Objectivity and Perspective in Empirical Knowledge

Abstract

Epistemologists generally think that genuine warrant that is available to anyone must be available to everyone who is exposed to the relevant causal inputs and is able and willing to properly exercise her rationality. The motivating idea behind this requirement is roughly that an objective view is one that is not bound to a particular perspective. In this paper I ask whether the aperspectivality of our warrants is a precondition for securing the objectivity of our claims. I draw upon a Sellarsian account of perception in order to argue that it is not; rather, inquirers can have contingent properties and perspectives that give them access to forms of rational warrant and objective knowledge that others do not have. The universal accessibility of reasons, on my account, is not a precondition for the legitimacy of any actual warrant, but rather a regulative ideal governing inquiry and communication.

1. Varieties of Objectivity

One traditional hallmark of the propriety of an epistemic practice is that it provides aperspectival warrant. That is to say, epistemologists generally think that our reasons must be available to anyone who is exposed to the relevant causal inputs and is able and willing to properly exercise her rationality. The motivating idea behind this requirement is roughly that an objective view is one that is not bound to a particular perspective or distorted by the intervening presence of the inquirer herself. The partial, political, personal, private, and anything else that fails the litmus test of democratic accessibility forms the unsavory and illegitimate underbelly of our epistemic practices. This pervasive view was given emblematic expression by classic mid-twentieth century epistemologists and philosophers of science such as Feigl, Popper and Hempel. Feigl writes:

The quest for scientific knowledge is regulated by certain standards or criteria … the most important of these regulative ideals [is] intersubjective testability… What is here involved is … the requirement that the knowledge claims of science be in principle capable of test on the part of any person properly equipped with intelligence and the technical devices of observation and experimentation.

(1953, 11)²

Here democratic accessibility of warrant is the privileged mark of legitimate empirical knowledge.

Historian and philosopher of science Lorraine Daston has undertaken a detailed study of the conceptual history of objectivity, in which she identifies several strands of objectivity with distinguishable but interwoven histories. Among these, facts or objects have ontological objectivity to the extent that they are real and independent of their appearance to us. Derivatively, a claim or judgment has ontological objectivity if it asserts such ontologically objective facts.³ Aperspectival objectivity, in contrast, attaches in the first instance neither to facts nor judgments, but to warrant. A claim has aperspectival objectivity to the extent that its warrant is independent of the contingencies of the claimant's personal character and context.⁴ Aperspectival warrant is what is left over when the contingent self is forcibly exorcised from the epistemic scene.

Now it seems that some ideal of ontological objectivity – perhaps a quite nuanced one – must at least implicitly govern inquiry. For if the deliverances of an epistemic practice are not held to the tribunal of an independent objective world that the practice purports to capture, then there is no reason to count it as an epistemic practice at all, as opposed to an elaborate dance or a mere social ritual. As John Haugeland has shown in detail, epistemic practices, in order to count as epistemic, are necessarily bound by two sets of norms: the norms of justification and the norms of truth, or fidelity to the objects of inquiry. What makes the former norms epistemic in the first place is that they are held to the tribunal of the second. But this will be so only if our doxastic judgments are open to correction and confirmation from the independent world they seek to capture. The legitimacy of our justificatory practices constitutively depends upon the reliable ontological objectivity of their deliverances.⁵

The central question of this paper is the following: Is the aperspectival objectivity of our warrants a precondition for securing the ontological objectivity of our claims? On the one hand, the burden of proof seems to lie with the aperspectivalists – for why should we believe that the democratic accessibility of a claim's warrant makes that claim any more likely to get at the truth about independent reality? On the other hand, being precise about just what kind of warrant might count as non-trivially contingent and perspectival will turn out to be a complex task. For instance, we do not have interestingly different perspectives just because we have access to different input data (because of differences in our sensory acuity or travel itineraries, for instance), nor because some of us are better than others at actually the drawing inferences to which they are entitled. One of my main tasks here will be to try to clarify what would count as a genuine challenge to aperspectivalism.

