“The world we live in is the world of sense data; but the world we talk about is the world of physical objects.”
(Wittgenstein 2016: 82)

1. Introduction

1Clarence Irving Lewis (1883-1964) and Roy Wood Sellars (1880-1973) are important figures of twentieth-century American philosophy. Lewis, the founder of axiomatic modal logic (and teacher of Willard Van Orman Quine), stood for a – partially Kantian inspired – “conceptualistic” pragmatism. Sellars, on the other hand, was the leading representative of “critical” realism (and father of Wilfrid S. Sellars). Both partook of what W. H. Werkmeister called “the naturalistic trend in American philosophy” (Werkmeister 1949: 577). However, Lewis’s pragmatist and Sellars’s realist approach were divided over several issues, the most prominent being Sellars’s rejection of analytic philosophy. In Sellars’s view, “so-called analytic philosophy […] did not seem to me very creative in either epistemology or ontology. American addiction to it and disregard of its own momentum struck me as a form of neo-colonialism” (Sellars 1969: 5; emphases omitted). Lewis would surely not have agreed with this assessment. Moreover, Lewis and Sellars defended divergent accounts regarding the issue of perception. Whereas Lewis located the very object of perception in the realm of sensorily given data, Sellars held that a thing is that which is perceived. This is very roughly stated, but it ultimately hits the nail on the head.

2The aim of the present paper is to comparatively reconstruct Lewis’s and Sellars’s respective conceptions of the object of perception. In systematic terms, an answer should be delivered as to the question what we talk about when we talk about this, for example, being blue. As a result, it will turn out that Sellars’s view has more to recommend it because it provides a more compelling framework for a comprehensive explanation of perception.

3I will proceed as follows. In section 2, Lewis’s conceptualistic pragmatism and its implications regarding the object of perception is going to be outlined. Section 3 is devoted to Sellars’s critique of Lewis’s approach. In section 4, Sellars’s positive account of the issue will be discussed at some length. Section 5 discusses exegetic and interpretative problems regarding Lewis’s theory of the given and the latter’s debatable status as a “myth.” Section 6 concludes the paper with a few comments on broader philosophical consequences.

2. Lewis on “the Given” and “the Pragmatic a priori”

5In order to better understand Lewis’s conceptualistic pragmatism, it is mandatory to clarify two points: first, what exactly is meant by “the given”; and second, how does the pragmatic a priori apply to it?

6Let us begin with Lewis’s account of the given. The given is, according to him, characterized by three features: it is (a) ineffable, (b) foundational regarding objects, and (c) essentially veridical. The first of these three features implies that what is immediately presented by the senses “remains untouched and unaltered, however it is construed in thought” (ibid.: 53). Thus the given forms the unshakeable ground of all our perceptual knowledge. It is for this reason that Lewis, referring to a paper by J. Loewenberg (see Loewenberg 1927), speaks of “preanalytic data” (Lewis 1929: 54) in this very context. Accordingly, the given element in perception figures as “an ultimate epistemological category” (ibid.). However, what exactly should be considered as the given element in perception? This question brings us to the second feature: On Lewis’s account, the given lies at the very bottom of objecthood. More precisely speaking, objects are seen by him as being constructed out of sensory elements. Therefore, the given not only has epistemological but also ontological impact and thus pertains to what Lewis defines as the “problem of metaphysics,” namely “’the problem of the categories’” (ibid.: 10). In a similar vein as did Bertrand Russell in his Our Knowledge of the External World (1914), Lewis assumes that sensorily given elements, such as particular patches of color or particular tastes, provide the foundation for a subsequent interpretative process. He points out: “An object such as an apple is never given; between the real apple in all its complexity and this fragmentary presentation [that it tastes sweet], lies that interval which only interpretation can bridge.” (Lewis 1929: 120). On the other hand, and this pertains to feature number three, the given element in perception is essentially free from error, i.e. it guarantees certainty. For Lewis, “the given itself is never misrepresentative; always it is true of the real, however partial” (ibid.: 179). If perceptual errors occur, then they are due to the interpretative (judgemental) part of cognition which, in turn, is always merely probable and thus fallible (see ibid., ch. X).

