Objectivity without Illusions, Ox- in: «Philosophy and Phenomenological Re- ford University Press, Oxford-New York 1993. search», vol. LV, n. 4, 1995, pp. 877-893). 24 See C. COADIE, Testimony. A Philosophical 32 See also A. PICKERING, Science as Cultural Prac- Study, Oxford University Press, Oxford 1992; E. tice, Chicago University Press, Chicago 1992. FRICKER, Telling and Trusting: Reductionism and 33 See W. DETEL, Wissenskultur, in: R. SCHÜTZEI- Anti-Reductionism in the Epistemology of Testimo- CHEL (Hrsg.), Handbuch Wissenssoziologie und ny, in: «Mind», vol. CIV, n. 414, 1995, pp. 393- Wissensforschung, UVK, Konstanz 2007, pp. 670- 411; E. FRICKER, Rational Authority and Social 679; J. FRIED, M. STOLLEIS (Hrsg.), Wissenskultu- Power: Towards a Truly Social Epistemology, in: ren. Über die Erzeugung und Weitergabe von Wis- «Proceedings of the Aristotelian Society», vol. sen, Campus Verlag, Frankfurt a.M. 2009. CXCII, 1998, pp. 159-177; A. GOLDMAN, Experts: 34 Similarly, it is possible to treat human beings Which Ones Should You Trust?, in: «Philosophy politically as physical things, for instance as mere and Phenomenological Research», vol. LXIII, n. instruments or as things belonging to the proper- 1, 2001, pp. 85-110; J. LACKEY, It Takes Two to ty of masters.
Epistemological Contextualism and Cultures of Knowledge 57
35 X in C is acceptable for S* in better C* if either X is and the Sociology of Knowledge, cit. valid in C* too or C is not valid in C*, but there is a X* 40 See J. MCDOWELL, Mind and World, Harvard Uni- in C* that can be substituted for X to justify b. So if X versity Press, Cambridge (MA) 1996, pp. xi-xvii. in C is not acceptable for S* in better C*, X is not valid 41 See J. ANDERSON, Kognitive Psychologie, Spect- in C* and cannot be substituted by an X* in C* either rum Verlag, Berlin-Heidelberg 2007. For a full in attempts to justify b. blown psychological theory of perception see D. 36 The way these different states are described MARR, Vision. A Computational Investigation into here relies heavily on the claim that some context the Human Representation and Processing of Visual parameters depend on others, i.e. that context pa- Information, Freeman, San Francisco 1982. rameters are not completely independent of each 42 In recent times a detailed semantics for sublin- other. guistic entities like perceptions (called teleoseman- 37 See points (i)–(iii) in section 4 tics) has been worked out, see R. MILLIKAN, Lan- 38 It should be noted that the notion of acceptance guage, Truth, and Other Biological Categories, The used in K(c) is the one explained in footnote 33 MIT Press, Cambridge (MA) 1984. (see supra). 43 See supra, fn. 40. 39 See B. BARNES, D. BLOOR, Relativism, Rationalism, 44 Compare supra, p. 51.