2. Boghossian's Attack on Epistemic Pluralism

Space for this epistemic pluralism seems to vanish, though, in the oscillation between two views that have largely shaped contemporary discussion: the view, on the one hand, that that science is a single, unified, discipline that discovers a single objective world according to a uniquely valid set of objective epistemic procedures, and the view, to the contrary that truth is relative to the interests, perceptions, background commitments, and values of disparate communal groups. The most influential advocate of the first view in recent times has been Paul Boghossian and of the second, Richard Rorty. In Fear of Knowledge, Boghossian meticulously constructs an argument against epistemic pluralism. The target of his argument is the relativistic view that ‘if our judgments about what it's “rational” to believe are to have any prospect of being true, we should not claim that some belief is justified absolutely by the available evidence, but only that it is justified relative to the particular epistemic system that we have come to accept’ (Reference Boghossian2006: 62). Such a view would seem to ‘give immediate support to the idea that there are many radically different, yet equally valid ways of knowing the world’ (62). A relativist ought not say that there are many radically different, yet equally rational, ways of knowing the world, because ‘that would amount to endorsing a use of ‘rational’ that is absolute, whereas the relativist view on offer is precisely that we cannot sensibly speak of what is rational, period, but only of what is rational relative to this or that accepted epistemic system’ (63n5).

The notion of an ‘epistemic system’, in Boghossian's usage, is that of a collection of epistemic principles, epistemic principles being defined as ‘general normative propositions which specify under which conditions a particular type of belief is justified’ (Reference Boghossian2006: 85). There are ‘generation’ principles, which generate a justified belief on the basis of something that is not itself a belief, and there are ‘transmission’ principles, which prescribe how to move from some justified beliefs to other justified beliefs (65). Again, there are ‘fundamental’ epistemic principles, principles ‘whose correctness cannot be derived from the correctness of other epistemic principles’ (65), and ‘derived’ epistemic principles, whose correctness can be so derived. The way of fixing beliefs that we call ‘science’, Boghossian suggests, is but a rigorous application of certain ‘ordinary, familiar’ fundamental epistemic principles. In particular:

And perhaps also

Might there be epistemic systems other than the one for which these ‘ordinary, familiar’ epistemic principles provide a conception of justification? Perhaps, for example, premodern Christian societies took as fundamental the epistemic principle Revelation:

Likewise, the Azande, it would appear, employ a different epistemic principle, Oracle:

The three epistemic systems, modern science, Christianity, and Azande, appear to employ divergent underived epistemic principles, and that might seem to motivate epistemic relativism, a view that Boghossian defines as the conjunction of three claims (Reference Boghossian2006: 73, emphases original):

1. A. There are no absolute facts about what belief a particular item of information justifies. (Epistemic non-absolutism)

2. B. If a person, S's, epistemic judgments are to have any prospect of being true, we must not construe his utterances of the form ‘E justifies belief B’ as expressing the claim E justifies belief B but rather as expressing the claim According to the epistemic system C, that I, S, accept, information E justifies belief B. (Epistemic relationism)

3. C. There are many fundamentally different, genuinely alternative epistemic systems, but no facts by virtue of which one of these systems is more correct than any of the others. (Epistemic pluralism)

Two objections to Epistemic relationism are now countenanced. First, propositions of the form E justifies belief B are normative, they make claims about what one should believe given certain evidence, whereas propositions of the form According to the epistemic system C, E justifies belief B are purely descriptive, they merely document the logical implications of a given epistemic system (Reference Boghossian2006: 75). A purely factual remark about what an epistemic system requires has come to replace a normative claim. Second, according to the relativist, the absolute claim E justifies B must be false, because justification is never absolute but only relative to an epistemic system. The objection is that epistemic principles are general normative statements about what beliefs are justified by what sorts of evidence. If, therefore, particular normative statements of this sort express something false, then the epistemic principles, too, must be false: ‘Given that the propositions which make up epistemic systems are just very general propositions about what absolutely justifies what, it makes no sense to insist that we abandon making absolute particular judgments about what justifies what while allowing us to accept absolute general judgments about what justifies what. But that is, in effect, what the epistemic relativist is recommending’ (2006: 87). Again,

The view I defend here affirms pluralism but rejects relativism, and I thus agree with Boghossian in his arguments against (B). Note, though, his merging two separate claims under the general label ‘Epistemic relativism’. Boghossian's argument against relativization is an argument only against what he terms ‘Epistemic relationism’ and does not yet speak to Epistemic pluralism (C). Boghossian does present an independent argument against pluralism, which he defines as the claim that there are many fundamentally different, genuinely alternative epistemic systems, but no facts by virtue of which one of these systems is more correct than any of the others. Suppose that one epistemic system, C1, employs epistemic principles that imply if E, B is justified, while another epistemic system, C2, employs epistemic principles that imply it is not the case that if E, B is justified. How can it be, in this circumstance, that there are no facts by virtue of which one system is more correct than the other, Boghossian asks. If there are no absolute facts about justification, then C1 makes a false claim, and C2 claims something true. More generally, if we take any contradictory pair of epistemic systems, ‘if one of them is deemed to say something false, the other will have been deemed to say something true. Under those circumstances, it's hard to see how it could be right to say that there are no facts by virtue of which one epistemic system could be more correct than any other’ (Reference Boghossian2006: 91). Boghossian's target is the thesis he terms Equal Validity: ‘There are many radically different, yet “equally valid” ways of knowing the world, with science being just one of them’ (Reference Boghossian2006: 2). His argument is that the very idea of a plurality of epistemic systems, each encoding a particular conception of epistemic justification, is incoherent.

