how to be a kantian and a naturalist about human knowledge
Post on 03-Jun-2018
225 Views
Preview:
TRANSCRIPT
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
1/33
JOURNAL OF PHILOSOPHICAL RESEARCH
VOLUME 36, 2011
I
HOW TO BE A KANTIAN ANDA NATURALIST ABOUT
HUMAN KNOWLEDGE: SELLARSS MIDDLE WAY
JAMES R. OSHEA
UNIVERSITY COLLEGE DUBLIN
ABSTRACT:The contention in this paper is that central to
Sellarss famous attempt to fuse the manifest image and the
scientific image of the human being in the world was an at-
tempt to marry a particularly strong form of scientific naturalism
with various modified Kantian a priori principles about the unity
of the self and the structure of human knowledge. The modified
Kantian aspects of Sellarss view have been emphasized by
current left wing Sellarsians, while the scientific naturalist
aspects have been championed by right wing Sellarsians, the
latter including William Rottschaefers constructive criticisms
of my own reconciling interpretation of Sellars. In this paper I
focus first on how (1) Sellarss Kantian conception of the neces-
sary a priori unity of the thinking self does not conflict with his
ideal scientific naturalist conception of persons as bundles
or pluralities of scientifically postulated processes. This then
prepares the way for a more comprehensive discussion of how
(2) Sellarss modified Kantian account of the substantive a priori
principles that make possible any conceptualized knowledge of
a world does not conflict with his simultaneous demand for an
ideal scientific explanation and evolutionary account of those
same conceptual capacities. Sellarss own attempted via media
synthesiswhat I call his Kantian scientific naturalism
merits another look from both the left and the right.
I. INTRODUCTION
t is no surprise that Kantians have often been hostile to comprehensive
forms of philosophical naturalism, whether in theoretical or practical philosophy.
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
2/33
328 JAMES R. OSHEA
For it is characteristic of the Kantian to argue that there are substantive conceptual
truths about human cognition and action that can be uncovered by the philosopher
independently of the results of particular empirical observations and of ongoing
natural scientific inquiry. Suppose we take naturalism, as I shall take it here, to be
the particularly controversial thesis that all truths about the human being and the
world are in principle exhaustivelyexplainableby the methods of ongoing scientific
inquiry. Admittedly this familiar thesis requires much clarification, each step of
which would likely be as controversial as the thesis itself; and furthermore, this
is only one among many ways of understanding naturalism. With naturalism so
understood, however, it is certainly not hard to understand the traditional hostility
between the Kantians, with their supposedly substantive a priori conceptual claims
about the self and any possible knowable world, and the philosophical naturalists
with their patient and ever knowledge-increasing scientific fallibilism.
In a recent book, Wilfrid Sellars: Naturalism with a Normative Turn(2007), I
have argued that Wilfrid Sellars was both a Kantian anda thoroughgoing scientific
naturalist in something like the traditionally mutually hostile senses described
above. We might put it in crude terms, in relation to recent philosophical history, as
follows. In my view, Sellars held that, on the one hand, the sort of updated Kantian
and later Wittgensteinian views defended by Peter Strawson, for instanceor in
another tradition, the transcendental phenomenology of Husserlwere right in
spirit and in fact true in central respects. But they were conceived in a way that
rendered them incompatible with the ideally all-comprehensive scientific natural-
ism that Sellars also sought to defend. On the other hand, the broadly scientific
naturalist outlooks of positivists such as Rudolf Carnap, Herbert Feigl, and Carl
Hempel, and including the sophisticated successor naturalism of W. V. O. Quine,
were also right in spirit and correct in central respects. But on Sellarss view both
positivism and Quinean naturalism ultimately failed to recognize (among other
things) that genuinely Kantian and Wittgensteinian views about human knowledge,
intentionality, and agency are in fact essential to achievingin a subtle roundabout
wayan ideally complete scientific explanation of the nature of the human being
in the world. In his Autobiographical Reflections Sellars interestingly remarks
upon his roundabout way of defending a naturalistic materialism as follows:
Feigl and I shared a common purpose: to formulate a scientifically oriented,
naturalistic realism which would save the appearances. . . . We hit it off
immediately, although the seriousness with which I took such ideas as causal
necessity, synthetic a prioriknowledge, intentionality, ethical intuitionism,
the problem of universals, etc., etc., must have jarred his empiricist sensi-
bilities. Even when I made it clear that my aim was to map these structures
into a naturalistic, even a materialistic, metaphysics, he felt, as many have,
that I was going around Robin Hoods barn [i.e., taking an unnecessarily
long route to the same endJOS]. (AR, 289290)
I hope to show that understanding what Sellars retained from his lonely walk
around Robin Hoods barnaccompanied by few if any other scientific naturalists,
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
3/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 329
past or presentis crucial to understanding the unique nature of his naturalistic
metaphysics.
In my view, then, Sellarss famous synoptic vision and stereoscopic fusion
of what he called the manifest image of man-in-the-world and the scientific
image of man-in-the-world (PSIM) was an attempt to happily marry, by means
of some deft and persistent matchmaking, the two ostensibly antagonistic figures
of the Kantian and the scientific naturalist.
In a judicious and probing examination, Why Wilfrid Sellars is Right (and
Right-Wing): Thinking with OShea on Sellars, Norms, and Nature, William Rott-
schaefer (2011) makes a strong case for rejecting central aspects of my interpreta-
tion of Sellarss synoptic vision, in particular my understanding of the latter as an
attempt to marry enduring Kantian and later-Wittgensteinian conceptual analyses
(in a sense of conceptual and analytic that is not restricted to narrowlyanalytic
truths) with an all-comprehensive scientific-explanatory naturalism. Rottschaefer,
by contrast, ably defends the so-called right-wing interpretation of what Sellars
certainly did embrace, namely, the primacy of the scientific image (PSIM, 32).
The right-wing interpretation typically understands Sellarss synoptic vision in such
a way that it involves the ultimate scientific replacement rather than the preservation
of the Kantian and later Wittgensteinian aspects that I emphasize in my reading of
Sellars. Rottschaefer argues in impressive detail that not only is this a more accu-
rate reading of Sellars, it also harmonizes with various currently widely embraced
naturalistic philosophical views that reject Kantian style a priorism in philosophy
in general. Thus, Rottschaefer argues, there is reason to resist OSheas attempt
to turn Sellars transcendentally to the left (33);1and on Sellarss account of the
structure of knowledge, in particular, he contends that OSheas reading runs into
massive problems, since . . . Sellars rejects Kant, arguing that we can know things
in themselves because of the successes of the project of scientific knowing that has
as its ideal the completed scientific image (36).
