Abstract
A long tradition in the philosophy of education identifies education’s most fundamental aim and ideal as that of the fostering or cultivation of rationality. In this article I relate this tradition in philosophy of education to recent work inspired by Wilfred Sellars on ‘the space of reasons’. I first offer a very brief overview of the tradition just mentioned, after which I briefly lay out Sellars’ notion and discuss its place in the work of some of those influenced by Sellars, especially John McDowell. I next address recent work in philosophy of education that suggests that there is a tension between Sellars’ notion and the traditional educational ideal as I have developed and defended it in my own work, or that the Sellarsian view as developed by McDowell resolves outstanding difficulties with my version of the traditional view. I will argue that there is less tension than there appears to some of my critics to be and that the Sellarsian notion is in fact compatible with the traditional view as thus developed, but that it leaves out an important aspect of that view that should not be lost.
Introduction
A long tradition in the philosophy of education identifies education’s most fundamental aim and ideal as that of the fostering or cultivation of rationality. In recent decades this aim has been advanced under the banner of critical thinking. Because this story is well known I won’t dwell on it in detail here. Instead, in what follows I will relate this tradition in philosophy of education to recent work, inspired by Wilfred Sellars, on ‘the space of reasons’. I will first offer a very brief overview of the tradition just mentioned. I will then briefly lay out Sellars’ notion and discuss its place in the work of some of those influenced by Sellars, especially John McDowell. I will next address recent work in philosophy of education that suggests that there is a tension between Sellars’ notion and the traditional educational ideal as I have developed and defended it in my own work, or that the Sellarsian view as developed by McDowell resolves outstanding difficulties with my version of the traditional view. I will argue that there is less tension than there appears to some of my critics to be, and that the Sellarsian view is in fact compatible with the traditional ideal as thus developed. But I will argue as well that the Sellarsian view leaves out an important aspect of the traditional ideal that should not be lost.
The place of reasons in education: A very brief overview
Philosophical issues concerning education have enjoyed substantial attention throughout the history of Western philosophy. From Socrates’ conversations with his fellow Athenians, Plato’s Republic and Aristotle’s Politics to the great philosophers of the Middle Ages, the Enlightenment and the Modern era, through to the twentieth century, issues concerning education, both epistemological and ethical/social/political, have loomed large on the philosophical scene. In this long history, there has been an unusual degree of agreement among the central figures in the philosophy of education concerning the place of reason, or the fostering of rationality, as the most fundamental aim of education. 1
These days, the story is often told in terms of critical thinking. While there are a range of theories of critical thinking in the contemporary literature and the usual amount of philosophical controversy, theorists are mainly agreed that critical thinking is best understood as involving two components: a reason assessment component involving the ability to evaluate the quality or epistemic strength of reasons, and a critical spirit component involving a range of dispositions, habits of mind and character traits such that the critical thinker is actually disposed to engage in reason assessment and to be guided by it in belief, judgment and action. On my own view of the matter the ideal is justified by several considerations, but most importantly is compelled by our moral obligation to treat students, children, and indeed everyone with respect as persons: failing to educate for critical thinking constitutes a failure to treat students with respect. On my view the successful articulation and defense of the ideal requires as well a principled rejection of epistemological relativism, a non-epistemic conception of truth, an internalist conception of justification, and a ‘transcendental’ justification of reason or rationality itself. (Bailin and Siegel, 2003; Siegel, 1988, 1997, 2003, 2007a, 2010a; cf. Robertson, 2009) These are of course highly contentious matters, and I cannot undertake a serious defense of them here. But I hope to have said enough to provide a proper background for the discussion of the ‘space of reasons’ to follow.
Sellars and ‘the logical space of reasons’
As is well known, Sellars’ overarching project – and, according to him, the project of philosophy itself – was that of reconciling the ‘manifest image’ of ourselves, as traffickers in the intentional contents of language and thought and the normative character of belief, judgment and action, with a thoroughgoingly naturalistic ‘scientific image’ of human beings as complex but scientifically explicable denizens of the natural world.
