Wilfrid Sellars and the task of philosophy

Abstract

Critical attention to Wilfrid Sellars’s “Philosophy and the Scientific Image of Man” (PSIM) has focused on the dubious Peircean optimism about scientific convergence that underwrites Sellars’s talk of “the” scientific image. Sellars’s ultimate Peircean ontology has led Willem deVries, for instance, to accuse him of being a naturalistic “monistic visionary.” But this complaint of monism misplays the status of the ideal end of science in Sellars’s thinking. I propose a novel reading of PSIM, foregrounding its opening methodological re ections. On this reading, the central point of the paper is to accuse figures like Wittgenstein and Strawson, whom I call “analytical quietists,” of taking the unity of intellectual endeavor as somehow given. Such unity as is forthcoming is, Sellars tells us, a task. I conclude by noting that a structurally similar accusation of too easily presumed unity emerges at the end of the paper, against a familiar sort of anti-relativistic moral theorizing. Thus, Sellars’s conception of the task of philosophy is, at least potentially, a point of surprising ethico-political significance as well.

1 Introduction: The scientific image

Wilfrid Sellars’s “Philosophy and the Scientific Image of Man” (PSIM) has a curiously mixed reputation. Its opening pages, what I’ll call its “metaphilosophical prologue,” contain some of the most familiar passages in Sellars’s oeuvre, starting with the first line: “The aim of philosophy, abstractly formulated, is to understand how things in the broadest possible sense of the term hang together in the broadest possible sense of the term” (1). And his depiction of the central drama of modern philosophy as a clash between “scientific” and “manifest” images, though still challenged, is a philosophical commonplace.¹

But the connection between the prologue and the central themes of the paper can seem tenuous at best. For instance, while the prologue is broadly pluralistic, Sellars insists there will be no piecemeal synthesis of various scientific images with the manifest image. With relatively little fanfare he “telescopes” them into one, the scientific image, and proceeds to consider the clash as a clash of idealized images. Perhaps the relation of the prologue to this problem is simply that, as the aim of philosophy is to understand how things hang together, if the table of physics and the table of common sense do not “hang together” very well, we have a problem.² This, however, makes it look like a bit of intellectual throat-clearing, mere prologue. The paper also ends with a curious meta-ethical discussion, which symmetry begs us to call a meta-ethical epilogue, and which, on the standard line, must seem like an afterthought.

Moreover, what argument the paper has strikes many as dubious. The centrality of the scientific image makes it seem to presuppose a dated (and undefended) monism about science. In a passage I discuss in detail later, Willem deVries urges that “science may end up more of a patchwork of locally profitable schemes than Sellars would ever have countenanced” (2016, 14).³ Part of what he has in mind is Sellars’s notorious claim that the conceptual development of science makes the noumenal world in principle accessible at the Peircean end of inquiry: Sellars is a “monistic visionary.”⁴ The suggestion is that there is a strand best ignored in Sellars’s thinking, connecting some out of date scientific monism to his Peircean ontological monism.

It is unsurprising, then, that sympathetic readers prefer to focus on the systematic role of PSIM: it is a useful overview and introduction to Sellars’s thinking, its argument aside.⁵ I shall argue that this reception betrays a failure to detect the central claim of the paper. It is announced in the prologue: “to the extent that there is one picture to be grasped reflectively as a whole, the unity of the reflective vision is a task rather than an initial datum” (PSIM 4). Those who assume that all is one, that there is just one world, are Sellars’s target, not his allies. I urge that Sellars never recants this pluralism.

It is noteworthy that Peirce does not appear in the paper, and that when Sellars does invoke the end of inquiry elsewhere, it is not quite as the objection above primes one to expect. Our ideal Peircean descendants will be able to picture “any part” of the one and only world, but “this does not require a single picturing of all parts” (SM 142). (“Part” is of course not to be understood spatially.) Thus, Sellars’s conception of the end of inquiry itself does not require telescoping the images. This raises the question just what role the regulative ideal of a single scientific image plays.

On my account, Sellars supposes this ideal to license his insistence on the reality of disunity. In the passage I quoted from the prologue, Sellars targets a view I call “analytical quietism,” which he associates with the ordinary language tradition. (He mentions both Strawson and Wittgenstein.) This is a tradition friendly to pluralism, so Sellars’s critics are right that he distinguishes himself from would-be pluralists. The error is to think he is moving towards an anti-pluralistic monism. Instead, Sellars maintains paradoxically (see §7 below) that one who rejects the regulative ideal of monism cannot really be pluralistic.

It is not some presumed unity—to which we need only be recalled—but dis-unity that gives philosophy its task. Quietism, Sellars charges, is a presumptuous failure to detect a real intellectual demand. With this thought in mind, the “epilogue” looks very different: rather than a mere postscript, it reflects the place where philosophy’s “task” is most pressing. I conclude by highlighting the way disunity informs Sellars’s meta-ethical thinking. In a slogan that applies to politics as well as theoretical activity, Sellars maintains that unity is a task, not a given.

2 Scientific realism and the clash of the images

I begin by recounting the standard reading of PSIM, both to situate the discussion that follows and to motivate the need for an alternative. I don’t mean anything in this reconstruction to be controversial. I merely want to draw attention to a set of familiar problems that derive from situating it in the standard way.

We pick up from the introduction of the clash of the images at the end of the prologue. One reason Sellars thinks the competition between scientific and manifest images has been overlooked is the failure to recognize the historicity of the latter. To counterbalance this tendency, he gives an historical fiction of the development of the various images from an “original” image in which personhood is everywhere. The manifest image is the result of depersonalizing the non-human world, so its fundamental ontological feature is the presence of persons and non-persons. The scientific image is a result of postulational sciences, and arises from explanatory pressures within the manifest image.

Though there is an abiding temptation to construe the manifest image as “the world of common sense,” this is not quite right: it is a particular, refineable and hence changeable, way of conceiving the world. As we shall see, Hegelian and Wittgensteinian ideas about sociality and publicity are essential to its articulation. It is a domain of philosophical contestation in its own right. Thus, “the so-called ‘analytic’ tradition in recent British and American philosophy, particularly under the influence of the later Wittgenstein, has done increasing justice to the manifest image” (15).

This image is the object of theoretical focus for the “perennial tradition” in philosophy, which emphasizes a dualism of causation. In addition to being in the physical world, and so subject to physical causation, persons are located in rational space, and so moved by rational considerations. The perennial philosopher sees Hume’s associationist psychology, for instance, as an abortive attempt to reduce rational causation to the physical one studied in the natural sciences (PSIM 15–18).

But the irreducibility insisted upon by the perennial philosopher is itself a source of tension, for, as Sellars notes later, “the scientific image presents itself as a closed system of explanation” (PSIM 36): science does not allow for another kind of causation than what figures in scientific explanation. And thus the perennial philosophy contains within itself an antinomy, insofar as it embraces scientific explanation as emerging out of it. “[A]lthough methodologically a development within the manifest image, the scientific image presents itself as a rival image” (PSIM 20).