The Right-Wing of Rottschaefers title refers, of course, to the distinction
frequently informally made by Sellarsians between left wing Sellars-influenced
philosophers, such as (in different ways) Robert Brandom and John McDowell,
and right wing Sellars-influenced philosophers such as (in different ways) Ruth
Millikan and Paul Churchland. The left/right distinction, whatever its merits or
demerits, can roughly be understood in terms of the two sides of Sellarss philoso-
phy that I have highlighted above. The left wing emphasizes the Kantian and later
Wittgensteinian (in fact, also Hegelian) aspects of Sellarss views while rejecting
Sellarss alleged scientism and reductionism. The left has no problem with
garden varieties of scientific realism about theoretical entities such as electrons,
etc. (as opposed, for example, to instrumentalism). But the left firmly rejects
the all-comprehensive scientific naturalist thesis as I articulated it above and as
expressed in Sellarss notorious scientia mensura claim in Empiricism and the
Philosophy of Mind (1956) that 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
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
4/33
330 JAMES R. OSHEA
not that it is not (EPM 42; see also Sellars SSIS, 396397, for Sellarss descrip-
tion of himself as an Extreme scientific realist even with respect to explaining
the ultimate nature ofpersons).
The right, by contrast, fully embraces Sellarss all-comprehensive scientific
naturalism, but as exemplified by the case of Ruth Millikan, likewise tends to ar-
gue that there is ultimately a crack in Sellarss system due to his treatment of
the nature of linguistic rules and the relation of these to conceptual roles and thus
to intentionality (Millikan 2005, 78). (Note, however, that Rottschaefer himself
does not argue that there is any crack or tension in Sellarss overall system.) In
particular the right wing tends to be suspicious of the Kantian a priori and Witt-
gensteinian conventionalist elements in Sellarss view, ifthese are simply taken
and left at face value, and are not ultimately fully scientifically naturalized, as
the right wing argues is frequently the case with Sellarss left-wing admirers. As
Rottschaefer correctly suggests, right-wing Sellarsians such as Millikan and many
other contemporary naturalistic philosophers are engaged in developing philosophi-
cal insights based on evolutionary biology and other domains of scientific inquiry
in the attempt, inter alia, to naturalistically explain the normative dimensions of
human language, thought, perception, and action, in the context of wider theories
of human and other animal cognition and behavior. Rottschaefer argues that while
Sellars initially articulates those normative dimensions in terms of insights from
Kant and the later Wittgenstein within the manifest image, Sellars himself envi-
sioned and would thus embrace the sorts of right-wing naturalistic theories that
seek ultimately to better explain those normative dimensions in scientific terms.
All of these left vs. right tangles in the wake of Sellarss philosophy seem to
suggest that it is difficult if not impossible to bring together the two sides of his
philosophy into one coherent picturewhich after all was Sellarss own central
goal (see PSIM passim). On Rottschaefers sophisticated right-wing view, Sellarss
scientific naturalism is in the endall that we need to make sense of Sellarss ideal
synoptic vision of the nature of the human-being-in-the-worldand this is a good
thing, Rottschaefer further contends, given the promise of philosophical naturalism
in light of recent developments.
In what follows I take Rottschaefers welcome and insightful criticisms as an
opportunity to continue the defense of my reconciling middle way (some might
say, Polyanna) interpretation, according to which Sellarss actual position is such
that, if successful, it enables us to sustain the centralpositiveclaims of boththe left
and right wings. This will require further clarifying the nature of Sellarss novel
attempt to bring about the unlikely marriage described above, as well as responding
to some more specific objections that are helpfully raised by Rottschaefer.
II. SELLARSS SYNOPTIC VISION AND HIS
NORMATIVE FUNCTIONALISM ABOUT THOUGHTS
I begin with one important issue that has wide ramifications. Here, as in several
other places in his article, I think that Rottschaefer appropriately raises some pivotal
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
5/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 331
questions concerning the interpretation of Sellarss philosophy. In his section II on
Sellarss Functionalism, Rottschaefer comments as follows on the momentous
character of Sellarss claim concerning the theoretical character of our knowledge
of both our own and other persons inner thought episodes, which in Empiricism
and the Philosophy of Mind Sellars had famously defended in terms of the myth
of genius Jones:
From the point of view of introspection and philosophical reflection, the
manifest image of thought, broadly construed, with its propositional attitudes
of belief and desire what has come to be called folk psychologymay
appear to be a given. But if Sellars functional role account of intentionality
is correct that entire framework is a theoretical one replaceable by a more
adequate scientifically based framework. On the Right-wing account, not
only is Eddingtons ordinary table replaceable by a scientific table, but so also
is the manifest image thinker seated at that table. As we shall see, OSheas
take on the matter is different. (295296)
I think I am right to have a different take on this matter, for I think that there is a
subtle, and quite common, yet crucial mistake embodied in the central comparison
made in the passage from Rottschaefer above. Roughly put, the faulty comparison
is this: just as (i) the entire manifest image ontology of ordinary, colored physi-
cal objects is ultimately to be replacedby a more adequate scientifically based
framework in which, to take a notable example, colors in some sense end up inthe perceiver and notin whatever scientifically conceived processes replace the
manifest table in the ultimate scientific ontology; in the same way, (ii) with respect
to the manifest image of thought. . . with its propositional attitudes of belief and
desire, and including the manifest imagethinker, that entire framework is a
theoretical one replaceable by a more adequate scientifically based framework
(italics added).
The problem with Rottschaefers comparison here, I contend, is that on Sellarss
view (i) is true, but (ii)in the most important sense to be explained belowis
false. And I hold that this is essential to understanding Sellarss attempted synopticvision of the nature of persons and their thoughts within an ideal scientific image
of man-in-the-world. There are crucial respects in which the fate of the manifest
image conception of persons and their thoughts and intentionsand therefore also
the communally shared normative standards and ought-to-be rules that make any
logicalspace of reasons possible, to use Sellarss famous phrase (EPM VIII,
36)differs from the fate of the manifest image conception of colored physical
objects. The fate of so-called folk psychology is radically different, on Sellarss
view of the matter, from the fate of folk physics.
In particular, there is a primary sense in which, with crucial insights fromperennial philosophers from Kant and Hegel to Wittgenstein and Strawson, and
culminating in Sellarss own normative functionalist or conceptual role account
of meaning and intentionality, we have already achievedan adequate conception
of what it is to be a thought and what it is to be a thinker in general, and we have
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
6/33
332 JAMES R. OSHEA
done so without any reliance on the Myth of the Given. We have it now, in the
philosophically refined manifest image, in a way that is notthe case with respect to
our (strictly speakingfalse) manifest image conception of what it is to be a colored
physical object out there in space. And again, this difference is not due to any
alleged immediate awareness of the ultimate nature of the self and its thoughts as
inner mythical Givens.