2. Standpoint, Perspectives, Locations: Some Distinctions and Place Setting

Standpoint epistemologists are a loosely and contestably unified group of theorists who have all defended some version of perspectivalism, arguing that some inquirers have contingent properties that give them access to kinds of knowledge that are not available to others.⁶ Most standpoint theorists have insisted upon two further claims: (1) that some contingent features of knowers can give them not only different but better, more objective knowledge than others have, and (2) that social positions of marginalization and structural disadvantage, such as those inhabited by women, African-Americans, or the working class, yield epistemological advantages, giving those who occupy them the potential to see truths that are inaccessible from the point of view of the dominant center. In this paper, as in my two earlier papers on this general topic (Kukla and Ruetsche 2002, Kukla 2003), I am interested in making sense of and defending (1). I in fact believe in (2) as well, and I have defended it in the past (Kukla 2003); however, for purposes of this paper, I can remain agnostic with respect to (2). My concern is with revealing the conditions for the possibility of some social positions yielding better, more objective perceptual possibilities than others, by attacking the presumed link between ontological and aperspectival objectivity.⁷ This leaves open the question of which social positions do in fact yield such an epistemic advantage.

There has been vigorous debate among those sympathetic with some version of standpoint theory over just what a standpoint actually is. Classic formulations of standpoint theory looked to Marx's account of class-consciousness, and described standpoints as belonging to groups in virtue of their structural position within systemically differentiated societies. In response, critics such as Susan Hekman (1997) accused such theories of 'essentializing' these groups, and glossing over the differences in standpoint between various group members. A particularly vindictive set of responses to Hekman accused her in turn of depoliticizing standpoint theory, removing its liberatory and strategic potential, and replacing structural social difference with a facile individualism.⁸

I take myself here to be asking a question that is prior to these disagreements, namely, whether it is coherent to claim that any social location – found or forged, idiosyncratic or shared – yields genuine warrant that is not universally accessible. Alison Wylie follows Hekman's critics in objecting to "a recurrent tendency to reduce the notion of a standpoint to the social location of individuals, a move that is inevitable … if it is incomprehensible (to critics) that social structures, institutions, or systemically structured roles and relations could be robust enough to shape what epistemic agents can know" (2003, 29). I by no means find it incomprehensible that such structural forces could shape what epistemic agents can know.⁹ At the same time, I believe that epistemic agents are individuals, regardless of the extent to which their epistemic agency is marked by their group memberships, decentered, achieved in collaboration with others, and so forth. Hence I think that it makes sense to ask about how the location of such individuals can enhance or compromise this epistemic agency. I will use the term 'perspectives' rather than 'standpoints' in order to focus my attention on individual knowers, while opting out of the debate over the best understanding of standpoints. But I do not think that doing so commits me to rejecting or underselling the extent to which such locations are systematically constituted by race, gender, and other institutionalized social identities.

This paper will have no direct political implications, in the sense that it does not seek to advocate for the epistemic value of the standpoint of any particular group, or of marginalized groups in general. However, I believe that it has two indirect political implications. First, it will create a theoretical opening for those who do wish to advocate for the strengths of a particular standpoint, or to strategize about how to forge one, by providing grounds for believing that such a project is coherent, and a rebuttal to those who have equated standpoint epistemology and the rejection of aperspectivalism with giving up on the ideal of objectivity altogether.¹⁰ Second, it will reveal that aperspectivalism itself is not a 'politically neutral' option in epistemology; rather, it is a piece of ideology, fueled by a history of specific interests and usually accepted without any argument or critical interrogation.

3. Rationality and Second Nature

One traditional version of aperspectivalism is grounded in a particularly restrictive conception of rationality, according to which we can distinguish between the contexts of discovery and justification, and locate our rational epistemic achievements in the latter. The definitive task of epistemology, on this view, is to capture these achievements with a confirmation theory, which is supposed to encode all and only proper exercises of rational justification. Anything that counts as a proper exercise of reason for anyone counts as a proper exercise of reason for everyone, and the criteria for what counts as a proper exercise of reason are decidable in advance, ready to be diligently applied to whatever evidence happens to show up within the context of discovery.

Now Aristotle, as we know, argued that our capacity for moral perception – that is, the capacity to see moral reasons for action in the situations we encounter – is a contingently inculcated second nature, cultivated through history and education and unevenly distributed, even among those whose organs are equally capable of processing sensory inputs. Moral perception, on this picture, is not only a contingently inculcated capacity, but it is one whose development and exercise take work. Our capacity for moral perception will vary widely depending on our personal history of experiences and practice, and, as John McDowell emphasizes, this capacity, though receptive, is never merely passive. Moral perception is a matter of skilled discernment and it is often hard, and the variation between individuals' abilities and dedication is large. As McDowell puts it, it can be "a manifestation of reason … recognizable as such only from within the practice" (1979, 345).