Boghossian mischaracterizes the view of an epistemic pluralist. He is wrong to claim that pluralism about epistemic cultures is reducible to a pluralism about epistemic systems, as these have been defined by him, namely as sets of general normative propositions which specify under which conditions a particular type of belief is justified. Reflection on the nature of the epistemic pluralism of the Sanskrit knowledge disciplines makes this evident. Indian epistemology in general is an analysis of pramāṇas, methods for interrogating reality, sources of warranted belief. A pramāṇa is, more or less, what Boghossian means by an epistemic principle. The Indians were perfectly aware of the distinction between generative principles and transmission principles, and they would have chastised Boghossian for failing to mention an important transmission principle, Testimony:

Their names for Observation, Deduction, Inference to the Best Explanation, and Testimony are, respectively, pratyakṣa, anumāna, arthāpatti, and śabda. Yet they might have forgiven him, for they also discussed and disagreed among themselves whether Testimony is a fundamental or a derived epistemic principle, and they were, in general, fully cognizant of the importance of establishing a basic set of underived epistemic principles. Other putative epistemic principles, pramāṇas, were entertained, and much discussion took place around the question of their status—for example, whether they are derivable from more basic epistemic principles and whether they ought to count as epistemic principles at all. Indian versions of Revelation and Oracle, for instance, were largely dismissed. Yet the crucially important point is that although the different Sanskrit epistemic cultures disagreed with each other about what the underived epistemic principles are, they agreed that there is just one correct set of such principles. That is, they agreed about there being just one epistemic system, even though they disagreed about what constitutes it. Thus the epistemic pluralism that the Indian tradition displays cannot correctly be described as a pluralism about epistemic systems, in Boghossian's sense of the term.

3. Stance Pluralism

In fact, the nature of the epistemic pluralism on display has already been analyzed for us, and by the Sanskrit tradition itself. The remarkable Jaina philosophers make a distinction of fundamental epistemological significance when they say that as well as and in addition to pramāṇas, epistemic principles, there are also nayas, epistemic standpoints or stances, and that both are essential constituents of an epistemic culture (Umāsvāti, TS 1.6, in Tatia Reference Tatia1994); (Siddharṣigaṇi, NAV 29.28, in Balcerowicz Reference Balcerowicz2001: 124). A naya is not a proposition but a practical attitude, a strategy or policy which guides inquiry: it is an approach to the problem of producing knowledge, not a thesis about the sources of justification. Thus: ‘Among these, the [general] definition is as follows: “The reflection of one facet of an object recognised by a pramāṇa is the standpoint”, because this [general definition] pertains to all particular standpoints and because it is capable of distinguishing [among standpoints of] different forms’ (NAV 29.12, in Balcerowicz Reference Balcerowicz2001: 97). One such policy might be to attend only to what is immediately present in experience, another might be to enumerate everything one encounters without making any categorical distinctions, another to attend to stasis rather than flux, or vice versa. To see that stances are not propositions, we need only reflect on the epistemic stance adopted by Nāgārjuna, the Buddhist Mādhyamika, who denied that there is any way to say what nature is in itself (svabhāva). Nāgārjuna was accused of refuting himself, for if his epistemic proposition that everything is thus empty then that proposition should be itself empty, that is, without meaning in itself. His response was that he held no proposition, that emptiness is not a proposition, indeed that it would be a fatal error to mistake adopting emptiness as a philosophical position for belief in any philosophical proposition (Nāgārjuna, VV 29 in Bhattacharya, Johnson, and Kunst Reference Bhattacharya, Johnson and Kunst1978: 113). And this might remind one immediately of Bas van Fraassen's argument that, as a position in the philosophy of science, empiricism is not a propositional thesis, for it if were then since it claims that every thesis is open to empirical confirmation or disconfirmation it would itself be open to empirical confirmation or disconfirmation (van Fraassen Reference van Fraassen2002; see also Doctor Reference Doctor2014). To put it in Boghossian's terminology, someone who claims that Observation is the only underived epistemic principle would have to regard Observation as itself rationally justified on the basis of observation. van Fraassen's response is to say that ‘[a] philosophical position can consist in a stance (attitude, commitment, approach, a cluster of such—possibly including some propositional attitudes such as beliefs as well). Such a stance can of course be expressed, and may involve or presuppose some beliefs as well, but cannot be simply equated with having beliefs or making assertions about what there is’ (Reference van Fraassen2002: 48).