Sellars is a scientific realist. As he says in EPM, “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” (§41, 173). Thus, he suspects the scientific image will in some sense win out in this rivalry. It is worth stressing, though, that the precise sense is hard to pin down. For we shall need “stereoscopic vision,” which means at least that the manifest image makes an ineliminable contribution to an ultimate view. Much more on this as we proceed.

Be that as it may, the thesis that the scientific image will win out in the clash is a central theme of the paper. The reader looking for an argument for this thesis is bound, then, to be frustrated by Sellars’s next move. Distinguishing two alternatives—the instrumentalist response that would have the manifest image win out, and the quietistic view that there is no real conflict—he sets aside instrumentalism with the acknowledgment that it “has been defended by able philosophers” (PSIM 26).⁶

A charitable reading of this point is available. Sellars locates his thinking—including his scientific realism—in an historical tradition of attempts to engage the clash of the images. So the paper has two themes: an historical one, developed at length with respect to Descartes, and in passing for Spinoza and Hume, that modern philosophy has been obsessed with the clash of the images; and a first-order one, that Sellars’s own position has unique advantages over its historical antecedents. Sellars wants to advertise and explore his revisionary approach to the manifest image, and as such he is not concerned with views like instrumentalism that, from the standpoint of the manifest image, are more conservative.

Sellars is explicit as to who his target is in the paper: the quietist who sees no conflict. But if all he tries to do is show that his realism is defensible, it is not clear in what sense even the quietist is needed as a dialectical foil. If, however, his goal is to foreground the clash itself, as opposed to his resolution of it, the role of quietism becomes clearer. This hint is the key to my alternative reading. First I want to raise the familiar problems on the received view.

3 Problems for scientific realism

When Sellars sets aside instrumentalism, he is discussing Descartes’s scientific realism. He claims the “considerations which led philosophers to deny the reality of perceptible things led them to a dualistic theory of man” (29). Cartesian dualism is a first attempt at reconciling the scientific and manifest images, a consequence of Descartes’s inability to see a scientific treatment of conceptual thinking “as a live option” (30). This is where Sellars’s positive account begins.

He considers three areas where philosophers are rightly concerned with the prospects of the scientific image capturing everything about the manifest image: conceptual thinking; the qualitative character of experience; and the concept of a person as a subject “who finds himself confronted by standards” (PSIM 38) that go beyond his own immediate desires. He addresses each in turn, mobilizing thoughts he developed at greater length (and more persuasively) elsewhere. This culminating discussion earns the paper’s reputation as a “flagship” article: one does get an idea of how disparate strands in Sellars’s thinking fit together. However, there is something approximating critical consensus that Sellars’s only persuasive reply (here) is to the first. I take the three in turn.

(i) Sellars takes his proto-functionalist theory of meaning to open the room for a scientific treatment of the conceptual. As he summarizes, “if thoughts are items which are conceived in terms of the roles they play, then there is no barrier in principle to the identification of conceptual thinking with neurophysiological process” (34). This is a compressed depiction of a highly controversial account, but I don’t suppose anyone denies that account is worth taking very seriously. Crucially, Sellars self-consciously construes this as contributing to “the primacy of the scientific image” (32).⁷

The historicity of the images is important to this thought: Hume, for instance, failed to articulate a coherent scientific picture of humankind for lack of the appropriate theoretical tools, and so there was justice in the Kantian recoil. Post-Kantian emphasis on the social nature of conceptual thinking plays a decisive positive role. It is no coincidence that Sellars characterizes the scientific image that supplants Hegel’s thought as evolutionary. Darwin represents a crucial intellectual turning point, making a scientific account available. I return to this theme in §8 below. Regardless, conceptual thinking as such is not supposed to pose any principled problem for scientific realism.

(ii) About the qualitative character of experience, what Sellars says is at best less clear. So long as physics remains particulate in its fundamental metaphysics, it is inconsistent with the ultimate homogeneity of experience. Sellars suggests, though, that perhaps an alternative physics will emerge, based on what in the Carus Lectures he refers to as a non-Whiteheadian process ontology (FMPP 59, II §124). It is perfectly in keeping with Sellars’s metaphilosophy for a philosopher to generate scientific hypotheses like this.⁸ However, whether defensible or not, all agree this is a dark part of his thinking.⁹

More to the point, there is merit in Stanford’s remark that it is “difficult to regard this hopeful invocation of an imaginary physics so different from our own as anything more than whistling past the graveyard” (2012, p. 34). If we confine ourselves to PSIM,¹⁰ Sellars’s purported response to the problem of the qualitative character of experience sounds like the thought that maybe someday scientists will solve the problem. And this hardly constitutes a solution in its own right. So it is unclear how it is responsive to the problem.

(iii) Having addressed the problem of the qualitative character of experience to his satisfaction, Sellars finally turns directly to the concept of a person, subject to norms. Where to this point Sellars has zigged, here he zags. Unlike with conceptual thinking and the qualitative character of experience, Sellars sees no hope for a scientific account of personhood, and concludes that the framework of normative evaluation can be “joined” to the scientific image. This is why the concluding discussion can seem like a postscript: the normative can somehow be set aside, exempted from the problems he has been considering.

It is far from clear just what this joining consists in.¹¹ Moreover, while Sellars distinguishes his view from Spinoza’s, according to which the scientific image “dominates the stereoscopic view” (PSIM 9), the first two responses seem to suggest that the scientific image would be all that is needed. As I discuss in §9 below, this is precisely how Sellars talks elsewhere, in discussions restricted to the first two problem areas. Thus, a crucial point of PSIM—that we need a stereoscopic view—emerges in what on the standard reading looks like a problematic postscript.

I return to these issues at the end. But I want to raise a different concern for now. Consider this comment:

To say that a certain person desired to do A, thought it his duty to do B but was forced to do C, is not to describe him as one might describe a scientific specimen. One does, indeed, describe him, but one does something more. And it is this something more which is the irreducible core of the framework of persons. (PSIM 39)

Though this characterization is practical, it also echoes a famous line of thought from EPM captured in passages like these:

For to say that a certain experience is a seeing that something is the case, is to do more than describe the experience [...] (§15, 144).

[...] 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. (§36, 169).

Characterizing an episode as a standing in the space of reasons is not (merely) describing that episode.¹² But while in PSIM this claim is limited to practical contexts, the reference to EPM should remind us that covert conceptual episodes are, quite generally, located in the space of reasons.

If attributions that are not merely descriptions cannot be accommodated in the scientific image (so they need an exemption), where does that leave the earlier functionalist account of conceptual thinking? Wasn’t it supposed to show that certain episodes’ standings in the space of reasons can be accommodated to the scientific image by functional-role characterizations, so they do not need an exemption? This point has not been sufficiently appreciated: Sellars’s treatment of persons at the end of PSIM seems to suggest that his own proto-functionalist theory of meaning should be impossible.¹³

Waiving this worry, if the normatively articulated framework of desires and intentions can be joined to the scientific image (whatever that means), why can’t other episodes constitutively located in the space of reasons also be “joined” to it? Why did we need the functionalist account of conceptual thinking? Clearly, addressing this question puts a lot of pressure on the metaphor of “joining,” which—I think any reader would concede—is at best hastily sketched in the concluding section of PSIM.