But it is unclear why we ought to restrict the Aristotelian account of perception as a contingently inculcated virtue to the domain of morality. In his epistemological writings about empirical knowledge broadly construed,¹¹ McDowell describes perception in general as an inculcated capacity to see reasons. Oddly, however, though his accounts of the metaphysics of moral and natural perception seem to be the same, McDowell never talks about the work that might be required in the case of natural perception, nor about how different inquirers might differ in their capacities to perceive non-moral reasons. Quite in contrast to his rhetoric and emphasis on the moral domain, McDowell's examples of perception in texts such as Mind and World are highly passive and automatic – he speaks only of seeing everyday colours and objects, and he refers to what "we" see with "our" second natures, rather than what individuals might be able to see thanks to their work and virtue. And yet, outside of the moral domain, our history of interests and experiences clearly gives us abilities and inclinations to notice some facts and distinctions and not others. Musicians notice features of musical performances that I cannot. Reading X-rays, ultrasounds and MRI results are skills that I don't have. I am more likely to notice sexist and classist behavior than are most of my male friends from more privileged backgrounds. Changes in our own bodies, or even in the bodies of loved ones, will change our perceptual dispositions. Two summers ago I helped to care for my father, who has Parkinson's disease. A common symptom of Parkinson's is the inability to move into a space unless that space is much bigger than what the body actually will use. As the summer progressed, I came to see the geographies of rooms and passages quite differently; manageable and unmanageable spaces for my father simply presented themselves to me as salient parts of the landscape.¹²

Our interests and our contingent histories shape and develop such dispositions. On the one hand, the dispositions we develop might bias or distort our inquiries: someone's investment in her research program may hamper her ability to recognize evidence that tells against the viability of that program. Another might interpret all of his intimate relationships through a lens distorted by an abusive relationship in his past. On the other hand, our interest-driven second nature dispositions might enhance our rational inquiries by enabling us to see salient patterns of evidence. Margaret Little, for instance, points out that observational vigilance is not always sufficient to enable us to find relevant evidence. This is because:

There is no exhausting ahead of time what one should be on the watch for. There are indefinitely many things that may be … relevant in a situation … the [properly attentive] person, then, is not someone who approaches each situation with some conscious grocery list of things to check for. The required attentiveness is a background disposition for relevant details to come into your consciousness, to emerge as salient.¹³

However, the fact that our second-nature perceptual dispositions can play such an epistemically valuable role in inquiry does not necessarily challenge the idea that rationality per se consists in the proper application of domain and agent-independent principles of justification to evidence.¹⁴ The purist could still distinguish between our second nature perceptual dispositions to attend to or ignore certain kinds of evidence, on the one hand, and our rational response to the evidence, on the other. She may allow that that our interest-inflected second nature perceptual capacities influence what evidence we notice, but deny that their exercise is itself part of rationality.¹⁵

Aristotle and his contemporary heirs, on the other hand, claim that perceptual capacities are second nature, contingently inculcated rational capacities. Virtues, for Aristotle and neo-Aristotelians such as John McDowell, Martha Nussbaum, and Nancy Sherman, are not mere habits of feeling or action. Rather, they are receptive sensitivities to reasons, born out of proper upbringing, experience, and practice. If this is correct, then since perceptual sensitivities operate in the context of discovery, they thereby provide warrant that could not be captured in a theory of the appropriate relationship between evidence and belief. Indeed, such rational perceptual capacities would confute the boundary between the contexts of discovery and justification, and thereby provide a serious challenge to the restricted account of rationality with which we began.

Nonetheless, even granting all this would not disprove aperspectivalism. The aperspectivalist could still argue that in order for someone to have any properly rational perceptual capacities she must have the same ones that everyone has. Certainly, this line goes, different people will have physical sense organs capable of processing different inputs, and different opportunities to be exposed to information, but given the same exposures and the same inputs, everyone who is genuinely rational will perceive the same things, and their perceptions will entitle them to same warrants (whether or not they are insightful enough to notice these warrants). Our perceptual capacities are inculcated over the course of our contingent histories, but to the extent that they succeed in being rational capacities these contingent histories all end up at the same place. This view would amount to a sophisticated version of aperspectivalism – it allows for an enriched conception of our rational capacities, upon which rationality is contingently inculcated and operates within the context of discovery, but insists that genuine receptivity to reasons manifests itself as the same capacity in everyone who has it. This position has some vocal supporters in the domain of moral perception, including Iris Murdoch and Martha Nussbaum.¹⁶ I think that one could argue that this is Donald Davidson's view, insofar as he argues for the incoherence of multiple conceptual schemes, and in turn Davidson's view arguably has its roots in Kant's argument in the first Critique for the a priori universality and necessity of the table of judgments (Davidson 1973-4).