7Which brings us to Lewis’s conception of the pragmatic a priori. Here, too, three features are worth mentioning: the pragmatic a priori is (a) the source of conceptual order, (b) historically dynamic, (c) essentially connected to practical needs. Feature number one brings in the Kantian element of Lewis’s approach. Though not at all identical with Kant’s original contribution (see Boyer 1958), conceptualistic pragmatism proceeds from a similar central idea. Lewis writes:

In experience, mind is confronted with the chaos of the given. In the interest of adaptation and control, it seeks to discover within or impose upon this chaos some kind of stable order, through which distinguishable items may become the signs of future possibilities. Those patterns of distinction and relationship […] are our concepts. These must be determined in advance of the particular experience to which they apply in order that what is given may have meaning. (Lewis 1929: 230)

9It should be noted that the position so far outlined significantly diverges from the “classical” pragmatist approach. To be sure, Lewis’s reliance on practical needs of adaptation clearly stands in the tradition of Peirce, James, and Dewey. But it must be seen that Lewis’s is a conceptualistic pragmatism. In sharp contrast especially to James and Dewey, Lewis focuses on questions from traditional epistemology, dating back at least to Kant. His Mind and the World Order is subtitled “Outline of a Theory of Knowledge”; and the interpretative function of concepts regarding the given plays the crucial role in his account (see ibid.: ix). Classical pragmatism, in his view, is too trustful in “the validity of rather superficial and capricious attitudes” (ibid.: 267), such as personal motives and desires. For Lewis, the actual epistemic situation must be clarified in the first place. He therefore declares: “We must all be pragmatists, but pragmatists in the end, not in the beginning.” (Ibid.: 267).

11So much for Lewis’s conceptualistic pragmatism. It should have become clear that, according to Lewis, perception is a complex interpretative and at the same time practice-oriented process. Furthermore, it has turned out that the given element in perceptual knowledge is essential for Lewis’s whole approach. It still remains to be determined, though, how the status of the given is to be assessed. The following section should prove instructive in this regard.

3. Sellars’s Critique of Lewis’s Approach

12Sellars’s critique of Lewis’s approach is very interesting because the two thinkers shared several common concerns. Sellars himself was no doubt aware of that fact. Thus, for example, in his paper “Critical Realism and the Independence of the Object,” published in 1937, he frankly states:

Perhaps I can not better express my sense of the logical keenness of Lewis’s analysis than by saying that, if a fatal weakness in my analysis of the categorical distinctions of cognition could be shown, I would regard his position as the most tenable alternative. It strikes me as much superior to the common versions of positivism, logical or otherwise. (Sellars 1937: 548)

13Already in his seminal The Philosophy of Physical Realism from 1932, Sellars had made a similar, albeit more polemic, statement in that direction. There, he writes: “In many ways his [Lewis’s] approach is similar to mine; but there remains a theoretic subjectivism which is obscured by language and yet reveals itself in passage after passage.” (Sellars 1932: 186). How is this to be understood? A first step in answering that question is provided by Sellars’s claim that Lewis’s approach is “an attempted blending of Berkeley and Kant in the setting furnished by pragmatism” (Sellars 1937: 545). Sellars himself was neither a Berkeleyan, nor a Kantian, nor a pragmatist. All of these currents were for him expressions of “subjectivism.” Accordingly, it might be expected that his critique turns out to be quite fundamental. Let us see if this is really the case.

14Confining ourselves to the exposition in The Philosophy of Physical Realism (the work in which Sellars expresses his critique most clearly) three principal objections against Lewis’s approach can be identified. The first of these objections, let us call it the representationalist objection, says that the roles of both sensory presentations and active attitudes are overrated by Lewis. In Sellars’s words:

The Berkeleian note [in Lewis’s approach] is dominant. Experience is the ultimate term. The logical problem is to conceptualize presentations properly. Don’t bother with questions of what lies beyond presentations, for they are essentially meaningless. In the terms of Carnap these are Scheinprobleme. To all this, I reply that it is independent and enduring things which we suppose ourselves to perceive, and not presentations. […] When I ordinarily assert that I am biting into a peach, I do not mean that I am biting into a presentation, a sensory datum. […] In knowing we are spectators and not primarily agents. (Sellars 1932: 146-7; emphasis in original)

15What Sellars is implying here is that, while sensory data are presented, things are re-presented; hence the label “representationalist objection.” However, his insinuation that for Lewis we are “primarily agents,” is exaggerated. We have already seen that Lewis’s pragmatism is conceptualistic in the first place and thus not to be confused with classical pragmatism. Accordingly, Lewis can (at least partially) be defended against Sellars in this connection.