For interests sake and to avoid repetition, I will begin by exploring this claim
in relation to some less well-known remarks of Sellarss on the topic. Eventually
this will bring us to Sellarss conception of what it is to be a thinker, and thereby
back to general questions concerning the Kantian dimensions of Sellarss scien-
tific naturalism.
Compare the passage quoted above from Rottschaefer with the following
comments from Sellars, which are taken from a Question and Answer session fol-
lowing the delivery at the University of Notre Dame in 1969 of what became his
important essay on The Structure of Knowledge (SK). Fortunately the original
audio recordings have recently been transcribed and made available as Wilfrid
Sellars: Notre Dame Lectures 19691986the Bootleg Version (WSNDL). I
do not claim that there is any surprise or smoking gun in these remarks, or that
they contain any claim that is not also clearly made by Sellars in his published
writings. But the remarks do help to illustrate the point I want to emphasize.
(I have added the original audience question, taken from the audio version, to
Amarals transcript.)
Audience question put to Sellars: [Y]ou liken the classical philosophical
postulation of thought episodes, as explanatory of our propensities to speak,
to what some day, perhaps, will be the last word that neurophysiology has
to say on why in fact we really do have these propensities to speak. No mat-
ter how nicely you polish it, it seems as if you make the manifest image a
nice tryyou know, you have to begin with some error, so well use this
oneand that ultimately thats going to be just phased out, you know, its
going to be long gone.
Sellars reply: I assure you that thats false. I think that human beings are
always going to think and know that they think. . . . I understand thinking
to be fundamentally a functional notion, governed by correctnesses and
rules and validity; the most that the scientific image can do here is to give
us some notion, in Aristotles sense, for the material cause of thinking
but theformal cause of thinking is surely afunctionand this is a function
which exists now and which we think of well now, we understand it well. I
think that what science can add here is trivial. For me, to say that thought is
neurophysiological is like saying English contains noises like and, or,
but, and so on. The actual function of thinking is to be found in the rulesthat govern inferences and the rules that govern the conceptual structures of
language. . . . We have an adequate notion of what thinking is in its formal
cause, the most the science can do, if I can use this terminology, is to give
us the material cause and as I said that is really quite unexciting as far as Im
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
7/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 333
concerned and that is why I think that as far as human living and the person
is concerned, the manifest image contains the formal truth and that science
is going to give an account of the material substructure. (WSNDL, 221222)
The views expressed in offhand fashion in these remarks are central to Sellarss
ultimate solution to the problem of integrating theoretical science with the
framework of sophisticated common sense into one comprehensive synoptic vi-
sion (PSIM, 19); and they are especially important for understanding his insis-
tence that, as I see it, the manifest image is not overwhelmed in the synthesis
(PSIM, 9). To explore this further we first need to see how the case of physical
objects (i.e., (ii) in the comparison above) differs from the case of thoughts and
persons gestured at in these remarks; and then in the next section we shall need
to explain in a bit more detail Sellarss Kantian conception of persons and their
thoughts, which is the formal truth about human beings that Sellars clearly
takes to survive all the way into the ideal, final, Peircean scientific image of
man-in-the-world.
In the same Question and Answer session Sellars takes it to be crucial to clearly
distinguish cases (i) vs. (ii), as discussed above. Specifically with regard to the
idea that the scientific image will replace the manifest image, Sellars comments
that the most I have ever said is that in its descriptive aspects the scientific image
could in principle replace the contentualaspects of the manifest image (and he
here again repeats that we are not going to replace the notion of thinking, all weare going to do is have a better understanding as to what specifically it is that is
doing those functions) (WSNDL, 223). It is at this point that he remarks that it
is very important not to suppose that sensation and thought are going to be handled
in the same way (ibid.). The reason for this is that sensation is quite a different
sort of thing and it is, in a way, a contentnot a function, like thoughtthat is
going to remain in the world picture regardless; and it will turn out that the locus
of color and sound . . . is not in the physical world, but in ourselves (ibid.). In the
case of color and the other sensible qualities:
It is not just [that the scientific image is] going to throw a light on it becauseI think it literally would involve a replaceability in the material aspects. I
think that putting it in Kantian language . . . the world of commonsense solid
colored objects is a phenomenal world in Kants sense of the term; it is an
appearance of scientific reality. . . . Let me emphasize that I have had rela-
tively little to say about values, and standards and norms and obligations . . .
because, putting it very crudely, I am talking here about the is of the world
and my whole theory of ethics hasnt been touched on at all. . . . When I talk
about the in principle replaceability of the manifest image by the scientific
image, I do so with respect to the contentof the world, its material and not
with respect to those forms which concern the normative, the obligatory,
the correct, the incorrect, the valuable. . . . I think Kant is essentially right,
not only in many of the things he said in a theory of knowledge but also in
ethics. (WSNDL, 223224)
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
8/33
334 JAMES R. OSHEA
Following this passage, on which I shall comment below, Sellars proceeds to discuss
animal cognition as opposed to human conceptual cognition (a topic which I shall
also briefly mention), before finally summing up as follows:
The formal components of the manifest imagethat remains. The formal
features of the manifest imagewhich are the important features, the features
that concern the normative, the evaluative, the matter of personal intention
and so onthese are going to remain in the scientific image. What is going
to change is the contentualaspect. (WSNDL, 226, slightly re-punctuated)
Having laid out Sellarss remarks, let us now compare the manifest image claim
(i) that Smith sees the pink ice cube over there, with the manifest image claim (ii)
that Smith is thinking that it is raining.