2
Of special relevance here is Sellars’ view of the ‘irreducibly normative character of epistemic discourse’ (Rosenberg, 2009: 10) and of thinking with, or in terms of, concepts: . . . anything which can properly be called conceptual thinking can occur only within a framework of conceptual thinking in terms of which it can be criticized, supported, refuted, in short, evaluated. To be able to think is to be able to measure one’s thoughts by standards of correctness, of relevance, of evidence. (Sellars, 2007: 374)
It is in the context of his celebrated critique of ‘the myth of the given’ and his positive characterization of epistemic normativity and its incorporation in his ‘stereoscopic’ reconciliation of the manifest and scientific images that Sellars introduced the notion ‘the logical space of reasons’ in Empiricism and the Philosophy of Mind: The essential point is that in characterizing an episode or a state as that of knowing, we are not giving an empirical description of that episode or state; we are placing it in the logical space of reasons, of justifying and being able to justify what one says. (Sellars, 1997: VIII, 36, p. 76, emphasis in original)
The details of Sellars’ view are complex and would require extended treatments of fundamental issues in the theory of meaning, the philosophy of language more generally, metaphysics and the philosophy of mind, and other core areas of philosophy. Such treatment would extend well beyond the present effort. Happily, for present purposes such an extended excursion is unnecessary. Our task here is to examine the relationship between Sellars’ naturalistic view of the normativity of epistemic discourse, thought and judgment as captured in his metaphor of the logical space of reasons and the traditional educational aim of the cultivation of reason.
Naturalism, normativity and transcendence
Sellars’ account of normativity is naturalistic; the scientific image is that given by the theoretical/explanatory sciences. What does Sellarsian ‘naturalism’ come to? Perhaps as good an expression of Sellars’ naturalism as any is the widely cited passage: ‘. . . in the dimension of describing and explaining the world, science is the measure of all things, of what is that it is, and of what is not that it is not’ (Sellars, 1997: XI, 41, p. 83). 3 This is enough to see that Sellars’ naturalism is metaphysical: he takes the deliverances of the theoretical/explanatory sciences to be the final court of appeal with respect to questions concerning the constituents of the world (‘of what is and what is not’).
A central aspect of Sellars’ project is that of reconciling the normativity of the manifest image with the naturalism of the scientific image. 4 That is, his stereoscopic combining of the manifest and scientific images aims to show, in part, that epistemic and other sorts of normativity can find their legitimate place in the scientific conception of the world. Such a reconciliation can be attempted in various ways; recent decades have seen several such attempts. Quine is famously, though controversially, often understood to have championed the abandonment of normativity in favor of a thoroughgoing non-normative naturalism. This interpretation of Quinean naturalism was eventually explicitly rejected by Quine himself in favor of a version of instrumental efficacy (Quine, 1986: 655); in any case it is untenable (Siegel, 1984, 1995, 1996a). Sellars, of course, took normativity to be central; he was certainly not an advocate of its abandonment.
Another influential attempt, popular among both epistemologists and philosophers of science, goes under the banner of ‘normative naturalism’. According to the normative naturalist (including the Quine of Quine (1986)), epistemic normativity is a matter of instrumental rationality. On this view, cognitive ends or goals are presupposed or ‘given’ and not criticizable on the basis of reasons; normativity is a matter of the judicious fitting of means to such ‘given’ ends. This view too is open to serious criticism (Siegel, 1989, 1990, 1996b, 1998, 2006). It is in any case not Sellars’ view either. 5
Rather, Sellars’ view is that reason is ‘both naturalistic and nonreductive’ (Hanks, 2008: 204); that is, reason is a part of the natural world but nevertheless operates autonomously from it and cannot be reduced to it. This aspect of Sellars’ view is developed at some length by both Robert Brandom and John McDowell; I concentrate here on McDowell’s discussion.