16So what about the second objection? Here the central claim is that the place of sensorily given elements in perception is falsely located by Lewis. Let us call this the anti-positivist objection. In Sellars’s words, it reads as follows:

The independence of reality – its independence of the knowing mind – means [for Lewis] the givenness of the given, the truth of “If-then” propositions, and the accrual of something in the future which cannot now be determined. In short, the rejection of co-existential transcendence is complete. […] In opposition to this positivism, the critical realist argues that knowing is, in intention, the interpretation of physical things as these in common measure appear, or are manifested, in sensory presentation and that, in critical knowledge, we achieve a complex content which is regarded as revealing the independent, physical thing. (Ibid.: 152; emphasis in original)

18Sellars’s third objection, which might be called the semantic objection, is tightly connected with this point. In a nutshell, it says that in Lewis’s account the generation of stable meanings is fundamentally obscured. Or as Sellars puts it himself:

[P]erceiving is always thicker than the awareness of sense-data. Sense-data are enveloped in a directed complex of meaning in the perceptual experience; and this fact is not hard to understand when we realize that perceiving involves organic response. As I see it, Lewis realized this supplementation; but his empiricism led him to make the sense-data the objects of interpretation rather than the guides of an interpretation of affirmed external things with properties. And this stand made it impossible for him to make a physical organism the basis of repeated meanings. (Sellars 1932: 202)

19To better understand Sellars’s point, let us come back to the example of this being blue. To be sure, Lewis would not deny that it is, for instance, a ball of which we actually say it is blue (see, for example, Lewis 1929: 57-9). However, as Sellars correctly observes, by making sense-data the objects of conceptual interpretation Lewis becomes confronted with the problem of how to specify the range of application of such a term as “blue.” If, in our example, the ball itself is the object of perception, then a stable meaning can be established to the effect that it can repeatedly be said of the same ball that it is blue. If, on the other hand, sense-data are conceived of as the object of perception, then we have no sameness anymore. Sense-data are always in flux and thus not stabilizable. The predicate “blue” would apply to as many different objects as there were different sense-data whenever we see the ball. Consequently, we would stand in need of a “private language,” since each individual person has its own sensational data. I dare say that Lewis’s account ultimately breaks down at this very point.

20To summarize, Sellars’s critique of Lewis’s approach is indeed fundamental and in more than one respect convincing. Regarding the object of perception, the two thinkers stood in radical opposition to each other. For Lewis, the sensorily given is the object of interpretation and, thus, of perception. For Sellars, in contrast, it is affirmed external things that are interpreted and thus perceived. Sense-data, according to Sellars, merely serve as guides of interpretation. They are by no means foundational (or terminal). Accordingly, the question of what we talk about when we talk about this, for example, being blue is answered in two completely different ways. What needs to be done next is to figure out whether Sellars’s positive account has more to offer than Lewis’s one.

4. Sellars’s Positive Account

21To begin with, Sellars defended the critical realist point of view already as early as 1908 (see Sellars 1908). In 1916, he published a book explicitly titled Critical Realism. In that book he still held that “we perceive percepts […] and not physical things” (Sellars 1916: viii). Four years later, Sellars co-published the Essays in Critical Realism (see Drake et al. 1920). In his own contribution to that volume, “Knowledge and Its Categories,” he made it clear that he now thought that “[w]e mean independent objects and we interpret these objects in terms of ideas” (Sellars 1920: 194). The principal implication of this change of view becomes particularly clear from the following statement in Sellars’s 1929 “A Re-Examination of Critical Realism”: “[D]irectness and mediation are not contradictory in the theory of knowledge.” (Sellars 1929: 440). Accordingly, perceptual knowledge is now conceived of as being both direct and mediated. As with the representatives of American “neo-realism” (Perry, Holt, and others), Sellars assumes that the perceiving individual is directed to mind-independent things from the very first. On the other hand, in contrast to the neo-realists, Sellars’s realism is critical insofar as it reflects upon the conceptual presuppositions of the perceptual situation. In his view, concepts and categories serve to disclose the characters of “the actual physical things to which we are reacting and adjusting ourselves” (Sellars 1920: 191). Just as for Lewis, concepts and categories are for Sellars means of interpretation. However, unlike Lewis, Sellars assumes that it is physical things (and not sensory presentations) that get conceptually interpreted. At any rate, concepts for Sellars have a representational role to play: they are the essentially mind-dependent vehicles through which the perceptual process is mediated.