Contingent differences between our perceptual capacities, I have argued, affect what we are able to notice. However, we've seen that this isn't enough, on its own, to provide a challenge to the belief that aperspectival objectivity is a condition for the legitimacy of reasons. In order to argue against aperspectivalism, then, we must be able to defend the subtle claim that the epistemic status of something as warrant can depend on the standpoint of the inquirer. Our contingent histories and resulting second natures must have the potential to make us not only more or less rational, but able to perceive different reasons and access different warrants when being rational in response to the same causal inputs.

4. Sellarsian Perception and Contingent Warrant

A neo-Aristotelian account of perception will pose a genuine challenge to aperspectivalism only if perceptual capacities are contingently inculcated yet genuinely rational capacities, whose deliverances are warranted only for certain kinds of perceivers. Until now, I have not given any substantive story about the nature of perception, and likewise my treatment of perception as a contingently inculcated rational capacity has been merely hypothetical. I am certainly not going to develop and defend a whole theory of perception here, but I will draw upon one highly influential account of perceptual knowledge that is friendly to my purposes, namely that of Wilfrid Sellars' "Empiricism and the Philosophy of Mind", and I will try to say enough about it to make it a plausible theoretical contender.¹⁷

Sellars argues that if perception is to be able to provide any warrant, its contents have to have conceptual structure sufficient to allow them to bear rational relationships to other conceptually articulated judgments. We must be able to perceive that x is F, rather than just taking in brute sense data. But, Sellars contends, our ability to perceive that some perceptual fact of the form 'x is F' holds requires that we grasp the conditions for the appropriate application of the concept F. That is, we must understand the conditions under which things that appear to be F are F, and vice versa. To use his example, I cannot see that a necktie is green unless I understand facts such as that green things look green under natural lighting, that they don't look green when seen on a black and white television, and so forth (Sellars 1997, §18). Now, grasping such conditions for property recognition involves understanding under what conditions various inferences (such as the inference, in a certain context, to x's actually being F) are or are not licensed by appearances. Without this normative and inferential mastery, we cannot distinguish between seeing that x is F and it merely looking as though x is F, in which case, according to Sellars, we could not drive the crucial wedge between appearance and reality that is necessary for our perceptual states to count as properly epistemic states. Hence for him, the ability to recognize a piece of evidence cannot be neatly separated from our ability to use it in inference, and hence perception cannot be taken as a capacity for discovery that lies outside the context of justification. In Sellars' terms, perceiving that x is F requires that our recognitional episode be placed "within the logical space of reasons, of justifying and being able to justify what one says" (1997, §36).

For Sellars, perceptual capacities are inculcated through our contingent histories, "involving a long history of acquiring piecemeal habits of response to various objects in various circumstances" (1997, §19). This is the history of our mastery of the ability to recognize instantiations of various concepts, in and through our mastery of their normative and inferential relationships. Thus, if x is indeed perceptibly F for a particular agent, this fact is dependent upon on the agent's contingent past. Only if she has the right history will she have developed the capacity to perceive that x is F, and only then will her empirical confrontation with x warrant beliefs and inferences based on the fact that x is F. An agent's particular history of observational situations and learned responses will inflect the topography of the recognitional concepts she brings to bear in perception, by giving these concepts their life and hence their content within differently inflected spaces of reasons. Our contingent history of concerns, experiences, and conditions of observation helps determine which facts and properties can show up for us and what counts as normal and aberrant behavior for objects of different sorts. Thus these contingent histories and the second natures they inculcate will help constitute what evidence is available and which inferences are warranted in the face of worldly objects and events.

But if our ability to perceive inferentially fecund facts is a contingently inculcated second nature capacity, then there is no prima facie reason to think that we share it in all of its details. We should expect the rational deliverances of perception to vary depending on the experiences and practices that gave form to an inquirer's normative grasp of standard conditions and appropriate inferences. Once again, there does not seem to be a good prima facie reason to distinguish between natural and moral perception here, for both involve learned receptive sensitivity to normative relationships and inference-licensing reasons, whether these are reasons for actions or for beliefs. If this is right, then our perceptual capacities can provide variant warrant that fails the test of aperspectival objectivity. I mentioned at the beginning that mere differences in our travel itineraries or inferential savvy, as such, do not constitute counterexamples to aperspectivalism; these are cases in which contingent differences between us impact our ability to access warrants to which we would both be entitled, were we to encounter the same worldly situation. But on the Sellarsian picture of perception that I have sketched, differences in our contingent histories of experience and inference can lead to differences in perceptual capacity such that we are entitled to different warrants even when faced with the same worldly situation and even when we are both inferring properly from what we see.