6
McDowell’s effort to reconcile the ‘realm of law’ (the realm in which Sellars’ scientific image reigns supreme) with the ‘sui generis character of spontaneity’ (the domain of the manifest image, in which human understanding, meanings, intentions and rational relations are central) hinges on his effort to ‘rethink our conception of nature so as to make room for spontaneity’ (McDowell, 1994: 77). This ‘rethinking’ is not an attempt to find some middle space between what McDowell calls ‘bald naturalism’ and ‘rampant platonism’ (McDowell, 1994: 72, 77; cf. 70–86); rather, it rejects the sharp distinction between them and builds on Aristotle’s notion of ‘second nature’ (84), according to which we are ‘animals whose natural being is permeated with rationality’ (85) and in which we can ‘reconcile reason and nature’ (86) by noting how our understanding develops in the course of ordinary human experience. It is worth quoting McDowell at some length: . . . Aristotle’s picture can be put like this. The ethical is a domain of rational requirements, which are there in any case, whether or not we are responsive to them. We are alerted to these demands by acquiring appropriate conceptual capacities. When a decent upbringing initiates us into the relevant way of thinking, our eyes are opened to the very existence of this tract of the space of reasons. Thereafter our appreciation of its detailed layout is indefinitely subject to refinement, in reflective scrutiny of our ethical thinking. We can so much as understand, let alone seek to justify, the thought that reason makes these demands on us only at a standpoint within a system of concepts and conceptions that enables us to think about such demands, that is, only at a standpoint from which demands of this kind seem to be in view. (McDowell, 1994: 82)
There are several things to note about this passage. First, McDowell is clear that this account of how ‘our eyes are opened to the very existence of this tract of the space of reasons’ is not limited to ethics: The point is clearly not restricted to ethics. Moulding ethical character, which includes imposing a specific shape on the practical intelligence, is a particular case of a general phenomenon: initiation into conceptual capacities, which include responsiveness to other rational demands besides those of ethics. Such initiation is a normal part of what it is for a human being to come to maturity, and that is why, although the structure of the space of reasons is alien to the layout of nature conceived as the realm of law, it does not take on the remoteness from the human that rampant platonism envisages. If we generalize the way Aristotle conceives the moulding of ethical character, we arrive at the notion of having one’s eyes opened to reasons at large by acquiring a second nature. (McDowell, 1994: 84)
McDowell’s account is thus not limited to ethics, but is rather a general account of how our eyes are ‘opened to reasons at large’. Second, the demands of reason are autonomous from and independent of us, in that the requirements and relations of reason ‘are there in any case, whether or not we are responsive to them’. Third, acquiring an Aristotelian ‘second nature’ is, on McDowell’s account, very much a matter of education, broadly conceived; it is ‘a decent upbringing’ that ‘initiates us into the relevant way of thinking’. As McDowell also puts it, ‘ordinary upbringing can shape the actions and thoughts of human beings in a way that brings these [rational] demands into view’ (83). Fourth, such an education is clearly a matter of initiation into the space of reasons; it is, as McDowell says, ‘initiation into conceptual capacities, which include responsiveness to . . . rational demands’, which initiation is ‘a normal part of what it is for a human being to come to maturity’. 7
Let me note three points concerning McDowell’s view. First, his proposed reconciliation of naturalism and normativity is accomplished by reconceiving nature in such a way that the normativity of language and thought are built into the very nature of human beings, who become aware of meanings, intentions and rational relations during the course of their ordinary upbringing and coming to maturity as rational animals; their rationality is, so to speak, a central aspect of their natural character. Second, McDowell explicitly rejects the idea that mind, reason, or normativity ‘transcend’ the natural world. On the contrary, such normativity is autonomous from nature in the sense that it is not determined by natural law, but it does not transcend nature in any metaphysically problematic way. As Chris Hanks characterizes McDowell’s view, on it ‘We can acknowledge that humans are rational animals and hold that reason can be exercised autonomously without placing this process outside of nature and setting up a dualism between mind and world’ (Hanks, 2008: 207). Third, and most importantly for present purposes, all this is completely compatible with the traditional view of the cultivation of reason as the fundamental aim of education and with my own development of the traditional view in terms of the fostering of rationality. The latter is completely in keeping with the conceiving of education as initiation into the space of reasons. In particular, when we ‘acquire appropriate conceptual capacities’ and are thus able to engage in what Sellars called ‘conceptual thinking’, our thinking ‘is under a standing obligation to reflect about and criticize the standards by which, at any time, it takes itself to be governed’ (McDowell, 1994: 81). Such criticism and self-criticism are central to my accounts of critical thinking/rationality. 8
Problems of compatibility?