In the terminology he used above, Sellars holds that there is a particularly prob-lematic material content or contentual aspect involved in (i), an ineliminable
sensory content that Sellars typically describes as thecube of pink, i.e., a certain
volume of perceived color. What happens to the iceis a different story, though an
equally interesting one, because being ice is a causal property rather than aproper
sensible (pink) or common sensible (cube) quality, in the Aristotelian terms Sellars
uses. It is the proper and common sensibles that notoriously raise what Sellars called
the sensorium-body problem, a problem that is closely related to what others call
the problem of qualia. On familiar, if highly complex and controversial grounds
(e.g., concerning hallucinations, etc.), Sellars and his mythical genius Jones arguethat the experienced cube of pink, needs to be relocated, as it were (i.e., its true
location is recognized), by being ontologically reconceived to be a sensory state
of the manifest image person who is Smith, rather than being an intrinsic content-
character of the physical ice cube over there. Sellars then leaves Jones and the mani-
fest image behind and argues that in the ideal scientific image there will need to be
afurtherreconception of Smiths state of sensing a-cube-of-pink-ly. In this case
the same sensible content (the volume of pink) is now ontologically reconceived as
a bottom-level and in some sense non-physical (i.e., physical1-but-not-physical
2,
in Sellarss technical terminology) absolute process (a sensing) that is takingplace in whatever Smiths visual cortex turns out to be, causally interacting with
whatever (physical2) quarkings, electronings, and other absolute processes will
ultimately be revealed to make up Smiths living body.2
In relation to claim (i), then, Sellars holds that the core contentthat is involved
in seeing (or merely vividly seeming to see) a pink ice cubenamely, the sensed
cube of pinkneeds to be preserved throughout a long journey of reconception as
to what ontological kind or category of item (quality of object? state of perceiver?
absolute pure process?) that same contentual cube of pink will ultimately turn
out to be. Hence the special importance of the rejection of the Myth of the Givenin this domain, since the Myth would allege, inter alia, that what kind of thing a
sensory appearance isits ontological nature and location, as it wereis Given
immediately with the conscious sensing of it.3Claim (i), then, when viewed in
light of the regulative ideal of the conjectured final scientific image, turns out to be
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
9/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 335
strictly speaking false (because there really is nopink ice cube of the kind conceived
in the manifest image). In another sense, of course, the manifest image claim (i)
is intelligibly and in some sense approximately true, in that the appearances
embodied in the manifest image conception will turn out to be explainedin terms
of the ultimately correct recategorization in the final scientific successor theory. (A
similar process occurs in scientific theory-replacementin general.)
What about claim (ii), the manifest image claim that Smith is thinking that it
is raining? Roughly put, this claim will be true, on genius Joness initial model, if
Smith is either candidly saying out loud, or has a short term propensity to say out
loud, It is raining, i.e., in a way that accords with the normatively rule-governed
conceptual role (functional role, inferential role) that gives that sentence its meaning
in English. And then on the theory of genius Jones, which starts from that rule-
governed verbal model, claim (ii) will be true if there was somethinggoing on in
Smiths soul or brain or heart (or whatever, as far as genius Jones, still within the
manifest image, is concerned) that is playing a functional role relevantly similar to
that played by the utterance-type, It is raining (more strictly, an it is raining),
but this time occurring silently in Mentalese, as Sellars calls it, i.e., in something
likea language of thought.
What does the ongoing scientific image then add to Joness initial theoretical
enrichment, an enrichment that takes place within the manifest image (i.e., Joness
postulation is of new inner statesof the same old manifest imagepersons)? Scientific
inquiry adds a series of more detailed, perhaps in some cases revolutionary theories
concerning how the conceptual roles or semantical rule-governed functions that,
on Sellarss view, are definitive of both inner and outer conceptual thinking, are
in fact realized in whatever will turn out to have been the real material vehicles
of our thoughts all along. We soon forget about hearts and souls, for example, and
concentrate on brains. Then we have ongoing scientific and philosophical inquiries
into the nature of the representational medium in the brainquestions that were
wisely left wide open by Sellars (as correctly noted by Dennett on Sellars in his
1987 book, The Intentional Stance). Perhaps, for example, the brain does not carry
information in the sort of symbol-processingway that Sellars might have projected
for his Mentalese, but rather in something more like the parallel distributed pro-
cessing way anticipated by connectionists, including eliminativists such as Paul
Churchland (who was Sellarss Ph.D. student). There are certainly complex issues
involved here, but the remarks above (along with his other writings, I believe) sug-
gest that Sellars would not accept the eliminativist conclusions that Churchland
attempts to draw from the projected neurophysiological facts.
Let us suppose, finally, that the Peircean ideal scientific image of man-in-the-
world is completed. What might we anticipate to be the conception of conceptual
thinking and intentional action in that final vision, on Sellarss view? Whereas with
respect to claim (i) and the manifest image conception of the pink ice cube or the
red apple, to recall the passage from Rottschaefer quoted earlier, it is true that
that entire framework is . . . replaceable by a more adequate scientifically based
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
10/33
336 JAMES R. OSHEA
framework, there is an important sense in which the same is not true with respect
to claim (ii). Whatever very different things people might be thinking aboutin that
future timei.e., however radically differently conceived the inner and outer objects
of their thoughts might be4and whatever they might have discovered to be the
neurophysiological or behavioral-cum-environmental vehiclesof their thoughts,
the manifest image account of what it is to be a thinking that p, on Sellarss view,
remains true and is not replaced. Conceptually contentful thoughts are inner and
outer inferential role-players, where the relevant roles (in the case of logic-using
animals) are determined by the socially maintained, rule-governed linguistic norms
that constitute a given logical space of reasons (EPM, 36).
The semantical rules, in Sellarss sense, and hence the contents of thinkable
thoughts, will of course change as new conceptual frameworks replace old ones.
But the ontology of thoughts, in the sense of accounting for what it is to be a
thought and what it is to be a thinker, remains the same account across changing
conceptual frameworks, and will remain the same in the ideal scientific image of
thinkers-in-the-world. Changes in the conceptual roles that constitute the contents
of thinkable thoughts, however radical such changes might be in the course of
scientific theory-succession, do not involve the replacement of the conception
of thoughts as normatively rule-governed inferential role-players. The process
of conceptual change in that sense must be clearly distinguished from the sort of
recategorization of the nature of objects that Sellars thinks is involved in scien-
tific theory-succession, whether the latter concerns the problem of the nature and
location of color in particular, or the ultimate categorial structure of physical
reality in general.
So in relation to the sensible qualities and claim (i), as Sellars remarked above,
we in some sense have a contentthat is going to remain in the world picture regard-
less (WSNDL, 211), but the ontological categorization of this preserved content
does indeed undergo radical replacement from the manifest image conception to the
ideal scientific image. In the case of thoughts and claim (ii), however, the opposite
is the case: the realizing contentual aspect or material vehicles of thought are
what undergo radical reconceptualization and replacement. But the correct ontologi-
cal categorization of what it is to be a thought, and of what it is to be a thinker, is
already at hand in the manifest image, and in particular in that image as it continues
to be refined by what Sellars calls the perennial philosophy (PSIM passim). The
crucial insights in this domain, as interpreted by Sellars, were: (1) Kants formal,
broadly logical conception of concepts as rules, together with his corresponding
formal accounts of what it is to be a thinking selfin aknowable world; (2) Ryle
and the later Wittgensteins so-called meaning as use conceptions of meaning and
thinking, with the focus on norm-governed linguistic behavior (communal aspects
of which were anticipated by Hegel and Peirce); and building on those insights,
(3) Sellarss normative-functionalist philosophy of mind as described above, as
well as his modified Kantian conceptions of self and knowledge to be explored in
the sections to follow.