None of this commits us to any opinion as to whether our contingent access to warrant simply varies from individual to individual, or whether it is interestingly rooted in systematic, structural differences in social position of the sort that may be shared by groups, as most standpoint theorists claim. Certainly Sellars does not use his own account to support any kind of standpoint epistemology attributing different access to warrants to different kinds of inquirers. But as a matter of empirical fact, it might turn out that there are systematic differences in perceptual capacities and available warrant for different types of agents. People living in different countries, or inhabiting different class positions, for instance, might have different enough histories of conditions of observation and inference that, as Marx claimed, they end up able to see systematically different facts about the world. Indeed, social identities that pervade our experience and opportunities, such as gender and race, seem to be just the kinds of factors that would likely play a systematic role in shaping our history of experiences, and especially our inculcated sense of what counts as 'normal conditions'.

5. Ontological Objectivity and the Plasticity of Perception

At the beginning of this paper, I argued that if a practice is to count as epistemic, fealty to the facts and objects it tries to capture has to be the tribunal of its adequacy; in other words, it must be substantively governed by the ideal of ontological objectivity. At this point, I have argued for the possibility of contingent, perspectival warrant. But warrant is an epistemic notion, and as such, nothing can count as warrant unless it can be rationally corrected by the testimony of independent objects. In claiming the perspectivality of warrant, therefore, I am not claiming that truth is relative to a perspective, but rather that different perspectives can yield different forms of rational access to the independent truth – or, to return to the language with which I began, that aperspectival objectivity is not a necessary condition for ontological objectivity. Hence my attempt to make room for contingent warrant within a Sellarsian picture will only work if our second nature epistemic practices are themselves open to rational revision in the face of new evidence.

To this end, notice that the contingent histories of our rational capacities are not simply strings of chance events that result in a second nature we have by happenstance. Rather, as Aristotle emphasized, our second natures are educatable through practice and experience. It is insofar as our perceptual capacities can be cultivated that these capacities count as Aristotelian virtues. The cultivation of virtuous perception is driven by the guided and attentive exercise of our rational capacities, including our receptive sensitivities. Thus, as Nancy Sherman puts it, Aristotelian perception can be "cultivated to yield dispositions that are enduring and responsive to appropriate objects" (1989, 47). Accordingly, our perceptual capacities and access to warrants will vary and (hopefully, though not necessarily) evolve over time and with experience.

The trick here is to notice that our contingent experiences can change not just our judgments, but our second nature epistemic capacities themselves. The exercise of our rational capacities in perception, judgment and inference modifies these capacities, and such modifications are themselves generally rational modifications rather than mere mutations. Consider a case where I discover that appearances are deceiving, because conditions are non-standard in a way I had not encountered before – perhaps I always thought that I could recognize expressions of happiness, but had not realized that happiness expresses itself differently in Norwegians. Such a discovery can do more than change my judgment as to whether this Norwegian is happy – it can also change my grasp of the normative conditions that govern evidence for happiness, the inferential link between such evidence and my other judgments, and, over time, my dispositional habits of perception. Indeed, Sellars' parable of John the Necktie salesman is just such a story: John is taught to perceive the colour green by his fellow salesmen, who help direct his attention to how conditions of observation affect the appearance of colours (1997, §16). Earlier I pointed out that our rationality can be compromised or enhanced by our second nature perceptual dispositions to notice, ignore, distort, etc. But on the picture I have now sketched, these dispositions determine not only what we will notice, but even what we are capable of perceiving. This means that in order to properly cultivate such dispositions, we have to alter not only our habits of attention, but our capacity to observe – we must, as Sartre put it, "change the structure of our eyes".

Now of course, not all second natures are educatable into all forms of rational capacity. There are all sorts of ways in which someone's second nature might be recalcitrantly uneducatable. Insouciance, inattention, impairment, and biases and prejudices that block or distort certain kinds of information all might hamper the cultivation of rationality, just as they can the moral virtues. But this is no argument against the rational modifiability of our contingently variable rational capacities. For not all dimensions of our second nature are rational. We might have second nature dispositions to irrationally distrust or ignore certain voices or kinds of evidence, to look to our left before we look to our right, to uncritically latch onto evidence that confirms our worldview, to commit various probabilistic fallacies, and so on. But the fact that we are irrational in various ways is neither here nor there when it comes to the question of whether we possess contingent rational capacities whose proper exercise can lead to their own improvement.