However, not all commentators agree that my view is compatible with the Sellars/McDowell reconciliation of naturalism and normativity. Why is my view thought to be problematic? The worry is that regarding epistemic criteria and standards of normative quality as reaching beyond or ‘transcending’ particular reasoners and communities – so that, for example, reasoning in accordance with the gambler’s fallacy is a mistake even in communities that regard it as a valid form of reasoning,
9
and slavery is morally objectionable even in communities that do not so regard it – forces me to abandon naturalism and thus to diverge from Sellars and McDowell. The worry is expressed most recently by Hanks, who suggests that my view founders on its ‘relation of reason to the natural world’ (Hanks, 2008: 194):
10
Siegel’s notion of rationality stands, in a fundamental way, outside the empirical realm, populated as it is by particular circumstances, cultural traditions, and the messy contingencies of everyday life. His fallibilism wisely acknowledges that any specific application of critical thinking must take place under such conditions, but staunchly insists that rationality itself must involve criteria that stand independent of particular circumstances . . . [He] adheres to the Kantian view that the grounding of philosophical (rational) truths must stand outside the causal, contingent realm of experience . . . Siegel clings to a kind of Kantian transcendentalism . . . Siegel maintains a commitment to rational autonomy, which motivates him to argue for reason’s transcendence of the natural frame. (Hanks 2008: 199, 200, 204)
There is much to say in response to Hanks’ complaint; I will do my best to be brief.
First: it is not true that on my view rationality itself ‘stands . . . outside the empirical realm’ – at least, not as it is manifested in the reasoning, beliefs, judgments and actions of actual agents – as Hanks acknowledges (199). Does my ‘notion of rationality’ stand outside that realm? It is not obvious how this question is best understood. On one way of reading it, lots of ‘notions’ – numbers, arguments, ideals, etc. – ‘stand outside the empirical realm’ in that they are abstract objects that, by their very nature, are not extended in empirical space. Of course we could deny that the empirical realm is limited to the so extended, as McDowell does. Understood in this way, insofar as the denizens of the space of reasons are in the empirical realm, my notion of rationality is as well. On yet another possible reading, such notions as these are ours, and so are in the empirical realm insofar as we ourselves are. So the claim that my notion ‘stands outside the empirical realm’ is ambiguous: in some senses it is, in others it is not. But in no sense do I embrace or commit myself to any sort of supernatural realm, a ‘realm’ outside or beyond the ordinary, everyday world.
What about my ‘staunch insistence’ that rationality involves criteria that are independent of particular circumstances? I certainly do insist on this – as does Hanks himself (210–11). Declaring that such criteria are ‘outside the empirical realm’ is likewise ambiguous: the criteria are ours, put forward in our theorizing, and meant by us to extend beyond the boundaries of their articulation and acceptance. This is why we can say, for example, that reasoning that commits the gambler’s fallacy is epistemically deficient independently of the context in which the reasoning occurs. Indeed, this is just McDowell’s view, as we have seen: on his view, as on mine, the demands of reason are autonomous from and independent of us in that the requirements of reason ‘are there in any case, whether or not we are responsive to them’. These demands, and the related criteria, are not ‘outside the empirical realm’ on McDowell’s view or on Sellars’ – the domain in which they reside, the ‘space of reasons’, is epistemically independent of/autonomous from the empirical realm, but not (metaphysically) ‘outside’ it. The same is true on my view.