22In order to adequately understand Sellars’s positive (mature) account, it is necessary to consider his rejection of the idea of a cognitive relation. In his article “Is there a Cognitive Relation?,” published in 1912, Sellars claims that “[t]he subject-object relation is a dogma which has been an article of faith” (Sellars 1912: 229). And in his contribution to the Essays in Critical Realism he emphatically states: “The critical realist holds that knowledge is a function of the knower rather than a peculiar, real relation between the knower and the known.” (Sellars 1920: 206; emphasis in original). Let us call this “the functional idea.” The gist of Sellars’s advocacy of that idea is the attempt to bring together two apparently divergent approaches in the philosophy of perception: his already mentioned representationalism and what nowadays is called enactivism (see, for example, Noë 2004). The enactivist aspect is provided by the insight that perception is an intention-based activity of an organism in its environment. Therefore, perceptual knowledge does not have the status of an abstract, static relation. Rather, it should be conceived of as the organism’s inter-action with its experienced environment. This is where Sellars, again, agrees with Lewis. However, as already outlined, the sensorily given element in perception is, for Sellars, not the very object of perception, but rather that which guides us in affirming and conceptually interpreting the actual physical thing. Accordingly, “the critical realist regards perceiving as a non-inferential selection and interpretation of a thing as an object. This selection is a guided intent, guided, in point of fact, by sense-presentations.” (Sellars 1937: 546; emphasis in original).

23It is important to note that Sellars’s functional idea is a more or less immediate reaction to both traditional (Lockean) representationalism and contemporary (American and British) neo-realism. Regarding the first, Sellars employs the enactivist aspect of his account in order to prevent the “Berkeleyan impasse.” It is commonly known that Berkeley had objected against Locke that by relying on mediating ideas, a direct knowledge of external things is made impossible (see, for example, Berkeley 1710, §20). Sellars agrees with this critique, but he also sees an alternative to Berkeley’s resulting plea for idealism (or “subjectivism”). As a matter of fact, Sellars attempts to account for the directness (and thus non-inferentiality) of perceptual knowledge by dynamizing the representational relation. Thus, he points out:

[T]he critical realist would not speak of a relation between external fact and the idea mediating knowledge of it. Rather would he regard the act of knowing as a descriptive interpretation of an external state of affairs now made the object of a cognitive intent. In such a situation, representation is a use made of an idea which is presumably such that it corresponds to the object of the intent. (Sellars 1932: 128; emphasis in original)

25Moreover, Sellars rejects the neo-realists’ assumption that perception be characterized as the literal presence in experience of the objects perceived. In Sellars’s view, this variant of “pan-objectivism” (Sellars 1920: 188) is doomed to fail (not at least because it is unable to explain perceptual errors). His alternative to it amounts to a sophisticated form of representationalism, or what he himself calls “epistemological dualism” (ibid.). According to epistemological dualism, “knowledge of objects is mediated by ideas which are in some sense distinct from the objects of knowledge” (ibid.: 190; emphasis in original). What exactly does this mean? It is the notion of content, which is decisive in this regard. For Sellars, objects are perceived directly. However, the process of perceiving is mediated by cognitive elements that – by being effected by the active mental being – constitute the content. Thus, in an article from 1918, Sellars explains:

[T]he very gist of the difference between neo-realism and critical realism is that the knowledge-content, or object of awareness, is […] numerically distinct from the existent or object of knowledge. The only justification of the phrase epistemological dualism resides in this fact. The existent acknowledged, but not given, is the object of knowledge, while the mental content given is the material and content of knowledge, but not the object. (Sellars 1918: 507)

26In his 1927 essay on “Current Realism in Great Britain and United States,” Sellars specifies this point as follows: “We make things objects, we mean, select, affirm them in a specific and definite way; and yet these objects do not literally enter our consciousness. Rather are they interpreted in terms of meanings and characters which stand out in our perception and in our thought of them.” (Sellars 1927: 507). In short, in perception objects do not become part of the brain, but are rather deciphered on the basis of representational contents.

27To sum it up: Sellars’s functional idea is focused on the active attitudes of an organism in its environment. Since the organism forms part of this environment, its being directed to “external” things is, in fact, an ingredient of the perceptual situation from the start. It is in the process of perceiving that representational contents turn out to be means of “making” things objects. Sensations play a functional role insofar as they serve as indicators of something non-sensational, namely physical things. Though not the very objects of perception, sensations help us “in our pressing need to adapt ourselves to our environment” (Sellars 1920: 191). They, in other words, guide us in generating perceptual contents which, in turn, serve to disclose the characters of physical things that are the subject of conceptual interpretation. The things themselves remain unaltered by this whole process. Our direct knowledge of them is an “achievement” (Sellars 1932: 79) being documented by our successfully referring to them. Nevertheless, it is the functionally interpreted contents that pave the way to this achievement.