The idea is helpfully elaborated by Anjan Chakravartty, who says that

So, for instance, ‘physicalism is not so much a factual thesis, but a deference to the claims of basic science’ (175). To adopt a stance is to resolve or commit oneself to acting or making decisions as described by it. Stances are open-ended, in terms of how they are interpreted and applied; their application requires discretion and judgement. They express and implement values, much as the policy of not lying implements a positive valuation of the truth (Teller Reference Teller2004: 166). Let me therefore say that a stance is a policy adopted towards the employment of epistemic principles. Epistemic pluralism is a commitment to pluralism about epistemic stances, not to epistemic systems in the Boghossian sense.

Boghossian's argument against pluralism about epistemic systems was that ‘if one of them is deemed to say something false, the other will have been deemed to say something true. Under those circumstances, it's hard to see how it could be right to say that there are no facts by virtue of which one epistemic system could be more correct than any other’ (Reference Boghossian2006: 91). This argument does not apply to epistemic stances, for it is possible for there to be pairs of genuinely alternative epistemic stances and no facts by virtue of which one is more correct than the other. We can see this most clearly if we remember that stances are action-guiding policies governing the application of epistemic principles. One can analogously think of a route as a guide to performing the action of reaching the summit of a mountain: there can be different routes up the mountain, perhaps with different benefits and drawbacks, but equally good for reaching the top. Here it is absurd to say that deeming one of the approaches ‘true’ necessitates deeming the other ‘false’, both because truth and falsity are not the norms according to which plans for action are evaluated, and because whatever that norm is, both approaches may satisfy it equally well. To give another example: it is often the case that a given mathematical theorem can be proved in two different ways, adopting in each case a different proof strategy, yet both equally ‘correct’, that is, sound as a proof of the theorem in question.

A fundamental distinction emphasized by the Jainas—and this is their second great theoretical achievement—is a distinction between inclusive and dogmatic ways of adopting a stance. A stance is adopted inclusively if its adoption does not prohibit others’ use of different stances; a stance is adopted dogmatically if its adoption does prohibit others’ use of different stances. Someone assumes a stance dogmatically if they believe that the stance is susceptible to rational support of the kind that makes it uniquely defensible as opposed to its contraries (see Teller Reference Teller2004: 162). It was, claim the Jainas, the great failing of the traditional proponents of the Sanskrit philosophical systems that they invariably took a dogmatic approach to the epistemic stances they articulate. The importance of the distinction is now clear: if we restrict our attention to stances adopted dogmatically, then a version of Boghossian's argument against pluralism about epistemic systems returns. For to adopt an epistemic stance dogmatically is to deem that its prescriptions for the use of the epistemic principles is correct and that any other prescription is faulty.

Siddharṣigaṇi, indeed, argues, in an exact parallel to Boghossian's argument against pluralism about epistemic principles, that the dogmatist claims that whenever a piece of evidence justifies a belief that an object has some one sort of property (e.g., stability), it also justifies a belief that it does not have other, contrary, properties (e.g., transformation). Epistemic stances ought not to commit themselves to the second conjunct of this claim but should remain silent on how things go with properties other than the one under investigation, and to that extent they are always incomplete ways of interrogating reality. Epistemic principles, on the other hand, are complete or absolute in their claims about justification, but they achieve completeness because there is a suppressed quantification over hidden parameters, which can be made explicit by attaching the operator ‘in a certain sense; somehow’ (syāt). This is how he explains the Jaina thesis that ‘this cognitive approach to a particular cognoscible may consist in standpoints (naya) and principles (pramāṇa). Among these two, a principle should be known to grasp completely, whereas a standpoint should be known to grasp incompletely’ (NAV 29.28, in Balcerowicz Reference Balcerowicz2001: 124).

Boghossian does consider a related idea, which is that epistemic systems, as he has defined them, are sets of imperatives. He rejects that idea on the grounds that an epistemic system encodes a particular conception of epistemic justification, but a set of imperatives does not, as well as on the ground that it cannot make sense of the relativist's relativization of justification to systems (Reference Boghossian2006: 91–93). But neither argument succeeds against Stance Pluralism. For, first, I have agreed that there is a unified epistemic system comprising a set of epistemic principles (pramāṇa), and I have defined an epistemic stance as a policy governing the use of that epistemic system; thus a stance does not itself encode a conception of epistemic justification. And second, I have separated the claim about relativization and the claim about pluralism, which Boghossian merges in his conjunctive definition of relativism. Having separated these distinct claims, it is evident that epistemic pluralism about stances is compatible with a rejection of relativization. Indeed, if it were true that epistemic principles are imperatives then stance pluralism would be impossible, for there cannot be distinct alternative approaches to the dictates of an imperative.