This constellation of interpretive questions leads O’Shea (2007) to characterize Sellars’s view as “naturalism with a normative turn.” But the problem is deeper than O’Shea concedes, more nearly an interpretive antinomy: in PSIM at least, Sellars suggests two different ways of accommodating the normative in nature, corresponding to his first and third problem areas. Sellars himself stresses that the replies are different. The question is whether they are compatible.

Even raising this question puts the paper in a very dissatisfying light. On the received reading, the purpose of the paper is to cast Sellars’s scientific realism as a plausible response to a longstanding philosophical problem, the clash of the images. To that end, Sellars considers three potential objections to scientific realism in general: objections from conceptual thinking, the homogeneous character of perceptual experience, and the normative status of persons. Of his replies to these concerns, only the first can be said to really be responsive. The second whistles past the graveyard, and the third simply concedes the objection. While the “joining” metaphor is supposed to mitigate this concession by allowing us to set aside the normative for separate treatment, it threatens to undermine the first, genuinely responsive, reply, which was a scientific account of normative episodes. Thus, it is hard to see that Sellars has given us much of a reply at all.

It is little surprise, then, that sympathetic readers refrain from characterizing PSIM as a defense of scientific realism. Rather, it is a “flagship article,” “a manifesto” without “much direct argument” (see n. 5 above). It provides an impression of how Sellars’s preoccupations fit together, but from this perspective it does little else besides.

4 Theory and practice in Sellars’s meta-philosophy

I have suggested that with a more careful look at the metaphilosophical prologue, the paper will appear differently. The problem is that Sellars’s meta-philosophy can seem quite banal.¹⁴

To get interpretive leverage, I begin with a prima facie plausible criticism of Sellars’s metaphilosophical view. P. Kyle Stanford (2012) observes that practical characterizations (philosophy is a species of know-how, for instance) give way quickly, via what he (Stanford) calls an “ocular turn,” to visual metaphors: having one’s eye on the whole, bringing a picture into focus, and ultimately constructing or employing images. Stanford claims that the problems derive from this unexplained transition. Had we remained in the practical key, we might have avoided the difficulties induced by “scientific realism” altogether.

Stanford takes the prologue to frame a basically first order defense of scientific realism. His claim is that this framework involves metaphilosophical blunder. When he quotes the following discussion of instrumentalism from Sellars, he makes much of the italics:

On this view, the theoretical counterparts of all features of the manifest image would be equally unreal, and that philosophical conception of man-in-the-world would be correct which endorsed the manifest image and located the scientific image within it as a conceptual tool used by manifest man in his capacity as a scientist. (PSIM 32, quoted in Stanford 2012, p. 35.)

The “ocular turn,” Stanford suggests, enabled Sellars to justify (to himself at least) the simplifying assumption that there is one scientific image. Had we not taken that turn, had we not begun thinking of the aim of philosophy in terms of the construction, ultimately, of a single image, however multi-faceted and multi-dimensional, we would not have been led to the conclusion that scientific realism is an all or none affair.

Stanford wants primarily to leave room for an instrumentalist reading of quantum mechanics. And the position Sellars announces at the outset can accommodate that. Consider: “the historian reflects not only on historical events, but on what it is to think historically” (PSIM 2). History is, to use a term of art he introduces elsewhere, a rational discipline, “a field of inquiry in which good reasons can be offered for answers to questions belonging to the field.”¹⁵ Thus, deference to the historian in his discipline extends not just to questions that are “in a primary sense historical questions”, i.e., the conclusions of historical reasoning, but to questions about method, about what counts as historical reasoning. It is part of the historian’s “business to reflect on his own thinking—its aims, its criteria, its pitfalls” (PSIM 2). While there is no sharp line between the “persistently reflective” historian and the philosopher of history, it is the practicing historian who is sensitive to the criteria of adequacy for historical thinking.

What goes for history surely goes for quantum mechanics. It is no part of Sellars’s methodological reflections to imagine philosophers dictating from on high how local questions need to be settled. He can be equipped with the thought that questions of realism are, at least in part, methodological questions for given rational disciplines.¹⁶ It is, then, a (socio-)empirical question whether instrumentalism is a live option in the interpretation of quantum mechanics. Whatever Sellars’s own opinion might have been, the Sellarsian who is not up to date on the current state of quantum mechanical theorizing is ill-advised to rule a priori on such questions.

Thus, at least Sellars’s opening methodological reflections provide latitude for methodological pluralism, and—I submit—local instrumentalist accounts.¹⁷ Sellars need only be opposed to a global instrumentalism of the sort that Stanford is not trying to motivate. But it is not as if Stanford has simply misread Sellars. He is tapping into the line of discontent deVries expressed, quoted earlier: “science may end up more of a patchwork of locally profitable schemes than Sellars would ever have countenanced” (2016, 14). Stanford and deVries disagree about just what problem Sellars’s (putative) inattention to methodological pluralism has induced—deVries is no instrumentalist. But they agree that a more discriminating, more pluralistic, and indeed, more practical meta-methodological picture would avoid whatever problems Sellars’s conception incurs.

This brings to the fore the relation between methodological pluralism and Sellars’s invocation of the scientific image. The critics are aware that the latter is an idealization: “There are as many scientific images of man as there are sciences which have something to say about man” (PSIM 20). So the question is what role this idealization plays. If the more pluralistic prologue gives way to an attempt to motivate the primacy of the scientific image, the idealization might license ignoring the official pluralism. Thus, even putting the question of instrumentalism aside, Stanford’s worry cuts deep.

5 Sellars’s pluralism

The objection rests, however, on a mistake. I argue now that Sellars never abandons the pluralism he announced in the opening paragraphs. Whatever the point of the idealization is—I return to that below—it does not license ignoring Sellars’s pluralism.

Perhaps the best evidence of Sellars’s supposed monism in PSIM is when he says: “With due precaution, we can unify the biochemical and the physical images; for to do this requires only an appreciation of the sense in which the objects of biochemical discourse can be equated with complex patterns of the objects of theoretical physics” (PSIM 21). Though this can sound problematically reductionist,¹⁸ Sellars’s claim is subtle. Some hedging is clear: we are invited “with due precaution” to appreciate “the sense” in which such an equation might be made out. Spelling out these caveats would exonerate Sellars from charges of an especially vicious sort of reductionism. But it goes much further than this.

Among the various scientific images of man, there are “man as he appears to the biochemist, to the physiologist, to the behaviourist, to the social scientist” (PSIM 20), all of which Sellars contrasts with the man of the manifest image. In the passage that talks of an equation, he considers only the easiest case. Thus we can wonder whether and to what extent the objects of physiology or behavioral psychology will be equated with the objects of physics.