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
11/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 337
It follows from the above that Rottschaefers thesis that, according to Sellarss
way of integrating the two global images, the claims of the manifest image are
false (7), is problematic with respect to certain definitive formal features of
persons, their thoughts, and their intentions. I shall say more about this difficult
issue in the next section. Rottschaefer also has many important things to say about
goal-directed normativity in nature during the course of his discussion of Sellarss
functionalism, and I shall briefly return to some of those issues later.
III. SELLARSS MODIFIED KANTIAN VIEW OF
PERSONS AS LOGICALLY UNITARY THINKERS
In the substantial middle sections III and IV of his article, Rottschaefer rightly
raises some pressing questions concerning the Kantian aspects of my interpreta-tion of Sellarss views on the structure of empirical knowledge. Here I will begin
by taking up a difficult question left hanging from the previous section: namely,
what is it to be a humanpersonon Sellarss overall stereoscopic integration of the
manifest and scientific images of man-in-the-world?
In one sense Sellarss various writings make clear that he wants to defend the
view that the scientific image will ultimately reveal the real truth about persons to
be as follows (see in particular Sellarss SSIS, i.e., his lengthy Reply to Cornman).
While in the philosophically refined manifest image we conceive a person to be the
single logical subject of its various thoughts and feelings, the ideal scientific imagewill reveal that a person is reallya complex system (a plurality, a series, a group)
of whatever logical subjects turn out to be explanatorily basic in that scientific
image. On Sellarss way of anticipating the latter, a person would be a bundle
of absolute processes (FMPP III.125). It might seem, then, that the conception
of persons in the manifest image is, after all, strictly speakingfalse, in the way
that the manifest image conception of the colored table is revealed to be strictly
speaking false, though ultimately scientifically explainable. Obviously this issue
bears directly on my interpretation of the comparison between claims (i) and (ii)
discussed in the previous section.Ifthe conception of the self within the philosophically refined manifest image
did consist in the positive ontological assertion that a person isa certain (conten-
tual) kindof thing, in particular, that the self isa single (substantial, persisting,
identical) objecti.e., a logical subject in the sense of being a basic subject of
propertiesthen the truth of the ideal scientific image of the person as a basic plu-
rality of items wouldentail the falsity of the manifest image conception of persons.
But Sellarss modified Kantian conception of persons within the manifest image
precisely does not have that consequence. The purpose of Sellarss endorsement
of central aspects of Kants arguments in the Paralogismssection of the Critiqueof Pure Reasonis precisely to hammer home the point that the idea of a unitary
logical subject occurs in two different senses when one is considering the idea of
a thinkeras a unified subject of its thoughts, as opposed to when one is considering
the nature of the thinker as a kind of object.
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
12/33
338 JAMES R. OSHEA
In the manifest image, Sellars holds, our concept of a person is that of a
system of capacities pertaining to the various modes of thinking (MP, 239), and
such a system necessarilyrepresents itself as a unitary thinking self: the I think
as the single logical subject of its various successive thoughts. (The crucial role
of those particular thoughts that are intentions, in Sellarss account of persons and
normativity, will be mentioned below.) Without attempting to reconstruct Sellarss
detailed analysis of Kants Paralogisms in his striking essay, . . . this I or he or
it (the thing) which thinks . . . (I; cf. MP), it is clear that Sellars endorses Kants
central arguments in the Paralogisms, while rejecting certain wider aspects of
Kants view in which those arguments are embedded. In the end this turns out to
have implications not only for Sellarss synoptic vision of the nature of persons,
but also for his account of the structure of human knowledge.
What Sellars primarily rejects are certain mistaken aspects of, and resulting
limitations imposed by, Kants conception of the material(for Kant, phenomenal)
world in Newtonian Space and Time. In particular, Sellars argues that we must
add to Kants analysis Sellarss own scientific-theoretical, analogical conception
of space and time as pertaining to ultimately real things in themselves (see SM
chapters one and two). In this way, Sellars also rejects what he characterizes as
Kants agnosticism about the ultimate noumenal reality of persons or thinking
selves. The incorrect aspects of Kants view force him into an empirical dualism
about the empirical self, as opposed to what Sellars takes to be the broadly correct
Strawsonian conception of persons (within the manifest image) as materially em-
bodied subjects possessing intellectual capacities. But what he thinks the correct
aspects of Kants analysis of persons as thinking selves deliberately and coherently
leaves open is the possibility that, while a thinker necessarily represents itself as
a single logical subject of its plurality of thoughts, such a thinker could turn out,
as a thing in itself, to be an ultimateplurality of basic logical subjects (Sellars
cites A363, for example). For Sellars, but not for Kant (given the incorrect aspects
of his view), this plurality can coherently turn out to be a materialplurality of
scientifically basic items. The following passage illustrates what Sellars wants to
endorse from Kant and then put to his own non-Kantian use:
[Kant] is suggesting that the logical identity of the I through Time, which
is an analytic implication of the knowledge of oneself as thinking different
thoughts at different times, is compatible with the idea that these thoughts
are successive states of different ultimate subjects. Compare the materialist
who argues that the thoughts which make up the history of an I are states of
systems of material particles which are constantly losing old and gaining
new constituents.
Thus although I do not represent my successive thoughts as successive
states of a series of different subjects of attributes, and do not need to doso in order to know my logical identity through the period in which these
thoughts occur, a being with suitable cognitive powers [JOS: God for Kant,
the ideal scientific image for Sellars] might know me to be such a series.