The Jaina distinction between principles and stances is enough to diffuse Boghossian's argument against epistemic pluralism. I have said that distinct stances may sometimes apply with equal correctness to an investigation. The Jainas argue that this does not entail that the distinct deliverances of stances are necessarily contradictory. Rejecting the idea that things have a single unique essence, the Jainas instead say—and this is their third theoretical innovation—that reality is in some sense manifold or multifaceted: the Jaina term is anekānta. Metaphysics tends to treat objects, qua targets of inquiry, as if they are simple points, like the peak of a mountain. Yet the mountain itself is metaphysically more complex, its variously shaped sides offering different aspects to the climber and so different potential routes to the top. So, to quote Siddhasena, ‘the real thing, whose essence is multifaceted (anekānta), [forms] the domain of all acts of awareness; an object qualified by one facet (ekadeśa) is known as the province of the standpoint (naya)’ (NA 29, in Balcerowicz Reference Balcerowicz2001: 83, modified). Siddharṣigaṇi elaborates, adding that ‘the real thing, both external and internal, endowed with a form that is under the sway of multiplex essential natures not separate from each other, unfolds itself to all epistemic principles (pramāṇa)’ (NAV 29.1, in Balcerowicz Reference Balcerowicz2001: 84). Mountaineers, whichever route they select, have the same tools and techniques available to them, but the mountain unfolds itself differently to each, and each aspect thus presented has as much of a claim to be the essence of the mountain as any other. Likewise, each nondogmatic epistemic stance is an approach to some one aspect of the world. Different stances are policies for warranting beliefs about different aspects of the world. That is, we might think that there are a plurality of special sciences, each special science having as its provenance some particular domain of properties, no such domain being reducible to any other (this is why it is better to speak of domains of properties, and not of levels). To say this is to deny that there is a single way the world is in itself, that there is some uniquely objective description of the world viewed sub species aeternitatis, from nowhere; rather, the multiplicity of different approaches collectively constitute a ‘view from everywhere’.

I have adopted this suggestion in my work on Sanskrit philosophical accounts of the nature of the self (Ganeri Reference Ganeri2012; compare NA 31; TS 5.29). The stance taken by the Naiyāyikas is that mental states must be owned by a subject, where ownership consists in endorsing the belief, preference, commitment, or aspiration represented in the state. That is, the aspiration, commitment, preference, or belief is made one's own. This stance is thus a normative approach to what is at stake in undertaking a first-person position. A quite different stance towards the human subject is taken by the Ābhidharmika Buddhists, whose stance is instead phenomenological. The Buddhist investigation of the first-person position is conducted with an approach that emphasizes the analytical dissection of the phenomenological character of presented mental states, most especially states of attention and attachment. Are these two philosophical approaches in contradiction with each other? That was certainly the story of the intellectual rivalry between Nyāya and Buddhism throughout the first millennium, the result, say the Jainas, of dogmatism on both sides. Yet we may instead take a more nuanced view about the multi-aspectual nature of human subjectivity, and what I have argued is that owning a mental state itself has several aspects, one of which is correctly explored by each of the two supposedly rival philosophical schools.

We might agree to say that Buddhists and Naiyāyikas employ different epistemic methodologies, as long as we are clear that this does not consist in the claim that they have rival conceptions of epistemic justification, only that their epistemic policies, approaches, cultures, are distinct. It is just because mental ownership, the foundation of subjectivity, is multi-aspected that there can be different approaches to it, in much the same way that it is just because objects have many properties that there can be different names with different senses for one and the same object.

To summarize: Boghossian unpacks the claim that ‘[t]here are many radically different, yet “equally valid” ways of knowing the world, with science being just one of them’ (Reference Boghossian2006: 2) as meaning that there are a plurality of equally correct epistemic systems. I have argued that we should construe the phrase ‘way of knowing’ as referring here, not to epistemic principles (pramāṇa) but to epistemic stances (naya). Boghossian's argument fails, in other words, because there is an ambiguity in the phrase ‘way of knowing’. Given that Boghossian is right that there is just one set of underived epistemic principles, something I have claimed philosophers in the Sanskrit cosmopolis would have agreed with, without agreeing what it is, it cannot be the case that epistemic cultures qua ways of knowing the world are what Boghossian calls ‘epistemic systems’.