Sellars does think that there is only one set of “objects”—not, indeed, objects at all but processes—such that in the long run everything that can be said to be will be identified in some sense with complexes of them. But in the passage I quoted, Sellars is not insisting on this Peircean ontological monism. Instead, he is noting the limits on the familiar partial identification of biochemical with physical entities: identifying theoretical entities appears relatively cheap. As this interpretive claim can seem surprising, it is worth spelling out.

I noted that both Stanford and deVries derive their worries from the observation that Sellars is insufficiently attuned to the practical dimension of scientific understanding.¹⁹ There is a fair point in this vicinity—Sellars’s tendency to focus single-mindedly on epistemological issues in science distorts what little he says about scientific practice. But given that, it is notable how pluralistic he is: continuing the discussion of the variety of scientific images, he writes: “each scientific theory is, from the standpoint of methodology, a structure which is built at a different ‘place’ and by different procedures within the intersubjectively accessible world of perceptible things” (20). Each has, he says “a certain autonomy.” The discussion of the equation of physical and biochemical entities occurs when he picks this theme back up. Here is his elaboration:

To make this equation, of course, is not to equate the sciences, for as sciences they have different procedures and connect their theoretical entities via different instruments to intersubjectively accessible features of the manifest world. But diversity of this kind is compatible with intrinsic ‘identity’ of the theoretical entities themselves, that is, with saying that biochemical compounds are ‘identical’ with patterns of subatomic particles. For to make this ‘identification’ is simply to say that the two theoretical structures, each with its own connection to the perceptible world, could be replaced by one theoretical framework connected at two levels of complexity via different instruments and procedures to the world as perceived. (PSIM 21)²⁰

When deVries (2016, p. 7) quotes from this passage, it is to show that while Sellars is “not totally naïve” about methodological pluralism, he is nevertheless firmly committed to the unity of science. But surely Sellars’s point is that even after identity claims are made—as, perhaps naively, he thinks they are being made in one case—and so theoretical frameworks are in some sense identified, we still need multiple levels of “complexity” corresponding to originally autonomous rational disciplines. “The” scientific image, a product of entity-wise identification, does not involve methodological identification or reduction. That is to say, he is warding off a potential over-reading of his idealization. Whatever epistemological unification might be forthcoming simply does not require methodological flattening of any sort.

Of course, one might still insist that a full recognition of the plurality of practices is incompatible with the epistemological unification Sellars foresees. But we should be clear about just what that unification is. For the point about not equating the sciences is not restricted to methodology. There is a pluralism, he says, about “principles” too:

Obviously a specific pattern of physical particles cannot obey different laws in biochemistry than it does in physics. It may, however, be the case that the behaviour of very complex patterns of physical particles is related in no simple way to the behaviour of less complex patterns. Thus it may well be the case that the only way in which the laws pertaining to those complex systems of particles which are biochemical compounds could be discovered might be through the techniques and procedures of biochemistry, i.e. techniques and procedures appropriate to dealing with biochemical substances. (PSIM 21)

Not only will biochemistry not be replaced by physics, but the emergent laws of biochemical substances will only be detectable via biochemical methodology.²¹ This is the import of a passage I cited at the outset: “The Peirceish method of projection must enable picturings [...] of any part, but this does not require a single picturing of all parts” (SM 142). The heterogeneous methods of the various sciences are ineliminable, he maintains, not only as a practical matter—we have different instruments for different levels—but as an epistemological one too.

According to deVries, “good explanations come in many sizes and shapes, and each one tells us something important about the structure of the world” (2017a, 167). Though he offers this in opposition to Sellars, it fits Sellars’s explicit view perfectly. Obsessive concern with Sellars’s (ideal-limit) monistic ontology has caused even his most sensitive readers to underestimate this pluralistic feature of his thought.

6 The unity of intellectual endeavor

Thus, Sellars agrees that science is a patchwork. Even at the ideal limit there will remain a sense in which this is so. There will not be one ultimate picture embracing all parts of the world; Sellars is only committed to the (also admittedly dubitable) idea that all “parts” of the world can be “ultimately” pictured.

I quoted him as saying, though, that a “pattern of physical particles cannot obey different laws in biochemistry than it does in physics.” If biochemists endorse an emerging identity between their entities and physical particles, but have the complex entities doing things that look physically impossible, we have a problem. Perhaps this assumption is anti-pluralistic. Dionysius Christias, for instance, offers Sellars the thought that “the ultimate raison d’être of the division of intellectual labour in empirical inquiry is the creation of a coherent (single, complex and stratified) picture of the world and our place in it” (2018, p. 1317). Sellars does seem to have the idea that a single unified “world story” is a transcendental condition of inquiry.²²

This line of thought would (as Christias agrees) require Sellars to oppose pluralism. We can put the pluralist reaction in deVries’s terms of local profitability: if biochemical research is profitable for biochemical purposes, why should apparent physical impossibility matter? And so long as we maintain that the sciences are distinct, a plurality of purposes seems inevitable. The pluralist might simply deny that the bringing into being of the Peircean millenium is the raison d’être of scientific activity.²³

Against Christias’s interpretation, then, it is worth recalling that in PSIM Sellars does not invoke Peirce, or, more generally, the idea that a single world story is a transcendental condition of inquiry. Sellars’s “telescoping” move is, I maintain, underwritten by a more quotidian observation. In a striking passage I quote at the very end of this paper, he appeals to the neutral purposes of “explanation and prediction” (SE 231).²⁴ Of course, interests in explanation and prediction often derive from other reasons, but such interests are also ends in themselves. We might talk about intellectual curiosity: “I was just curious” is a perfectly adequate answer to certain “Why” questions.

It is helpful to note that intellectual curiosity is unparadoxically self-applicable: we can—and should!—be curious about why, e.g., we find this an object of curiosity, and not that.²⁵ Thus one need not construe intellectual curiosity in a theoretically weighty way, as a brute explainer not itself to be explained. The only point I stress is that a desire for unification needs no further explanation than curiosity. Sellars’s appeal to “the” ideal scientific image need only be backed by the observation that intellectual curiosity won’t be sated without unification. Of course, any given case of investigation is likely to have other explanatorily relevant features too.

Moreover, while one (pseudo-Peircean?) problem is that physically impossible biochemistry would be incompatible with our best physical picture of the world, that is not the problem Sellars is drawing attention to. The point about emergent principles was that physics might have to defer to biochemistry. No doubt, the reality would be a complex negotiation involving interdisciplinary work (comparable, presumably, to that insinuated by the label “biochemistry”). As Christias stresses, there is no explanatory hierarchy here. Two intellectual endeavors, each of which counts as a rational discipline in its own right, can come into conflict, and when they do there is a problem. One can desire to solve the problem without meaning to introduce a candidate element of the ultimate Peircean picture: the concepts of (non-Peircean) science are local problem-solving devices, as deVries suggests.