28On the whole, then, perception according to Sellars is a form of “triangulation,” so that we have

1. the thing existing independent of the mind;

2. the sensation which guides perception;

3. the perceptual content which furnishes the basis for the conceptual interpretation of the perceived object.

5. Exegetic and Interpretative Problems Regarding Lewis’s Theory of the Given

29In light of the previous two sections, one might ask: Did Sellars’s critical realism “win out over” Lewis’s conceptualistic pragmatism after all? In historical terms, it is highly interesting to see that Roy Wood Sellars’s son Wilfrid, in his critique of the “myth of the given” (see W. Sellars 1956), argued along exactly the same lines as his father. The account of Lewis was one of Wilfrid’s main targets too (see O’Shea 2021). Already in a paper from 1954, he had put the issue thus:

[W]hen Jones sees a chair, although his perceptual experience is founded on, guided, and controlled by his sensations, there is nothing in the nature of aboutness or reference which requires us to say that his experience is primarily about the sensations, and only about the chair in some complicated or derived sense of “about.” His perception is “mediated by” the sensations, but his perception is not about the sensations. (W. Sellars 1954: 20; emphases in original)

30This is nothing but Roy Wood’s point of view, so that it can be claimed that the latter was the precursor of the critique of the myth of the given (see, in this connection, also Gironi 2017).

31However, the criticism of Lewis as a proponent of the myth of the given has not remained uncontested. For example, Peter Olen asserts that “R. W. Sellars radically misinterpreted conceptual pragmatism” (Olen 2015: 157). In a similar vein, Cheryl Misak categorically states: “One of the gravest injustices in the history of modern philosophy is the relegation of Lewis to the dustbin of foundationalism and what Sellars later called the Myth of the Given.” (Misak 2014: 334). And Tomasz Zarębski holds that “Lewis’s conception of the given is immune to Sellarsian criticism of the ‘myth of the given’” (Zarębski 2017: 200). Assessments such as these are surely motivated by passages like the following from Lewis’s Mind and the World Order:

It is indeed the thick experience of the world of things, not the thin given of immediacy, which constitutes the datum for philosophic reflection. We do not see patches of color, but trees and houses; we hear, not indescribable sound, but voices and violins. […] The given, as here conceived, is certainly an abstraction. (Lewis 1929: 54)

32To be sure, this very passage provides some exegetical support against seeing Lewis as an advocate (or “victim”) of the myth of the given. However, in systematic terms, there still remains a fundamental discrepancy in Lewis’s account: How could it be that “a world of sensa with nothing behind them” (Lewis 1929: 31; see above, sect. 2) turns out to be “the world of things?” Of course, there are “things” in Lewis’s ontology. But their status is that of constructions. And here the trouble begins. Take again our example of the blue ball: According to Lewis, our talk of its being blue would rely to particular sense-data, and their interpretation by concepts would imply a constructive maneuver. This is plainly at odds with the actual mechanism of perception (as well as with respective insights from Gestalt psychology). In (both Roy Wood and Wilfrid) Sellars’s view, on the other hand, conceptual interpretation amounts to the sensorily guided and perceptually grounded representation of particular states of affairs. It is things about which we talk when we talk about this, for example, being blue.

[T]here is an absolute certainty of the empirical which has been recognized – the immediate apprehension of the given. Such direct awareness is not indubitable knowledge of an object, but the content of it is an absolutely given fact. This immediate presentation is our confrontation with reality and is requisite to the distinction of particular empirical truths from falsehood. Immediate qualia constitute the ultimate denotation in experience of our concepts, and the specific character of the given plays its indispensable part in any verification. (Lewis 1929: 310)

6. Some Broader Philosophical Consequences

37Regarding these and other controversial issues of the Lewis-Sellars connection, it might be concluded that further exploration is worth pursuing. On a larger scale, the relationship between American pragmatism and American critical realism needs to be reconsidered. Beyond the shadow of a doubt, the two currents shared many aspects; but, as we have seen, there were also significant differences. Hopefully, the present paper will contribute to extended research on this revealing episode in the history of American philosophy.