4. Styles of Enquiry

I have been arguing for the viability of a position that combines epistemic pluralism, correctly understood as stance pluralism not system pluralism, with realism. I now turn to two recent attempts to carve out this same territory. G. E. R. Lloyd borrows the phrase ‘style of enquiry’ from Crombie and Hacking, but he uses it in a more relaxed sense than they. While Hacking ‘defined styles first in terms of their bringing new objects into existence, and secondly by their self-authenticating character’ (as quoted in Lloyd Reference Lloyd2007: 8) thereby creating relativized criteria of truth and falsehood appropriate to an investigation, Lloyd uses the term to ‘draw attention to the different possible foci of attention of enquiries within a single domain and to the corresponding differences in the manner in which the enquiry will be conducted’ (Reference Lloyd2007: 8n3). Different styles incorporate different criteria, but ones that ‘will be complementary to one another in so far as they relate to different aspects of the phenomena in question’ (Reference Lloyd2007: 8n3). Different styles of enquiry ‘do not constitute incommensurable systems of belief or paradigms … rather the notion of differing styles serves to underline the importance of evaluating individual theories, concepts, and explanations in the light of the wider complexes of assumptions that their proponents made’ (Lloyd Reference Lloyd2004: 78). So ‘we need to broaden the scope of what we may mean by a style of enquiry by reflecting on how substantive leading ideas, images, interests, and preoccupations help to create a perspective on the world’ (Reference Lloyd2004: 87). Lloyd provides a clear example in the contrast between the styles of enquiry of Aristotle and the Huainanzi. Aristotle looked for aitiai, causes or explanations, distinguishing the four types, material, formal, efficient, and final. Lloyd comments that ‘since explanation has to be of what is general, and in terms of stable forms, it follows that the transient, changing, particular falls out of the frame’ (88). The Huainanzi, on the other hand, depends on spotting the associations of things, marked with the use of the word thus (gu). For example, ‘Things within the same class mutually move each other: root and twig mutually respond to each other. Thus (gu), when the burning-mirror sees the sun, it ignites tinder and produces fire’ (88). The style of enquiry here is associative, not deductive. A leading principle is that associations between things govern the transformations they undergo. The focus is on correlations, not essences, and so on transformation rather than on stability. Lloyd concludes that differences in world-view can be associated with differences in styles of enquiry, which are ‘themselves constituted by different perspectives and different leading preoccupations, where there are undeniable influences from the side of culture, of values and of ideologies’ (Reference Lloyd2004: 91). Yet he wants to distance himself from the relativist claim that that there are different worlds in play. He wants to ‘uphold the claim that, despite the differences in their world-views, there is still a sense in which Aristotle and the writers of the Huainanzi inhabit one and the same world’ (91). He does this by claiming that the data are, in his particular use of the term, multidimensional:

Again, ‘the notion of the multidimensionality of the phenomena allows for different accounts to be given of different aspects or dimensions of a single domain of investiganda … [and] reflect objective features of the investiganda that may be the focus of attention of different modes of enquiry’ (Reference Lloyd2007: 8n2). The multidimensionality of the data follows, he argues, from the widely accepted theory-ladenness of observation statements. As he has summarized his position most recently (Reference Lloyd2015: 5; cf. Reference Lloyd2017), ‘recognizing that reality is multidimensional allows for a plurality of accounts, each dealing with a different aspect or dimension of the subject-matter’.

All this seems to me closely to parallel what the Jaina thinkers said about the epistemic pluralism of the classical Sanskrit intellectual world. Lloyd distances his notion of a ‘style of enquiry’ from that of Crombie and Hacking, and it seems fairly evident that the move he wishes to make is one away from thinking of a style of enquiry as a relativized truth-criterion and towards conceiving it as an epistemic stance, a policy in the guidance of inquiry in which values, preoccupations, and perspectives play a part, something ‘constituted by different perspectives and different leading preoccupations, where there are undeniable influences from the side of culture, of values and of ideologies’ (Reference Lloyd2004: 91). That is, Lloyd's ‘styles of enquiry’ simply are the Jaina nayas. In his claim that the explananda are multidimensional, he echoes also the third move the Jainas make, their claim that reality, though one, is multi-aspectual, anekānta, and that different aspects of this single reality may be the focus of different styles of enquiry. It is remarkable that the model of epistemic pluralism that the Jainas develop after observing the actual epistemic multiplicity of their lived world so closely converges with the model Lloyd arrives at through an imaginary meeting of the worlds of ancient Greece and China, in the context of contemporary battles between social constructivism and scientific universalism. The lessons are ones of pluralism and realism: a single, multi-aspectual world made available through a plurality of perspectival policies of inquiry.

5. Nature ‘As It Is in Itself’?

Hubert Dreyfus and Charles Taylor, in their recent book Retrieving Realism, are also in search of the elusive middle ground. ‘There are,’ they say,

This third possibility is what they designate ‘pluralistic robust realism’. According to this view,

The second clause rejects a relativization of the truth-predicate, and with it the idea that ‘worlds’ are social constructs. But is this a pluralism about epistemic principles or a pluralism about epistemic stances? The phrase ‘way of interrogating reality’, just like ‘way of knowing the world’ is prone to ambiguity. Here is how I think they intend the phrase to be understood, deduced by reconstructing their argument for plural realism. The argument is, roughly, as follows. Different epistemic cultures have given different answers to the question, ‘What is the essence of a thing?’ ‘In the seventeenth century our culture asked’, they say, using the first-person plural to aggregate common strands of European thought, ‘about the structure of the universe as it is in itself independent of all human interpretation and eventually developed a science that claims to be approaching a view from nowhere’ (148). All the currently available evidence is that, gold, for example, is a natural kind ‘and its essential property of having atomic number of 79 explains all its other properties that can figure in causal laws’ (151). Yet,