This insight, however, would not require him to say that my knowledge of
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
13/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 339
myself as logically identical through the period of time in question is an
illusion, but only that the logical identity of the I as I represent itis not an
adequate conceptualization of the nature of our thinking being. (I, 2728)
To begin with the final sentence, Sellars is notsaying that an ideally adequate
knowledge of what the self is in itself would reveal that the manifest thinking self,
like the colored manifest table, is a strictly speaking false but approximately true
appearance in the same sense as such manifest physical objectsas the colored
table or the pink ice cube discussed earlier. Kants analytic conception of the
knowledge of oneself as thinking different thoughts is what discloses asparalo-
gistic (i.e., a fallacious equivocation) any attempt to regard the logically necessary
unity of any thinking self as requiring or enabling the knowledge of the self as any
kind of objectat all. The correct aspects of Kants analysis of the necessary unity
of any thinking self, that is, does not involve the framework conception of a kind of
objectthat could even be coherently thought of as undergoing scientific correction
by being categorially reconceived in the ideal scientific image in the way that the
cube of pink does get categorially relocated (so to speak), thereby revealing
the manifest image conception of pink physical ice cubes to be, strictly speaking,
false.5What makes the scientific image conception of persons more adequateis
that it doesreveal the ultimate nature of our thinking being (namely, as an ulti-
mate plurality of items), while the necessary analytic unity of any diachronically
thinking self, which Sellars embraces from Kants analysis, does not even involve
taking a shot at such an object of knowledge.
So, in the ideal scientific image, will there be (1) any logically unitary thinking
selveswho are engaged in the business of fully adequately conceptually represent-
ing the world, and (2) doing so solely in terms of the representation of scientifically
conceived objects and processes? Yes, there necessarily will be, given: (1) the
necessary conditions that Sellars argues (following Kant and Wittgenstein) must
be satisfied for there to be such a thing as thelogically structured conceptual rep-
resentationof a world at all; and (2) Sellarss conceptionabout which, I think,
sophisticated right-wingers such as Rottschaefer are correctof what the ideallycomprehensive scientific image of man-in-the-world ultimately will represent
to be the fundamental contents or basic objects (or pure processes) in the world
(cf. note 4 above).
The crucial point is that the ideal scientific-naturalistic representation of all
objects and processes in the world involved in (2) does not conflict with (1), i.e.,
with the perennial Kantian conception of the self in relation to any conceptually
knowable world, a conception that I think Sellars clearly adopted and adapted while
subtracting what he takes to be Kants detachable errors. The fact that Kant was
ultimately wrong about the nature and status of the represented objects in the world,on Sellarss view, does not invalidate central aspects of Kants conception of what
anyconceptual representation of a worldeven in ideal scientific termsrequires.
It is ultimately a non-trivial analytictruth, on Sellarss reading of Kant (and in this
respect, according to Sellarss own view), that any such conceptual representation of
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
14/33
340 JAMES R. OSHEA
a world is possible onlyfor a being that represents itself as having the sort of logi-
cal identity that is embodied in the representation, I think. And that conceptual
representation, which takes place within the manifest image, is the key aspect of
the manifest image that is preserved(not overwhelmed) within the ideal scientific
image. That is, it is preserved ifwe can assume that there are any conceptual thinkers
representing a world in that ideal scientific imagewhich of course, we can, given
that an image, on Sellarss view, is precisely a conceptual frameworkthat represents
the world as being a certain way. Furthermore that conceptual representation, i.e.,
the logically unitary representation I think in relation to a manifold of thoughts,
is on Sellarss view not falsebut rather true, although it is also not an adequate rep-
resentation of the nature of the self since it remains agnostic about the underlying
noumenal plurality of scientifically conceived processes that, on Sellarss view,
such thinking selves ultimately really are. These are deep and controversial Kantian
waters, yes. But those are the waters in which Sellars swam, from start to finish.
At any rate, I hope that the above helps to clarify my understanding of the
passage from Sellars with which I ended my book (and for this opportunity I am
thankful for Rottschaefers criticisms):
The heart of the matter is the fact that the irreducibility of the I within the
framework of first person discourse . . . is compatible with the thesis that
persons can (in principle) be exhaustively described in terms which involve
no reference to such an irreducible subject. For the description will mention
rather than usethe framework to which these logical subjects belong. Kant
saw that the transcendental unity of apperception is a form of experience
rather than a disclosure of ultimate reality. If persons are really multiplici-
ties of logical subjects, then unless these multiplicities used the conceptual
framework of persons there would be no persons. But the idea that persons
really are such multiplicities does not require that concepts pertaining to
persons be analysable intoconcepts pertaining to sets of logical subjects.
Persons may really be bundles, but the concept of a person is not the con-
cept of a bundle. (PH, 101)
IV. SELLARSS MODIFIED KANTIAN VIEW OF
THE NECESSARY STRUCTURE OF HUMAN KNOWLEDGE
The preceding discussion of Sellarss modified Kantian conception of the thinking
self is connected with his theory of knowledge in ways that Sellars explored in a
variety of writings throughout his career. Here I will begin by continuing with Sel-
larss article, . . . this I or he or it (the thing) which thinks . . . The correct aspects
of Kants view of the thinking self and the transcendental unity of apperception,
as Sellars explains, begins with an
unrestrictedprinciple in the philosophy of mind, which transcends the dis-
tinction between the noumenal and the phenomenal self, to the effect that
an I thinks of a manifold
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
15/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 341
is not to be confused with
an I has a manifold of thoughts.
. . . The ways in which many thinkings constitute one thinking are the formsof thought, e.g., the categories. The recognition of the radical difference
between categorial forms and matter-of-factual relationships is the pons
asinorumof the Critical Philosophy. (I, 78)
Before continuing the analysis of the correct aspects of Kants view here, let me
briefly digress on an important point. For the final sentence in this passage reflects
another way in which Sellars criticizes Kants agnosticism about things in
themselves, one that Rottschaefer correctly describes in his section IV, but that he
incorrectly takes to be incompatible with my reading of Sellars and Kant.
Ultimately, in order show how language hooks up with the world, Sellars, onmy readingwhich in this respect is congenial to the right-wingers but notto the
left-wingersthinks that we must provide a naturalistic theory of representation
(or picturing) that explains how all of the Kantian and Wittgensteinian rule-
governed epistemic activity I am describing really succeeds in mappingmatter-
of-factual occurrences among things in themselves. John McDowell from the
left, for example, rejects this sort of sideways on attempt to naturalistically
explain (in the sense of naturalism I am using here) how our representations
ultimately succeed in corresponding to and correctly representing objects in the
world. On the importance both philosophically and to Sellars (but notto Kant)of this sort of comprehensive attempt to naturalistically explainhow our mental
representations hook up to the world, I am with Rottschaefer, Millikan, and the
other right-wingers, contrary to what Rottschaefer suggests. (Furthermore, I am
with the right wing in holding that, according to Sellars, even the content and
efficacy of those community-based conceptualized intentions that constitute ratio-
nal normativity itself (i.e., within a logicalspace of reasons) must be susceptible
to an ultimate scientific explanationin naturalistic terms. But here, too, Sellarss
account of those we-intentions that generate rational normative principles is
given a Kantianformaltwist, as we go around Robin Hoods barn, that is noteliminatedin the final synoptic vision of man-in-the-world. I shall comment
upon this point in the final section.)