This anti-pluralism derives directly from the (pluralistic) methodological deference announced in the prologue. Philosophy is the discipline in which “no intellectual holds are barred” (PSIM, 1). By implication, in other disciplines—sciences or not—intellectual holds are barred. For certain purposes of biochemical research, the question of the relationship between biochemistry and physics might be barred. A mismatch is a problem, no doubt, but Sellars stresses that it does not call the epistemic credentials of biochemistry (or physics) into question. Perhaps we should say that the potential for such mismatches reflects the rules constitutive of the discipline: in these disciplines holds are barred. For some problems we should expect that only the philosopher, as a matter of her professional identity, is obliged to be moved by them.²⁶

Sellars’s paper concerns philosophy and the scientific image of man. As philosophy is the discipline in which no intellectual holds are barred, it is from the philosophical perspective in particular that clashes between various local disciplines (including sciences) must seem like problems. That they look that way to the philosopher means neither that they look that way to the local specialist nor that they ought to. The extent of Sellars’s anti-pluralism is his insistence that “clashes” are problems.

7 Against quietism

Thus, Sellars’s characterization of philosophy as the discipline in which no intellectual holds are barred yields his putatively anti-pluralistic conception of the unity of intellectual endeavor. We cannot divorce the bracing if vague prologue from the rest of the paper. I want to go further now, and argue that the prologue frames the discussion of the paper as a defense of this conception of philosophy.

The bulk of the prologue concerns the relationship between philosophy and other rational disciplines, given especially that philosophy has no special subject matter. Perhaps we should conceive philosophy as analysis, as opposed to “the discovery of new truths” (3). The best metaphorical elaboration of analysis, he claims, is of “bringing a picture into focus.” If I am right, this view, which I’ll call analytical quietism, is the primary target of criticism for the paper.

This approach has two serious drawbacks. First,

It suggests that the special disciplines are confused; as though the scientist had to wait for the philosopher to clarify his subject-matter, bring it into focus. To account for the creative role of philosophy, it is not necessary to say that the scientist doesn’t know his way around in his own area. What we must rather say is that the specialist knows his way around in his own neighbourhood, as his neighbourhood, but doesn’t know his way around in it in the same way as a part of the landscape as a whole. (PSIM 4)

Sellars distinguishes two projects: the first-order project of working in one’s own area, and the higher-order integrative project of locating one’s area in “the landscape as a whole.” This is the point considered above: if problems induced by “intertheoretic” reflection do impose themselves on the day to day work of the researcher, this does not call into question her epistemic credentials.

The problems Sellars is interested in emerge when specialists look up from their specialties. A very different reply than Sellars’s would be to say that specialists ought not to do that; their temptation to raise such questions is an expression of confusion. On this alternative view, the clashes I have insisted Sellars construes as problems are instead a result of language “going on holiday.” There is no distinct integrative project; rather, the felt need for such a project results from a kind of conceptual confusion. Philosophical clarity would allow the specialist to avoid being bothered by “the landscape as a whole.”

Of course, this vaguely Wittgensteinian line of thought is not the only one that fits the rubric of analytical quietism. My point depends solely on the idea that it does fit.²⁷ When Sellars introduces the clash of the images, Wittgenstein and Strawson are singled out as modern-day representatives of the perennial philosophy that speaks for the manifest image. Here in the prologue, before the clash of images, we are already given reason to doubt the quietist refusal to detect problems. The idea that philosophical problems, such as those arising from the epistemology of micro-physics or history, are a result merely of language going on holiday suggests that those who tell us otherwise do not understand their rational disciplines. This, Sellars tells us, is patronizing.

Sellars’s second concern about analytical quietism derives from the status of philosophical problems. Quietism

implies that the essential change brought about by philosophy is the standing out of detail within a picture which is grasped as a whole from the start. But of course, to the extent that there is one picture to be grasped reflectively as a whole, the unity of the reflective vision is a task rather than an initial datum. (PSIM 4)

Unity—to the extent that it is forthcoming at all—is a task rather than a datum. On the alternative view, the unity is, we could say, given. Sellars’s fundamental complaint against analytical quietism is that its focus on diagnosis commits it to a subtle and interesting version of the myth of the Given.²⁸

That is, of course, a loaded charge. And while I am prepared to defend it at greater length than I can now, one point is clear. To see Sellars’s ontological monism as requiring of him a naive conception of the unity of science is to accuse him of precisely the view he means to be criticizing. That the world is one—if it is—does not mean that it appears, even in scientific thinking, as one. This pluralism is a central commitment of PSIM.

The quietist we are considering could spell out deVries’s “patchwork” imagery by insisting that there will be puzzles about the interaction between biology and physics, say, such that it is no good question how they should be resolved. This anti-Sellarsian pluralist agrees with Sellars about the essential disunity of intellectual endeavor, but disagrees as to whether this is a problem. Sellars’s disagreement with the pluralist derives not from his Peircean monism then, but from his belief that we should not be satisfied until we produce a Peircean monism.

Thus far, I have attributed to Sellars a relatively weak pluralism, consisting primarily in the thought that when all is said and done, we will need many pictures rather than just one.²⁹ But this invocation of Peirce suggests that a stronger version is the theme of PSIM. It is not given that all intellectually virtuous behavior will cohere, just as it is not given that the principles constitutive of various rational disciplines will be consistent. To say that philosophy has a task is not to say that success is assured.

The emergence of an inconsistency, of an incoherence, gives the philosopher (and the reflective specialist) a task, for it is a regulative ideal of intellectual behavior that such conflicts can be resolved. But I have argued that this regulative ideal need only reflect intellectual curiosity, rather than a transcendental condition of inquiry. Sellars sees no reason why we should be satisfied with a disunified picture but, we can add, there is no reason why we should expect satisfaction either. Of course, Sellars’s talk (elsewhere) of the end of inquiry forecloses this possibility—he seems to assume that a scientifically ultimate picture is a principled possibility. Indeed, this is arguably central to Sellars’s “systematic” thinking.³⁰ But nothing we have seen in PSIM requires this element of Sellars’s system.

The important point is that the appearance of conflict need not result from lack of clarity, or conceptual confusion. It can be evidence of the need for productive and creative work. Such work, Sellars maintains, is the task of philosophy.

All of this can be offered as an interpretation of what Sellars says before he introduces the clash of the images at the end of the prologue. This brings Sellars’s dialectical strategy into focus, for the clash of the images now figures primarily as proof that there is need for productive work. Where one might hope that some problems could dissipate with conceptual clarity, the fact that the scientific and manifest images are alternatives “of essentially the same order of complexity,” each purporting “to be a complete picture of man-in-the-world” (4), entails to Sellars that their clash cannot be dissolved this way. Thus, the purpose of the paper is to vindicate this conception of philosophy with specific reference to the case that ensures that it is required, namely the attempt to understand persons in the scientific image. This, in turn, points us to what I have called the metaethical epilogue. For it is only there that Sellars is convinced of the existence of a clash.