For the ancient Egyptians, on the other hand, the essential property of gold is the sacred property of shining with divine radiance, and this may only be accessible to Egyptian religious practices. Therefore, and this is the crux of the view,

Dreyfus and Taylor conclude that

From all this it is clear that Dreyfus and Taylor do not regard ‘ways of interrogating reality’ as epistemic principles, general normative propositions that encode a conception of justification. They consider them to be embedded cultural practices that provide a mode of access to some portion of reality, practices that encode values and goals, such as, in the case of western science, the value of discovering what things are in themselves as viewed from nowhere, and, in the case of ancient Egypt, a value inherited from Egyptian religious practices. A ‘way of interrogating reality’ is a perspectival practice concerning the proper way to set about accessing reality, not a proposition about a source of justification. Dreyfus and Taylor strikingly combine this thought with the thought we have seen expressed in different ways by both the Jaina theorists of Sanskrit intellectual pluralism and by the comparativist of ancient worlds, Lloyd, the thought that reality is anekānta, many-faceted, multidimensional, multi-aspectual. They formulate this thought by saying that epistemic practices gain a perspective on reality that corresponds to one aspect of reality, and more sharply, that what a thing essentially is relative to a culture's practices. This relativization of essence to practice should not be mistaken for a relativization of truth to practice, which makes practice into a way of world-making (as in Goodman or Hacking); rather, the multifacetedness of reality is a pluralism about essences (cf. NAV 29.1, quoted above). The view is robustly realist, while denying the claim of scientific realism that there is only one way the universe is carved up, the way described from the view from nowhere. The claim is simply, and this was the very point made by the Jainas, that distinct epistemic cultures have the capacity to reveal different aspects of a single reality, to access different orders of causal power.

Taylor and Dreyfus conclude by cautioning that the unification of perspectives may be only a regulative ideal in epistemic practice:

We might eventually build up a complete picture of reality, but such a picture would not be a view from nowhere: it would better be described as a view from everywhere.

6. Modern Science Is Itself Epistemically Plural

Karin Cetina has shown, in addition, that it is anyway a mistake to think of modern science as a single way of knowing the world, for the epistemic cultures of different parts of modern science are themselves different from one another. ‘Epistemic cultures’, she says in her seminal study of knowledge societies, are ‘cultures that create and warrant knowledge, and the premier knowledge institution throughout the world is, still, science’ (Reference Cetina1999:1). ‘The notion of epistemic cultures’, she goes on to clarify, ‘refers to the different practices of creating and warranting knowledge in different domains’ (246). Cetina uses the notion to study differences in ‘epistemic procedure’ in two prominent current sciences, high energy physics and molecular biology, and she finds them to consist in differences ‘between the liminal approach to truth in physics and ‘blind’ variation in molecular biology, or the difference between physics’ way of locating data at the intersection between signs, simulations, and theory and molecular biology's experimental conception of measurement, or the difference between communitarian mechanisms in one case and individuation in the other’ (246). To the suggestion that the notion of culture, as operative in the idea of an epistemic culture, is ‘a particular take on an ensemble of practices and preferences, a take that brings out their characteristics in relation to other such ensembles’, she responds by saying that this notion, though popular among anthropologists, should be supplemented by the idea that culture also deals with the symbolic (247). Cetina powerfully makes use of the idea of an epistemic culture, so understood, to resist the idea that science disenchants the world:

Cetina, evidently, is a pluralist about epistemic cultures, but she is not a relativist or social constructivist. Her claim is not that the truths of high energy physics are true for the high energy physicist alone while the truths of molecular biology are true only for molecular biologists, that there is no univocal truth predicate but only a spectrum of relativized ‘true-fors’. Indeed she employs a ‘notion of ontology and multileveled analysis’, which, while rejecting the view that things have an immutable essence, refers instead to ‘a potentially empirical investigation into the kinds of entities, the forms of being, or the structures of existence in an area’, for by not fixing an ontology from the start ‘one can see the configurations of several ontologies side by side and investigate their relationship’ (253). The distinct knowledge-yielding procedures of high energy physics and molecular biology carve the world up according to their particular epistemic cultures, but not in ways that render comparison and commensuration impossible.

7. Classifying and Evaluating Epistemic Stances

An epistemic stance is a policy governing the employment of the epistemic principles. Other names for an epistemic stance include ‘epistemic culture’, ‘style of enquiry’, ‘way of interrogating reality’, ‘use of reason’, ‘mode of argumentation’. The traditional proponents of the Sanskrit śāstras fell into dogmatism in their attitude towards the epistemic stances they articulate. Bracket the dogmatism and what remains is a viable mode of accessing some one aspect of reality. Each of the Sanskrit ‘knowledge systems’ is thus an epistemic stance (not an epistemic system in Boghossian's use of that term—and that is why the phrase is potentially confusing), practiced in a distinctive way to produce knowledge in a distinctive domain. For example,

The emptiness of Madhyamaka Buddhism is also, as we have seen, an epistemic stance:

‘The early Jainas, in their survey of the variety of epistemic stances employed in classical Sanskrit intellectual culture, identified seven distinct types of stance and discussed their application. ‘There are [the following] standpoints: comprehensive, collective, empirical, direct, grammatical, etymological, and factual’, says Umāsvāti (TS 1.33; Tatia Reference Tatia1994). But Siddharṣigaṇi thinks rather that there are indefinitely many: ‘according to the number, however, [standpoints are] infinite, because the real thing is endowed with infinite properties and because [various] outlooks confined to [one] property of this [real thing] are standpoints. Nevertheless, ancient preceptors taught that there are seven standpoints, by means of assuming seven outlooks that collect together all [possible standpoints]’ (NAV 29.12, in Balcerowicz Reference Balcerowicz2001: 97).

Lloyd is sensitive to a corresponding pluralism within European intellectual cultures, and in his early pre-comparative work identified in broad outlines two epistemic cultures or ‘modes of argumentation’, which he termed polarity and analogy. The analogical stance questions reality by appeal to resemblances, models, images and related notions; ‘analogies apprehend or postulate similarities or connections, often suggesting inferences and extensions of the similarities apprehended’ (Reference Lloyd2015: 3). He has shown in his later comparative investigations that the analogical stance is to be found in many Chinese writers, including those of the Huainanzi. My own studies of classical India bear witness to a similar distinction. They have led me to see that there is a fundamental contrast between two styles of reasoning, that of abstract syllogistic and formal deduction and that of particularist, case-based, ‘blueprint + adaptation’ extrapolation. The latter model—whose origins in India lie as much in the ritual reasoning of the Mīmāṃsā exegetes and the jurisprudence of the Dharmaśāstra as in explorations in the science of prediction in the medical treatises and, most especially, in early Nyāya logic—developed into a general theory of ethical and normative reasoning. The basic idea is that an object is inferred to have one, unobserved, property on the grounds that it has another, observed, one. The most distinctive aspect of the schema is the fundamental importance given to the citation of an example, a single case said either to be similar or else dissimilar to the topic at hand. Suppose I want to persuade you that it is about to rain. I might reason as follows: ‘Look, it is going to rain (pakṣa: proposed thesis). For see that large black cloud (hetu: sign). Last time you saw a large black cloud like that one (dṛṣṭānta: exemplary case), what happened? Well, it is the same now (upanaya: application). It is definitely going to rain (nigamana: decision)’. What does it tell us about the nature of reason when particulars are in this way made to work as exemplary cases? First, that methods of selection and adaptation are implied by the description of the particular. Like a flexible ruler, an exemplary case bends the standard of comparison to fit itself. Second, that the standards are context-sensitive and localized, because of the requirement that proper purpose is preserved, whether that be performing rituals that have their intended effects or making accurate medical or meteorological predictions. This implies that substitutions and comparisons remain close to the prototype, that the spread of the standards of selection and adaptation is localized. In the ‘blueprint + adaptation’ model, the standard is not absolute and universal but localized, sensitive to context, and open textured.

Thus Lloyd's ‘analogical stance’ is the epistemic stance of several strands within ancient India, most explicitly evident in the Ritual sūtras and in the Nyāya-sūtra, and I have used the terms case-based reasoning or case-based use of reason to describe it (Ganeri Reference Ganeri2014). The analogical stance is the epistemic stance according to which

This is not an epistemic principle derived from Induction and Inference to the Best Explanation, but a distinctive epistemic policy regarding their use. By polarities Lloyd meant ‘modes of reasoning that focus on pairs of opposites and use those oppositions as the basis of schemas of argumentation, as when two opposites are held to present mutually exclusive and exhaustive alternatives, and one proceeds from the rejection of one to the confirmation of the other’ (Reference Lloyd2015: 3). The polarizing stance, in which the epistemic principle Deduction and in particular the law of the excluded middle are prominent, is evident in the work of the great Buddhist logician Dignāga, whose ‘wheel of reasons’ encodes just such a view about argumentation. Dignāga's shift from the analogical stance to the polarizing stance has been heralded as marking a defining moment of transformation in the history of Indian philosophy (see Ganeri Reference Ganeri2017: 6).

Another epistemic stance, widely employed in India, appeals to visual thinking in mathematics, the use of external visual representations, such as diagrams, graphs, or symbol arrays, whose ‘epistemic roles include contributions to evidence, proof, discovery, understanding and grasp of concepts’ (Giaquinto Reference Giaquinto and Zalta2015). The same stance is operative in hermeneutics, when compactly formulated texts are treated as putting on display a certain array of concepts, rather than as presenting explicit narrative argument (Clooney Reference Clooney and Ganeri2017). We might describe the policy of such a stance crudely as follows:

The stance of scientism, as characterized by Dreyfus and Taylor, is dogmatic because it incorporates the belief that science and science alone explains all modes of being. Dreyfus and Taylor contrast the epistemic stances of post-seventeenth century European science and ancient Egyptian mythology. They fail to note, and this is precisely what is shown by Karin Cetina's careful contrastive study of the epistemic stances of high-energy physics and molecular biology, that what they describe is more a creature of mythology than of fact, that the idea of science as a unified quest for a view from nowhere is a piece of collective self-depiction rather than an description of actual scientific practice.