What apparently distinguishes my reading of Sellarss middle way from both
the left-wingers and the right-wingers is that I think the whole point of the subtle
Kantian aspects of Sellarss unifying stereoscopic vision is to argue that Sellarss
bold (and non-Kantian) attempt to comprehensively naturalistically explain the
ultimate nature of our intentional and epistemic activities within the ideal scientific
image is consistentwiththe enduring formal truth, as we saw him call it earlier,
of the Kantian and equally Sellarsian conception of thinkers and thoughts (andrational normativity itself) that I have been trying to articulate here and in the book.6
To return to the passage from Sellars quoted above, for our present purposes
I will continue to take its central Kantian claim concerning the necessary unity
of the I think in all conceptual thinking both to be true (it is of course not
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
16/33
342 JAMES R. OSHEA
uncontroversial) and to be endorsed by Sellars. Sellars immediately continues that
passage as follows:
In epistemology, which, as concerned with good thinking in its variousmodes, is a fortioriconcerned with thinking as such, this general principle
becomes the epistemic principle that any true content of thought, e.g., that
Socrates is wise, must, in principle, be an element in a certain kind of larger
context, e.g.,
an I thinks the true thought of a world in which Socrates is wise [italics
added].
Roughly, the form of empirical knowledge is: an I thinking (however sche-
matically) the thought of a temporal system of states of affairs to which any
actual state of affairs belongs.Thus, in the Transcendental Analytic, the above unrestricted principle
about thinking provides the clue to the form of the phenomenal world. This
world is a presented world, and Space, Time, and the Categories are its
forms. (I, 910)
On my interpretation, Sellars agrees with the Kantian epistemic principle that
any true content of thought must be an element in a certain kind of larger con-
text, which is roughly, an I thinking (however schematically) the thought of a
temporal system of states of affairs to which any actual state of affairs belongs. I
think Sellars, in this respect similar to Strawson, regards this as a truth about thepossibility of any discursive representation of an objective state of affairs. Where
he particularly disagrees with Kant is with respect to complex issues concerning
realism and idealism, in particular with respect to the possibilities for the sort of
world-in-itself that we might come justifiably to represent.
The issue is especially complex because Sellars agreeswith Kant that the world
as conceived in the manifest image is phenomenal or transcendentally ideal,
but not because of the Kantian epistemic principle described above. Sellars makes
clear in the second chapter of his Science and Metaphysicsthat he rejects most of
Kants ownreasons for being a transcendental idealist, and accepts only a versionof Sellarss own argument from the categorial relocation of colored physical ob-
jects when we move from the manifest image to the scientific image, as described
earlier (SM, II 5879). This fits nicely the reading I gave above in section II of
Sellarss views on the crucial differences between (i) the ultimate falsity (and for
Sellars, transcendental ideality) of the manifest image claims concerning manifest
pink ice cubes and other colored physical objects, as opposed to (ii) the perennial
formal truth of manifest image claims concerning thinkers and their thoughts.
Sellars thinks Kant was right to treat the manifest image of common sense objects
as phenomenal, but only for Sellarss reasons concerning the ultimate falsity ofthe framework-conception of the objectsin the manifest image.
Sellars also thinks Kant was right with respect to the necessary holistic and
systematic conceptual requirements on having any true thought about a temporal
state of affairs. But Kant also failed to see how the rule-governed holistic framework
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
17/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 343
of self-conscious human knowledge can (and must) itselfbe given a sideways on
scientific explanation in naturalistic terms, in terms of things in themselves. The
latter, boldly naturalistic explanatory task, which targets the nature of human cogni-
tion and action themselves as objectsa task which the left-wing Sellarsians regard
as misguided, but about the importance of which to Sellars I think Rottschaefer
and I agreewill itself be an instanceof the sort of conceptual representation of a
world, by self-conscious human thinkers (i.e., logically unitary subjects of thought),
that Sellars thinks Kant correctly analyzed.
Within Sellarss own view, what modified form do the holistic Kantian epis-
temic principles take concerning the necessary conditions for the possibility of the
conceptual representation of a world, which I claim Sellars endorsed for conceptual
representings in anyimage, whether manifest or scientific, while rejecting certain
of Kants limitations upon them? In his Autobiographical Reflections Sellars
remarks as follows on his early encounters with Kant at Oxford:
Kant wasnt attempting to prove that in addition toknowing facts about im-
mediate experience, one alsoknew facts about physical objects, but rather
that a skeptic who grants knowledge of even the simplest fact about an event
occurring in Time is, in effect, granting knowledge of the existence of nature
as a whole. I was sure he was right. But his own question haunted me. How
is itpossiblethat knowledge has this structure? (AR, 285)
Sellars then comments that it wasnt until much later that I came to see that thesolution of the puzzle lay in correctly locating the conceptual order in the causal
order and correctly interpreting the causality involved (AR, 285286). What the
latter is designed to naturalistically support is the former, Kantian epistemological
formal insight (based on a perennial philosophical analysis) that knowledge neces-
sarilyhas a certain systematic holistic structure: namely, that granting knowledge
of even the simplest fact . . . is, in effect, granting knowledge of the existence of
nature as a whole.
Rottschaefer carefully considers the attempt in my book to bring to the surface
a variety of Sellarss own modified versions of the Kantian epistemic principlesdiscussed above in several of his works, in particular with respect to my use of Sel-
larss Some Remarks on Kants Theory of Experience (KTE), More on Givenness
and Explanatory Coherence (MGEC), and the Structure of Knowledge (SK).