Indeed, it now looks as though a key element of epistemic responsibility is the ongoing cultivation of our perceptual capacities; being a responsible observer requires not just using our perceptual capacities to look carefully and fairly at the evidence, but working to develop these capacities so as to render our perceptual apparatus more sensitive and accurate. We need to take it as an epistemic failure and not just a piece of bad luck if we are unable to access warrant that others seem able to access. Likewise, occupying a perspective that enables accurate, objective perception is an achievement and not just a given feature of our found location.¹⁸ It is not fully under our control what we can learn to see, but there is also no reason why we must leave the direction of this development entirely up to chance.

Yet taking responsibility for rectifying deficits in our perceptual capacities is no trivial task. Mere observational vigilance will not do the trick, since we are seeking to fix the very perceptual capacities we exercise during such vigilance. We can work ourselves into a quasi-Meno's paradox if we focus on the worry that we can't learn how to perceive something if we can't already perceive it to learn about it in the first place. Luckily, the paradox is only apparent.

First, though perception is always receptive, this doesn't mean that it is always immediate. We might have the capacity to see something, but only with work. We may need to draw on our current epistemic resources in novel and creative ways in order to perceive, and in doing so we may incrementally rehabituate our perceptual practices. When we learn to look at maps or geometric or scientific diagrams, for instance, we bring reason and inference to bear in decoding what we see, and only later, once we have habituated ourselves to looking at them, will we be able to directly perceive the information they contain. The same may go for the emotions of Norwegians. We will be suspicious of this idea that we can move from inferentially mediated to direct perception only if we beg the question, and assume that our receptive capacities are not themselves rational second natures.

Second, we can recognize the limitations of our perceptual capacities and draw upon other resources in order to hone these capacities. Nancy Sherman points out that for Aristotle, sometimes "what is most crucial [to the refinement of our perceptual capacities] is a recognition of the limits of one's experience and an acknowledgment that one may need external assistance" (1989, 53). In order to accept and benefit from this assistance, though, we first need enough humility to acknowledge that perception is a contingently variable capacity, and that in some domains, others are well equipped to educate us if we let them. This sometimes requires us to make a provisional meta-commitment to trusting in someone else's perceptual capacities, for reasons other than our ability to directly verify their deliverances right now (for instance, Sellars' John the necktie salesman had to have such a commitment to trusting his colleagues' colour vision in order for him to develop his perceptual capacities under their guidance). Luckily, such trust, while fallible, need not be arational or arbitrary. Indeed, we all live within an epistemic division of labour whose structure is intertwined with the larger structure of social institutions. We look to various kinds of experts to help us form beliefs, set goals, develop preferences, and make choices. Our trustful reliance on such a fabric of expertise is so massive that if it seriously faltered we would have a social and epistemic crisis on our hands. One of the tasks that it is often rational to entrust to others is that of helping us to educate our own perceptual capacities. Reasons for such trust can be found not only in our institutionalized mechanisms for recognizing expertise – we trust music critics to teach us how to listen to music and doctors to teach us how to look at diagnostic images – but also in our already-inculcated capacity to see when others have a contingent history of experiences that is likely to make them sensitive to some things that we are not.¹⁹

Traditional epistemology presupposes that aperspectival objectivity is a condition for ontological objectivity. It is because I have separated aperspectival from ontological objectivity, rather than giving up on the latter, that I can coherently claim that some perspectives are more objective than others. But many critics of traditional epistemology and the aperspectival ideal retain the assumption that aperspectivality and ontological objectivity stand or fall together. Hence, in giving up on the aperspectival ideal, they also give up on the idea that we can rank the epistemic quality of different perspectives, arguing that such ranking would require a 'view from nowhere' from which comparisons between perspectives could be made. According to Susan Hekman, for instance, it is a "logical consequence" of embracing situated, perspectival knowledge that "no perspective/standpoint is epistemologically privileged" (1997, 351).