8 Quietism and the science of conceptual thinking

This framing of the paper puts several crucial points in a more satisfying light. For instance, if quietism is the primary target, that explains why Sellars puts instrumentalism aside: he isn’t defending scientific realism, but the existence of the conflict to which scientific realism is his preferred answer. The other major problem I raised earlier concerned the tension between the first and third problem areas. Before turning to the epilogue and, hence, the third problem area, it is worth noting how this tension looks from our new vantage.

Though Sellars claims that “the so-called ‘analytic tradition’ [...] particularly under the influence of the later Wittgenstein” (15) represents the culmination of the perennial tradition, the most “recent” insight he mentions occurred “at the time of Hegel,” when “the essential role of the group” was recognized for any account of conceptual thinking. We learn from Hegel that the “manifest image must, therefore, be construed as containing a conception of itself as a group phenomenon, the group mediating between the individual and the intelligible order.” Nevertheless “any attempt to explain this mediation within the framework of the manifest image was bound to fail” (17), because a more genuinely explanatory view would see this social status “as a matter of evolutionary development” comparable to the evolution of the bee’s dance.³¹ Without Darwinian scientific vocabulary, Sellars thinks the Hegelian will be stuck in the manifest image and “bound to fail.” According to deVries (2017b), this Darwinian move betrays a blindspot.

The tension we discussed above highlights the possibility deVries points at, of an undefeated Hegelian alternative here.³² For Sellars stresses that “the essentially social character of conceptual thinking”—the fundamental post-Kantian thought—“comes clearly to mind when we recognize that there is no thinking apart from common standards of correctness and relevance, which relate what I do think to what anyone ought to think” (PSIM 16–17). Thinking, understood socially, is essentially to be located in a normative space. But normative space is to be joined to the Scientific Image—whatever exactly that means. Given this, it is unclear what problem (his) Hegel faces. It is tempting indeed to imagine that Hegel represents a more thoroughgoing “joining” of the normative to the scientifically explicable.

It complicates this story to note that sociality mostly drops out of Sellars’s discussion of the first problem area. There, he maintains that developments in computing technology, combined with new insight into the complexity of the brain, show Descartes was wrong on the empirical question whether “complex neurophysiological process could be sufficiently analogous to conceptual thinking to be a serious candidate for being what conceptual thinking ‘really is’ ” (30). From the standpoint of neurophysiological complexity, the role of sociality in individuating thoughts can seem vestigial at best. Perhaps this suggests that even Sellars understands a post-Hegelian conception of conceptual thinking as to be dealt with in the “epilogue.” This is an intriguing possibility, but I don’t think it is how Sellars saw matters: he opens the discussion by reiterating that “It is no accident that one learns to think in the very process of learning to speak” (32); and in his discussion of psychology in the scientific image he makes passing reference to evolution: “large-scale variables” like stimulus and response are, he says, “explained by evolutionary theory” (23). Thus, a more likely thumbnail sketch is that the analogy of thought to language makes available an evolutionary theory of the normative standing of episodes that are (in turn) to be identified with complex neurological patterns.³³ But it is anyway notable that the social and normative dimension of this point goes missing in his discussion of the primacy of the scientific image, vis-a-vis conceptual thinking.

What is more immediately relevant than the justice of his discussion of Hegel, or the details of his account of conceptual thinking, is the light this point sheds on the over-arching perspective of the paper. Descartes’s dualism is made plausible by his ignorance of neurophysiological complexity, and Hegel’s idealism by his ignorance of Darwin. The post-Kantian tradition’s discovery of the social was instigated, he suggests, precisely by the failures of Hume’s scientific associationism. Thus, Sellars maintains that conceiving the history of modern philosophy in terms of the clash between the manifest and scientific images helps make sense of what can otherwise seem like very foreign views. The historicity of the images is absolutely central to Sellars’s story.

And the problem with quietism emerges from this historical perspective. It is “current practice” to capture the post-Kantian insight that conceptual thinking is essentially social by comparison to games (PSIM 17). This builds a bridge between the Hegelian who rightly notes the essentially intersubjective nature of conceptual thinking and the Wittgensteinian who reminds us of this fact, as if that reminder settles the matter. The suggestion seems to be that Hegel’s failure here is excusable in a way that twentieth century versions of the thought are not.³⁴ The prospects for a scientific account are very different now, and so Sellars complains that the quietism of Wittgensteinian analytic philosophy shuts down what is otherwise a fertile explanatory project.

There have been times when it would be appropriate for a philosopher to doubt whether a scientific picture of man is forthcoming: the post-Kantian tradition was right to reject Hume’s proto-scientific associationism. But maintaining the epistemic credentials of the manifest image does not require us to downgrade the explanatory pretensions of the scientific image. Thus, Sellars’s objection here is primarily methodological, directed at the blanket rejection of the possibility of scientific understanding.

9 The task of constructing the person

Without “critical reflection on the philosophical enterprise, one is at best but a potential philosopher” (3). If I am right, PSIM is an extended critical reflection on the philosophical enterprise. The thesis is that the central task of philosophy is to construct a vision of “man-in-the-world” that overcomes the inherent tensions in our self-image, starting with the idea that the scientific image emerges out of the manifest image, and purports to improve it. Sellars’s primary target is someone who does not see these tensions as needing to be overcome.

Putting aside the status of humankind in the world, Sellars sees no deep problem with simply endorsing the thought quoted earlier: “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” (EPM §41, 173). Indeed, this just is what his pluralistic deferential methodology requires. Postulational science is intrinsically revisionary: scientists construct novel concepts that will replace ordinary ones by explaining why they work to the extent that they do. So the question is whether we can simply adopt an optimistic view about scientific progress towards a more adequate conception of humankind. The most plausible alternative maintains that it is a mistake to look for lessons about “personhood” in developing science. On this quietistic alternative, science is no source of insight for humanity’s self-conception; to the extent that one takes science to have such implications, one is simply confused.

Sellars’s objection to this quietism is methodological: the epistemic credentials of developing science are basically unimpeachable. But the pluralism I have been stressing does not sit well with the idea that there would be a one-sided victory for the scientific image. Insofar as there are specialists of the manifest image—perennial philosophers—we should expect the upshot of a clash between the scientific and manifest images to respect the credentials of them as well. And indeed, Sellars insists on the need for “stereoscopic vision.”

It goes part way to acknowledging this point if we note that, though Sellars agrees with the eliminativist that it will be “reasonable at some stage to abandon the framework of common sense and use only the framework of theoretical science” (SRI 174),³⁵ he rejects Feyerabend’s more radical insistence that the time for such replacement has arrived. Until neurophysiology gives us the vocabulary to replace manifest color concepts (in their ultimate homogeneity), any adoption of a scientific picture will be premature, incurring unacceptable conceptual loss. In PH he explains that we should reject eliminativist replacement “not because it wouldn’t be a better way, but because the better is the enemy of the best” (97).