Alternative epistemic stances, such as Paradigm and Visual Thinking, have no place within the stance of Scientism, and are regarded as being incorrect. Yet they too represent modes of accessing aspects of reality. These are stances that do not use the epistemic principles with the intention of viewing the world from nowhere, for they are contextual in application and work through the extrapolation of local standards of comparison, drawing variably upon the individual cognitive capacities of specific viewers, readers, and thinkers.

Dogmatism about the actual practices and modes of production that constituted nineteenth- and early twentieth-century European science combined with a belief in the appropriateness of the use of violence to suppress other stances to constitute the epistemic stance of European colonialism:

While it is not the case that dogmatism per se entails the endorsement of violence, what distinguished European colonialism, in particular, was its belief in the justification of the use of violence against anyone who employs the epistemic principles differently. While Boghossian describes relativism as implying a ‘fear of knowledge’, the hallmark of this colonial stance is ‘fear of others’ knowledge’. Sheldon Pollock writes that ‘when colonialism made the norms of Europe the norms of India the Sanskrit intellectual formation melted like so much snow in the light of a brilliant, pitiless sun’ (Reference Pollock2001: 24). But, first, they were not the norms of Europe, because, I have argued, there is a common set of general normative epistemic principles that constitute a conception of justification, and indeed this was something agreed by the Sanskrit intellectuals as much as by the Europeans. It was not the epistemic system of Europe that was made India's by colonialism, but its epistemic stance, the policy of imposing its own provincial mode of accessing reality and actively undermining all others using violence instead of argument. Simone Weil, for example, would condemn the intrinsic absurdity evident in children in French Polynesia being made to recite ‘Our ancestors the Gauls had blond hair and blue eyes …’ while simultaneously forcibly denied the right to practice their indigenous custom, language, and tradition, forbidden even access to the libraries containing documentation relating to it (Reference Weil and Little2003: 110). It is the use of violence to enforce dogmatism with respect to an epistemic stance that lends the epistemic injustice of colonialism its distinctive hue.

Epistemic pluralism claims that there can be pairs of genuinely alternative epistemic stances and no facts by virtue of which one is more correct than this other. This is not necessarily so of every pair, however, and epistemic stances can be evaluated according to norms appropriate to them. Although stances are normatively evaluable, the appropriate norm of evaluation is not truth or falsity. Rather, a stance is ‘evaluated as being well or ill advised, conducive to certain ends, easy or difficult to administer, and in many other practical respects’ (Teller Reference Teller2004: 166). One path up the mountain may be steeper but shorter, another more scenic, another better served with teashops. So one can order stances, if only partially. There can also be bogus pseudo-stances (nayābhāsa), just as there can be bogus pseudo-principles (pramāṇābhāsa); for example, the ad hoc policy of being an empiricist on Mondays, a rationalist on Tuesdays, a Mādhyamika on Wednesdays, and so on. Moreover, in the case of policies, but not propositions, it makes perfect sense to advocate that one follows one policy (Plan A) unless it fails to deliver any guidance, or is in some other way inappropriate, in which case one follows another (Plan B), and so on. My general policy may be to tell the truth, but I may be fully aware that there are circumstances in which a different policy, lying, is the better policy to adopt. So ‘policies are generally not rigid in the sense that their recommendations may be overridden by other criteria or policies’ (Teller Reference Teller2004: 166). Thus, the partial ordering among stances is what we might call a ‘lexical’ partial ordering (i.e., ceteris paribus, do A, otherwise do B, otherwise C, etc.).

In all these ways, committing oneself to a stance is a matter for rational deliberation. That is because adopting a stance is a commitment, and commitments are undertaken on the basis of reflection about choices and values (Teller Reference Teller2004: 168). I might argue against {Emptiness} that it is self-defeating (like the approach to climbing a mountain by circumambulating its base). I might argue against {Scientism} that it rests on mythology (as if I could reach the top without having to take any path at all).

My argument against {European Colonialism} is that its dogmatism in denying the validity of alternative stances is underwritten by other aspects of the culture to which it belongs, specifically the endorsement that culture gives to the use of violence over reason to silence alternatives. One way to evaluate a stance, qua action-guiding strategy, is to consider whether its implementation requires or licenses, for instance by virtue of a larger ideology of which it is a part, the use of violence against others. It was indeed Gandhi the philosopher who invoked yet another foundational Jaina idea, the idea of ahiṃsā or non-violence, in a critique of the epistemology of colonialism. The philosophical function of the notion of ahiṃsā is, we can now see, to provide a nonaletheic norm on epistemic stances. Judged according to this norm, the epistemic stance of colonial Europe fares very poorly indeed.