In our earlier paper, Laura Ruetsche and I argued against such an inference from the absence of a 'view from nowhere' to the impossibility of ranking the objectivity of perspectives (Kukla and Ruetsche 2002). I will not rehash these arguments here, but I want to consider briefly what implicit picture of rationality might have made this inference tempting. Hekman writes, "All of Harding's talk of 'less false stories', 'less partial and perverse accounts', and more 'objective' research necessarily presupposes a shared discourse – a metanarrative, even – that establishes standards by which these judgments can be validated" (1997, 355). Hekman rejects the possibility of such metanarratives. But in assuming that the only way to judge some claims and practices to be more 'objective' than others is by using a 'metanarrative' that establishes universal, domain- and content-independent standards for judgment, Hekman ironically recreates traditional, positivist assumptions about the nature of rationality: She presupposes that the context of justification is fully separable from the context of discovery, and that any legitimate practices of rational judgment can be codified in the form of a set of discursive standards.

However, if the neo-Aristotelian picture of rationality and perception that I have been drawing upon is correct, then it challenges precisely these assumptions. A genuinely rational assessment of the comparative objectivity of different perspectives need not proceed by appeal to independent discursive principles. Our comparative judgments may be rational whether or not we can separate out a set of principles for making these judgments that are fully independent of the judgments themselves. By drawing upon the techniques and capacities that I have discussed – direct perception, the cultivation of perception through education and practice, rational trust in the judgments of others – we can learn to see shortcomings and strengths in others' epistemic practices, as well as in our own.

6. Epistemological Democracy

I have argued that ontological objectivity is a non-optional epistemic ideal. We strive to perceive and reason properly in order that we might discover the facts as they really are as opposed to how they subjectively appear. The aperspectivalist sets up a separate methodological ideal: warrant only counts as legitimate to the extent that it is democratically accessible. Conversely, the influence of the contingencies of the self on inquiry can only be either distorting, or external to rationality itself. Now it seems to me that we have good reason to accept this second methodological ideal if and only if following it essentially supports the first ideal. I have argued that on one rich picture of perception, with its roots in Aristotle and its contemporary development in Sellars, McDowell, and others, perception turns out to be an exercise of rationality that upholds the ideal of ontological objectivity but violates the ideal of aperspectival objectivity, providing contingent warrant that is available only from certain standpoints. I haven't given arguments that would compel us to accept this picture of perception. But as long as it is a serious contender, the burden of proof shifts to the aperspectivalist, who must show why only aperspectival, democratically accessible warrant can be counted upon to successfully yield rational, ontologically objective claims. Until such proof is forthcoming, we must treat the valuation of aperspectival warrant as a piece of ideology.

Lorraine Daston argues that the ideal of aperspectival objectivity has a specific social history:

Aperspectival objectivity became a scientific value when science came to consist in large part of communications that crossed boundaries of nationality, training and skill. Indeed, the essence of aperspectival objectivity is communicability, narrowing the range of genuine knowledge to coincide with public knowledge (1992, 600).

Historically, she claims, the association of objectivity with public accessibility, repeatability and communicability does not have its primary roots in epistemological or metaphysical pressures, but rather in what we might think of as epistemic etiquette. Aperspectival objectivity became an ethos for scientists for practical and political reasons, as the increasingly expanding and mediated community of scientists, sharing their results through journals and conventions rather than personal communications, needed to be "calibrated and organized". Public, aperspectival warrant served as a guarantee against the stains of corruption, bad character, and unregulated idiosyncrasy, as well as functioning as a democratic and anti-elitist epistemic measure.

In light of this history, we might well worry that in scrapping the ideal of aperspectival objectivity, we risk giving up something important. There are political and practical dangers involved in being ready to accept that some people have direct perceptual access to facts and inferential access to warrants that others are not equipped to access. We risk licensing exclusionary and elitist practices of judgment, taking some people as especially authoritative inquirers without requiring them to justify their judgments by appeal to publicly accessible facts and reasons. We can imagine whole groups of people at risk of disenfranchisement from the community of fully legitimate inquirers. Once we relinquish the goal of expunging the contingent, situated, perspectival self from epistemic practice, can we ensure that objective judgments will be public property in the right way?²⁰

The solution, once again, lies in the plasticity of our second natures. We would have the right to exclude inferior perceivers from the community to whom we are epistemically accountable only if we could be sure that their perceptual capacities were not educatable. Likewise, we could dismiss the epistemic claims of viewers who claim to perceive things that we cannot only if we could be sure that we could not learn to perceive what they perceive. But as it is, the differences in perspective and perceptual capacity that give us access to different warrants and reasons are never fixed in stone. Given the possibility of cultivating more sensitive, more inclusive perceptual capacities – capacities that allow us to be claimed by more of the reasons that the ontologically objective world can warrant – both ethical and epistemic responsibility demand that we attempt to bridge perceptual divergence through such cultivation, rather than writing off the other as simply mistaken or as trapped in a perspective incommensurable with our own. If my picture of perception is right, then perspectives can indeed license different warrants, but there is no reason to think that such epistemic inaccessibility is necessarily permanent.