As I said, this only goes part way to articulating the need for stereoscopic vision. It shows why we should continue to speak with the manifest image (for now), but it suggests that our need for articulating the manifest image is decidedly secondary. Ultimately the scientific image will drown out the manifest. And sometimes this appears to be Sellars’s position: having maintained that the “basic roadblock is the unity of the person as the subject of conceptual activities” (PH 100), he goes on to insist that “the irreducibility of the ‘I’ within the framework of first person discourse (and, indeed, of ‘you’ and ‘he’ as well) is compatible with the thesis that persons can (in principle) be exhaustively described in terms which involve no reference to such an irreducible logical subject” (PH 101). A similar passage occurs a few pages earlier: “science is making available a more adequate framework of entities which in principle, at least, could serve all the functions, and in particular, the perceptual functions of the framework we actually employ in everyday life” (PH 97).³⁶ But note that these two discussions correspond to the first and the second problem areas. Indeed, in PH there is no mention of ethics.

The thought one gets from these passages is that the manifest image produces explanatory demands, showing us that the scientific image is not yet ready to go.³⁷ But in PSIM at least, Sellars insists on a stronger point, that the scientific image cannot in principle accommodate persons as subjects of norms. And while a different philosopher might infer eliminativism about persons from this, Sellars himself insists that it is “in accordance with the drift of the argument of this chapter” (PSIM 39) to look for another option.

Thus, Sellars sees another task confronting the philosopher: not just the anti-quietistic one of understanding and accommodating new scientific developments, but also articulating our conception of the person. It is reasonable to query this suggestion: deVries, for instance, employing his distinction between the “practical” ontology implicit in our normative practice and the “descriptive” ontology of scientific-explanatory practice, urges that Sellars has given us “no reason to believe that the practical ontology of future science will differ significantly from the practical ontology of the manifest image” (2012, p. 15). Whereas there is every reason to think that the image of the world we get from “final science” will be different from that of contemporary science, deVries presumes that our moral world will, in broad outline, remain in place for our Peircean descendants.

In at least one place in PSIM, however, Sellars seems to insist that our “practical ontology” is itself a matter of controversy. In discussing the development from the original image, he stresses that to abandon the claim that, e.g., trees are persons is not just to abandon a belief.

The truth is, rather, that originally to be a tree was a way of being a person, as, to use a close analogy, to be a woman is a way of being a person, or to be a triangle is a way of being a plane figure. That a woman is a person is not something that one can be said to believe; though there’s enough historical bounce to this example to make it worth-while to use the different example that one cannot be said to believe that a triangle is a plane figure. (PSIM, 10)

I don’t know how much can be made of this passage: Sellars does simply set aside the example of the woman.³⁸ But what gives the example “historical bounce” is that there can be a clash vis-a-vis the status of woman as a person. Such a clash, Sellars maintains, would not be a mere clash of beliefs.

Sellars has prepared us, then, for the thought that the category of the person is a locus of political contestation, where such contestation is not resolved by mere appeal to conceptual clarity. It is not just a clash of beliefs. This is the line of thought I want to read in to the meta-ethical epilogue.

10 Meta-ethical monism?

The meta-ethical epilogue is, alas, undeniably compressed (“proper exposition and defence would require at least the space of this whole volume” (39)). But two thoughts seem clear. First, there is an apparently banal cosmopolitanism, as expressed in this claim about “the community”:

Once the primitive tribe, it is currently (almost) the ‘brotherhood’ of man, and is potentially the ‘republic’ of rational beings (cf. Kant’s ‘Kingdom of Ends’). [...] Thus, to recognize a featherless biped or dolphin or Martian as a person is to think of oneself and it as belonging to a community. (39)

Elsewhere, he puts this point slightly stronger (though see below for the sentence I elide here):

The recognition of each man everywhere as one of us was the extension of tribal loyalty which exploded it into something new. [...] Kant’s conception of each rational being everywhere as one of us is a still more breath-taking point of view which may yet become a live option. (IIO, 210)

This first point is that it is to the essence of the moral point of view to see it as underwritten by an extension of tribal loyalty to the cosmopolitan community of humankind as such.

Second, Sellars invokes his doctrine of community intentions, stressing that a community is those with whom one says “we”:

It follows that to recognize a featherless biped or dolphin or Martian as a person requires that one think thoughts of the form ‘We (one) shall do (or abstain from doing) actions of kind A in circumstances of kind C’. To think thoughts of this kind is not to classify or explain, but to rehearse an intention. (PSIM 39)

This is what supposedly allows the person as a subject of norms to be “joined” to the scientific image. More on that below. But it can read as if all it takes to construct an embracing ethical community is a clear-sighted use of the first-person plural. After all, he says, its non-metaphorical use “is no less basic than the other ‘persons’ in which verbs are conjugated” (ibid). Thus, we need only recognize the implications of our use of “we”; if we would all say “we” together, the Kingdom of Ends would be upon us. Moreover, Sellars’s repeated use of dolphins and Martians (along with the “breathtaking” Kantian view of rational beings as such) can suggest an argument for this claim: where there is rationality, it is conceptually unproblematic to extend the category of persons.

If this were right, then it would be unclear what if any task is left at the end of the paper. Sellars does think there is a task: he writes that “to complete the scientific image we need to enrich it not with more ways of saying what is the case, but with the language of community and individual intentions” (40). Even if we had all the “ways of saying what is the case,” there would still be something more to do. But the only task he mentions is “construing the actions we intend to do and the circumstances in which we intend to do them in scientific terms” (ibid.), and that sounds like merely applying new “ways of saying what is the case” to actions and circumstances, hardly a shocking “enrichment.” If it is true that our Peircean descendants will have different ways of saying what is the case, then undoubtedly they will express their commitments and intentions differently. But this does not entail that adding the “language of community and individual intentions” to the scientific image is a further and substantive step.

Moreover, deVries notes that this project seems to presuppose the implausible thought that “science itself offers the most practical and efficient scheme of classification, even for such tasks as deciding how to design one’s domicile, how to win the heart of a loved one, or what is the best way to spend a hot day in August” (2005, p. 280). Like deVries, I am not sure this is not Sellars’s view.³⁹ Given this, it is not hard to sympathize with deVries’s claim that Sellars has given us “no reason” to expect significant changes in practical ontology.

But there is a problem with this picture too. Danielle Macbeth, for instance, writes that

although we and (say) the rational Martians may converge in our mathematical physics because and insofar as that science has been purged of all the contingencies of our bodies, histories, and cultures, it hardly follows that we and the rational Martians could live together in a single community. Our forms of life may be just too different to sustain meaningful discourse outside of highly abstract mathematics—or even at all. (Macbeth 2017, p. 173)

Macbeth detects great moral danger in the assumption that expanding the community to rational beings as such is (“ontologically”) unproblematic.⁴⁰ Even granting that dolphins or Martians rehearse community intentions amongst themselves, the hurdles standing between recognizing that and constructing an embracing community are more serious than Sellars seems, officially, to recognize. On Sellars’s behalf deVries suggests what we might think of as a moral realism. At least in outline form we can tell from here how a community of all rational beings ought to be.⁴¹ Macbeth suspects this betrays a kind of ethical monism: as if “the contingencies of our bodies, histories, and cultures” were morally irrelevant, insofar as the question of living in a community is concerned.