Second: I have tried to be careful to distinguish stronger and weaker forms of ‘transcendence’. I do offer a ‘transcendental’ argument for justifying rationality (Siegel, 1988: Postscript; 1997: ch. 5 and Epilogue), but here ‘transcendental’ does not mean ‘outside the empirical realm’, but rather refers to the argument form according to which something (in this case rationality, or at least an acknowledgment of it and a commitment to it) is necessary if something else (in this case the serious calling of rationality into question) is possible. When I say that epistemic criteria transcend local contexts I do not thereby assign to criteria some special metaphysical status. Rather, I affirm just that what people in a given context say or think is criticizable on the basis of suitable criteria whose force is independent of and extends beyond the bounds of that context, and whose suitability is itself a matter open to reasoned criticism. This is more or less the same point Sellars makes when he says ‘to be able to think is to be able to measure one’s thoughts by standards of correctness’, presuming, of course, that it is possible that one’s thoughts might fail to measure up; the same point that McDowell makes when he says that our thinking ‘is under a standing obligation to reflect about and criticize the standards by which, at any time, it takes itself to be governed’, presuming, again, that our thinking can fail to meet such standards – which it clearly can on McDowell’s view, since according to it ‘rational requirements . . . are there in any case, whether or not we are responsive to them’. ‘Transcendence’ in my hands does not have the untoward metaphysical features that I join Sellars, McDowell and Hanks in rejecting (Siegel, 1997, 1999a, 1999b, 2001, 2007b, 2011).
I suspect that this tangle results from insisting that my epistemological account of rationality/critical thinking be understood metaphysically, and by failing to distinguish epistemological from metaphysical naturalism. Like Sellars and McDowell, I too embrace metaphysical naturalism, and reject all forms of supernaturalism. What I reject is epistemological naturalism, when that view is understood to involve or entail that ‘metaphysically naturalist’ standing – i.e. something’s standing as a thing that has ‘naturalistically emerged’, e.g. by way of McDowell’s ‘ordinary, decent upbringing’, or its status as something genuinely in the world – guarantees positive epistemic standing. 11 The reason is simply put: The ‘natural’ includes or gives rise to both the epistemically good and the epistemically bad, both truths and falsehoods, both justified and unjustified beliefs. So once we agree that everything arises or emerges naturally, i.e. as a result of natural processes, or is natural in that it is ‘in the world’, the problem of distinguishing the good from the bad remains. Naturalist metaphysical status – i.e. being a product or a part of nature – is no indication of epistemic status or quality. Once we have agreed that minds generally, and the minds of critical thinkers in particular, emerge naturalistically and are part of the natural world, we need to deal with the fact that the minds of uncritical thinkers likewise so emerge and reside. Doing so requires the rejection of epistemological naturalism as here understood, but is of course compatible with metaphysical naturalism. Sellars, McDowell, Hanks and I are agreed that critical thinking is embodied and that human nature is part of nature. We agree, further, that language, minds, ideas and ideals all emerge naturalistically, i.e. from natural processes. However, we might disagree on what follows from that emergence. For the reasons just rehearsed, I reject the claim that such naturalistic emergence or status establishes epistemic quality. Sellars, McDowell and Hanks should agree, given their embrace of the independence and autonomy of the ‘space of reasons’ from the ‘realm of law’, and the ‘non-reductive’ character of reason/rationality. That is, my rejection of epistemological naturalism and my positive accounts of rationality and critical thinking are completely compatible with the metaphysical naturalism embraced by all four of us. 12
A key missing ingredient