Our second nature capacities are internally limited and partial even to the extent that they are rational. We can imagine, however, that these internal limitations might be self-overcoming given the proper experience, work and education, as long as we exercised our rational capacities correctly on their own terms. That is, it might be the case that were an individual to be given the right experiences and training, and were she at each stage to respond fully rationally given her current capacities to this experience and training, then she could, through the exercise of her rationality, retool her second nature so as to come to be able to grasp any proper reasons that anyone can grasp. In this case, all rational agents would in principle have the capacity to be educated into a maximally inclusive rationality, given the disciplined use of the rationality they already have. We can call this the thesis of the 'optimistic plasticity' of our rational capacities. On this view, to the extent that we are uneducatable, it is because of irrationalities built into our second natures and not because of the incompleteness and partiality of our rationality.

Ideally educated, optimistically plastic subjects would inhabit a maximally inclusive perspective, which is not that of an abstract 'ideal observer', but rather that of an observer whose capacities are ideal given what actual humans with contingent, sensuous, receptive faculties can come to perceive. This notion of an ideally educated observer does not provide us with a measuring stick for judging the adequacy of particular perceptual capacities, but rather a theoretical picture of what fully objective perception would be, when we begin from our finite, sensuous nature rather than from a divine standard of transparent vision. I think that the deliverances that would be available from such a standpoint can be read rather nicely as what Kant called the deliverances of common sense, which he thought aesthetic education helps us approach, and whose ideally universal accessibility he took to be a regulative presumption underlying communication and objective perception.

Now if everyone had the perspective of the ideally educated, optimistically plastic inquirer, then all warrant would in fact be democratically accessible, or part of 'common sense'. The fact that it is not so now need not undermine the legitimacy of our warrant, as the aperspectivalist thought it would. But if this regulative picture of ideal human objectivity is correct, then the attempt to cultivate one's faculties to allow a maximally objective view is equivalent to the attempt to have access to warrants that would be democratically accessible in an ideal human epistemic community. Where the aperspectivalist advocated restricting the tools of epistemic inquiry to those that are already shared, I am suggesting instead that democratic accessibility is a project that we undertake in the shared pursuit of knowledge. The ideal exercise of the rational capacities that any of us has now, from an epistemic as well as a political point of view, is one that not only disciplines the use of these capacities to prevent distorted uses of them, but also, in good Aristotelian fashion, seeks to cultivate and educate these same capacities in the direction of maximal inclusiveness.

There is no reason to think that our rational capacities are in fact optimistically plastic, nor that the perspective of the ideally educated perceiver is an attainable goal. Indeed, even if it is attainable in principle, we will generally fall far short of it in practice. People often cannot be brought to see certain kinds of truths – certainly not on the spot during a disagreement, and perhaps never. But simply consigning oneself or others to a partial and inaccessible standpoint has to count as a failure of responsibility. Through mutual education and attention we can strive to cultivate a maximally inclusive shared perspective in which all warrants that could be available to anyone are available to everyone. Thus the aperspectival ideal of democratically accessible warrant need not and should not be scrapped. The universal accessibility of warrant is a built-in ideal of inquiry and rational discourse. But we need to transform the logical place of that ideal, from a necessary condition placed on counting any actual judgment as objective, to a regulative principle governing our rational attempts to work towards a maximally complete and accurate grasp of the character of the empirical world.

I have reinvested in the ideal of ontological objectivity, while arguing that the contingencies of the self need not be excised in its pursuit. By abstracting away from those features of the self that are contingent and perspectivally bound in inquiry, we reduce our knowledge to the lowest common denominator, and accept as legitimate only those reasons that happen to be accessible to all rational inquirers in all of their contingent variety. On the picture of rationality and second nature that I have explored, this would require cutting ourselves off from crucial resources that aid our ability to perceive and reason as objectively and accurately as we can. The universal accessibility of reasons should serve as a regulative ideal governing our shared project of cultivating our rational capacities, rather than as a condition upon the objective adequacy of any actual warrant.