Characterizing this view as monistic should put us in mind of the debate we saw in the context of science above: there, the thought had been that Sellars’s invocation of the scientific image required him to efface intersubdisciplinary clashes. Here, Macbeth complains that invoking cosmopolitanism effaces intercultural clashes. If this is right, then the same response suggests itself again. Earlier I used talk of intellectual curiosity as the subjective dual to the regulative ideal of unity: so long as there is problematic plurality, intellectual curiosity will push us forward. But putting the point this way works best for the theoretical key in which Sellars’s writings usually stay. I want to conclude with some remarks transposing this thought into the practical domain.

11 Sellars’s meta-ethical pluralism

My leading idea has been that failure to acknowledge the merely regulative status of Sellars’s monism can result in inadvertently attributing to him precisely the view he means to criticize, according to which apparent tensions are a result of mere confusion. The task of philosophy as Sellars sees it is to confront apparent tensions as real, to treat them as genuine problems. On the framing of PSIM I have developed, this points us to the meta-ethical epilogue as the place where, more than anywhere, the tension between the scientific and the manifest images is irresolvable. The question is, in what sense does Sellars see philosophical work to be done here?

On the more traditional reading, on which the goal of the paper is to historically situate and advertise Sellars’s scientific realism, the concluding section looks like a postscript. A conception of persons as subjects of ethical norms cannot be reconstructed scientifically, but this is no great problem: the framework of normative evaluation can be joined to the scientific image of man. But pulling the threads of my argument together now, this concluding section takes on a new importance. The “strictly logical” reason why a reconstruction of persons in the scientific image is impossible (38) is, ultimately, the fact that the conception of a person is the conception of a member (actual or potential) of “an embracing group each member of which thinks of itself as a member of the group” (39). This is what Sellars calls a “community”, and is the immediate lead-in to the cosmopolitan remarks quoted above. Sociality has been a subtheme of the paper all along—Darwinian resources, recall, were supposed to enable us to explain some of this. Though it will produce no new scientific results, Sellars insists that the post-Kantian picture emphasizing the essential sociality of persons, is one the scientific image itself must accommodate.

This is still consistent with the claim that there is no work to be done here. After all, if the project is simply to restate our “community and individual intentions” in “final” Peircean language, that is not something we can do now. As he said in PH, the better is the enemy of the best. But earlier I suggested that Sellars is sensitive to a different potential task: the manifest image itself has work to do, articulating the concept of a person. If the drift of my argument is right, the work within the manifest image is not primarily descriptive-theoretical but ethico-political.

A familiar refrain from those hardy few who read Sellars’s meta-ethical writings is that he is really unspecific about just what ethics demands.⁴² Perhaps this is by design: “in different cultures and in different persons in the same culture, ‘pure [practical] reason’ makes contradictory deliverances” (SE 216). To see through to a conception of morality as a rational discipline, Sellars tells us we must “plumb the depths” of the skepticism motivated by this relativism. Thus, the basic truth of moral relativism is that there are clashes in “practical ontology.” If this is right, we should not be able to tell from here what moral constraints will figure in “the Kingdom of Ends.”⁴³ The moral “realist” who assumes she is in a position to say how that ideal community would be structured is comparable to the quietist we discussed earlier, shirking the real task at hand.

What made the realistic interpretation tempting was the suggestion that it is a matter of grammar whether, e.g., rational martians are “non-metaphorically” within the extension of the embracing “we.” If this were right, the problem would simply be one of conceptual clarity, and it would have been more precise had Sellars said that we are coming to realize that all humans are among “us,” and trying to figure out who else might be. Instead, he depicts the most embracing community as expanding. The sentence I elided from the quotation from IIO in the previous section tempers his otherwise apparently triumphant cosmopolitanism: “It [the recognition of each man everywhere as one of us] has a precarious toehold in the world, and we are usually a far smaller group” (IIO, 211). Arguably Sellars is overly optimistic about the “toehold” a genuine cosmopolitanism has in the world—compare his optimism about scientific convergence. But in neither case does he pretend that it leaves no important work.⁴⁴ Thus, he can take Macbeth’s insight in stride. It is not clear what it would mean to say “we” with the martians. This corresponds to the scientific pluralist’s insistence on an ultimate patchwork. Macbeth leaves the discussion there. On my reading, Sellars again urges that this is insufficient: in such a case, we would need a successor “we.” And as I quoted him (in §9 above), this is not merely a point in principle about fanciful thought-experiments: familiar examples give the point “bounce.” Perhaps philosophy can contribute to the construction of an ever-more embracing community, in a way that requires not just conceptual clarity but ingenuity, the construction of novel concepts.

I understand these “concepts,” like their counterparts in science, as local problem-solving devices rather than utopian would-be features of a guessed-at Peircean epoch. We should not, as our moral realist does, imagine that we can see from here what demands might face us (there). On the contrary, from here there might well be conflicting and contradictory moral demands. The tensions that stand between us and a more embracing community are real, not mere product of irrationality or confusion. As I read him, Sellars urges us to confront these tensions as problems to be solved not confusions to be dissipated.

The closest Sellars ever comes to giving substantive normative ethical advice is when he cites Josiah Royce.

The only frame of mind which can provide direct support for moral commitment is what Josiah Royce called Loyalty, and what Christians call Love (Charity). This is a commitment deeper than any commitment to abstract principle. It is this commitment to the well-being of our fellow man which stands to the justification of moral principles as the purpose of acquiring the ability to explain and predict stands to the justification of scientific theories. (SE 231, his emphasis)⁴⁵

This is the conclusion Sellars comes to by plumbing the depths of moral skepticism. He uncovers not a moral truth to which we ought all to be committed, but instead a commitment “deeper” than that: loyalty/love/charity is not an abstract principle. He goes on, “This concern for others is a precious thing, the foundation for which is laid in early childhood” (SE 231); it is the product of a good upbringing.

He compares this concern to “the purpose of acquiring the ability to explain and predict,” which is the characterization that led me earlier (§6) to talk of intellectual curiosity. Of course, there is no guarantee that two intellectually curious subjects will agree about everything. It is just a regulative ideal of the intellectually curious that a genuine disagreement can be resolved. Writ large, this is the task of philosophy we encountered in the prologue to PSIM.

Sellars never denies what seems evident from his invocation of relativism, namely that Royce’s loyalty can underwrite different, and conflicting, first-order moral commitments. Such conflicts, we can now say on Sellars’s behalf, need not be the result of mere confusion, simple moral error. The unity of an embracing ethical community is regulative in just the same sense as is the unity of intellectual vision towards which the Sellarsian epistemologist struggles. Indeed, when at the end of PSIM he connects the possibility of an embracing community with the eventual overcoming of the dualism of manifest and scientific images, it is just possible that Sellars sees these two projects as two sides of the same coin.