While there is then no incompatibility between the two, there is something missing from the Sellars/McDowell view that is central to (my version of) the traditional ideal with which we have been comparing it. Suppose that Sellars and McDowell are right about what is involved in humans’ becoming competent language users and conceptual thinkers. It does not follow that as a matter of fact competent language users/conceptual thinkers are good at making normative epistemic judgments. Epistemic discourse may be ‘irreducibly normative’, as Rosenberg puts it; Sellars may be correct that ‘to be able to think is to be able to measure one’s thoughts by standards of correctness’. But one can ‘measure one’s thoughts by standards of correctness’ well or badly; one can, once one’s ‘ordinary decent upbringing’ opens one’s eyes to the space of reasons and enables one to maneuver in that space, maneuver well or badly; one can be ‘initiated into the relevant way of thinking by a decent upbringing’ and yet engage in such thinking either well or badly; one can be a competent concept user and player of the game of giving and asking for reasons and yet play it either well or badly, epistemically speaking, in that one can, despite one’s conceptual competence, judge badly the probative force of candidate reasons. If so, the ability to reason well is not guaranteed by an ordinary decent upbringing or the acquisition of appropriate conceptual capacities, unless having mastered a range of concepts somehow prohibits mistakes in reasoning or leads inevitably to epistemically high-quality thinking. This is of course not the case. Being ‘in’ the space of reasons does not ensure that good reasoning will result. Every competent language user uses it in the space of reasons, yet not every such user is a critical thinker. More is needed. 13
Consequently, while the Sellars/McDowell view is compatible with the traditional ideal, that ideal requires more: it requires that ordinary decent upbringing and the acquisition of appropriate conceptual capacities – that is, normal human development as conceived by Sellars and McDowell – be supplemented by explicit educational intervention aimed at enhancing students’ abilities to reason well. This is not surprising, since the mental life of every normal human being is lived ‘in’ the space of reasons but irrationality and uncritical thinking abound nonetheless. Simply acquiring appropriate conceptual capacities during the course of learning language (McDowell, 1994: 125) is not enough, educationally speaking. At any rate, it is not enough once we are agreed that (1) one can learn a language yet reason badly, evaluate reasons badly, and/or fail to be disposed to believe, judge and act in accordance with one’s evaluation of the strength of candidate reasons, (2) we are obliged to treat students, as everyone, with respect as persons, and (3) treating students in this way requires doing our best to foster their rationality (Siegel 1988, 1997).
Conclusion: Education as initiation into the space of reasons
R. S. Peters famously argued that education is fundamentally a matter of initiation. 14 I do not want here to defend Peters’ view generally. But I do think that the idea that education importantly involves initiation is basically correct, and that it is readily adapted to the idea that education fundamentally involves the initiation of students into Sellars’ space of reasons. However, in so far as we embrace the thesis that a central aim and guiding ideal of education is the fostering of reason or rationality, initiating students into the space of reasons as Sellars and McDowell conceive it – by way of learning language and in that way acquiring appropriate conceptual capacities – is, as we have seen, not by itself enough. Such an ordinary, decent upbringing needs to be supplemented with explicit educational intervention aimed at fostering students’ critical thinking – that is, aimed at fostering their ability to evaluate reasons well and their dispositions to seek reasons, evaluate their power and convicting force competently, and be guided by the results of such evaluation.
There is thus no reason to think that there is any serious conflict between the traditional educational ideal concerning the fostering of rationality and the Sellars/McDowell view. Education can indeed be seen as inherently involving initiation into the space of reasons. But such initiation, to satisfy the ideal, must do more than rest on ordinary decent upbringing and language learning. It must pay explicit attention to the fostering in students of the ability to assess reasons well, including a sophisticated understanding of the epistemology underlying such assessment, and to the fostering of the suite of dispositions, habits of mind and character traits constitutive of the critical spirit. Only thus enhanced will ‘initiation into the space of reasons’ fulfill our deepest educational ideal.
Footnotes
Acknowledgements
A version of this paper was presented at the Congress of the German Society of Philosophy (Deutscher Kongress für Philosophie), Munich, September 2011. I am grateful to Krassimir Stojanov for the invitation to participate in this event, to him and its other participants, especially Randall Curren, and to Jonathan Adler, Catherine Elgin, Israel Scheffler and Amie Thomasson for very helpful comments, suggestions and advice.
Notes
Biographical note
Harvey Siegel is professor of philosophy at the University of Miami. His interests include epistemology, philosophy of science and philosophy of education.