Here is a characteristic snippet from SK (see Rottschaefer, 305): We have to be in
this framework[of epistemic meta-principles] to be thinking and perceiving beings
at all; and furthermore,
the exploration of these principles is but part and parcel of the task of
explicating the conceptof a rational animal or, in VB [verbal behaviorist]
terms, of a language-using organism whose language is aboutthe world inwhich it is used. It is only in the light of this larger task that the problem
of the status of epistemic principles reveals its true meaning. (SK, 4546,
first italics added)
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
18/33
344 JAMES R. OSHEA
Or as Sellars puts it in the more explicitly Kantian context of KTE, with Kants
views modified by Sellarss linguistic turn:
Transcendental linguistics . . . is not limited to the epistemic functioning ofhistorical languages in the actual world. It attempts to delineate the general
features that would be common to the epistemic functioning of any language
in any possible world. As I once put it, epistemology, in the new way of
words, is the theory of what it is to be a language that is about a world in
which it is used. Far from being an accidental excrescence, Kants tran-
scendental psychology is the heart of his system. He, too, seeks the general
features any conceptual system must have in order to generate knowledge
of a world to which it belongs. (KTE, 41)
What Kants formal philosophical analysis correctlyrevealed, according to Sellars,
is that any conceptual system must, if there are to be any such things as concep-
tualized thoughts of a sensible world of temporally locatable events at all, have a
structure according to which there are warrantedmeta-principles pertaining to the
correctness of language entry perceptions and material inference principles in
general (see, e.g., KTE, 37; and SM: chapter four, 61, and chapter five, 30).7
Where we must go beyond Kants agnosticism about things in themselves, ac-
cording to Sellars, is in relation to the ongoing scientific naturalist project conceived
as extending to both of the following tasks: (a) Explaining how rule-governed
linguistic behaviour and the (rational and non-rational) animal cognition of objectsin an environment actually works, as for example including a naturalistic theory
of mental and linguistic representation. And ultimately, (b) providing a natural-
istic answer to the question How did we get in to the framework? constituted
by conceptual thinking and hence by epistemic meta-principles in the first place,
which Sellars says [p]resumably . . . has a causal answer consisting in a special
application of evolutionary theory to the emergence of beings capable of conceptu-
ally representing the world of which they have come to be a part (MGEC, 79).
Upon closer examination I still find, despite Rottschaefers critical comments
on my reading of the Kantian-formal dimension of Sellarss theory of knowledge,that each of the articles KTE, SK, and MGEC strongly confirms the reading of Sel-
larss complex relationship to Kant that I have given in this paper and in the book.
I will try to be brief here, although Rottschaefers analysis is admirably detailed.
Many of Rottschaefers criticisms here and throughout his article have to do
with my alleged view that, on Sellarss view, nature as pictured in the scientific
image is bereft of goal-directed normativity. In fact, however, I hold that Sellars
stressed the importance to present and future science of ideal evolutionary explana-
tory accounts of both kinds of goal-directed animal representational systems, as
Sellars called them in his important late article, Mental Events (MEV, passim):namely, the rational or Aristotelian logic-using kind of animal representational
system, involving conceptual representationproper(and hence, a logical space
of reasons); and the Humean, non-logic-using yet quasi-propositional kind of
animal cognition of objects. Sellars thought that in the ideal scientific image there
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
19/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 345
would be complex evolutionary and behavioural-learning explanations(rather than
conceptual analyses) of bothkinds of goal-directed processes in nature. On this I
am at one with Rottschaefer, and I have tried to explain above how I think Sellarss
modified Kantian conceptual analyses are consistent with the idea of such an ideally
complete naturalistic explanation. It should be pointed out, however, that for Sellars
the goal-directedness of non-logic-using Humean animal representational systems
is cashed out in terms of the wider context of natural selection in a way that does
not require the additional considerations that pertain to persons and logic-using
Aristotelian animal representational systems. Accounting for the overall coher-
ence and justificatory structure of the latter systems requires not only a complex
causal account of how language and thought evolved by natural selection out of the
former representational systems, but also the independent philosophical account
of the necessary structure of conceptual knowledge that Sellars is providing on the
backs of his colleagues from the perennial philosophical tradition.
So while such evolutionary explanations in the ideal scientific image would
explain how it is possible that human beings came to possess the sort of conceptual-
linguistic abilities that we do possess, I do not claim that Sellars regarded the
justification of his modified Kantian epistemic principles, which pertain to logi-
calrepresentational systems, as being provided by those projected ideal scientific
explanations (see Rottschaefer on the split personality of my Sellars, 302).
Sellarss account of the ultimate justification of those fundamental Kantian epistemic
principlesin Sellarss view, principles that concern (as Rottschaefer recounts)
the necessary elements in a conceptual framework which defines what it is to
be a finite knower in a world one never made (MGEC, IV, 73; and similarly in
KTE and SK, III)comes from theperennial conceptual analysisside rather than
the complementary ongoing scientific explanatoryside of his subtle stereoscopic
view. Such justification is ultimately due to updated versions of what Sellars took
to be right about Kants philosophical analysis of the conceptually necessary
conditions for any conceptual representation of an empirically mind-independent
world. Sellars explicitly endorses such an account when he refers, for example, to
the pure pragmatics or transcendental logic of empirical knowledge as such,
where the former, of course, refers to one way in which he framed his own overall
philosophical framework (TTC, 51). But he carefully disagrees with Kant in all
those key places where he argues that Kant shut off the possibility of Sellarss own
scientific naturalism about things as they ultimately are in themselves.
Where my view is relatively silent, and where Rottschaefer has helpful further
proposals to recommend in relation to recent philosophical and scientific theorizing,
is where I think Sellarss view remained prudently quiet, too: namely, in relation to
any details concerning the projected ideal scientific explanation of how linguistic-
conceptual capacities evolvedin the human species in the first place (see OShea
2007, 85 and n13; and Rottschaefer 2011, 296ff.). I did say quite a lot in the book
about closely related matters concerning which Sellars himself said quite a lot, in
particular about the possibility of an ideal behavioral-functionalist account of what
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
20/33
346 JAMES R. OSHEA
it is to acquire the ability to haveconceptualized thoughts and norm-instituting
intentions within and across inherited conceptual frameworks in general. This is an
account according to which both naturalistic and normative elements track one
another in complex relations of interdependence that I briefly encapsulated in what
I called Sellarss norm-nature meta-principle. But Rottschaefer and I agree on the
right-wing view that Sellars envisioned both kinds of overall explanation, referred
to above, within the ideal scientific image. Rottschaefer rejects my attempt to argue
that the formal epistemological truths about thinkers and their knowledge that Sel-
lars embraced and modified from the perennial philosophy are consistent within
fact synoptically fused withthat projected ideal scientific explanation. What I
dont see is why my attempt to interpret Sellarss synoptic vision as preserving both
the substantive neo-Kantian conceptual analyses and the projected ideal scientific
explanationsboth of them concerning the same rule-governed linguistic-conceptual
phenomena, one from the side of broadly empirically warranted scientific explana-
tion and the other from the side of perennial philosophical analysisis guilty of
producing a Sellars with a split personality in any sense that involves inconsistency,
as Rottschaefer seems to suggest. I think Sellarss stereoscopic vision is precisely a
delicate simultaneous combination of those two dimensions, the ongoing scientific-
explanatory and the philosophical-analytic, in one coherent view.
What is the ultimate justification for the Kantian epistemic principles that I
have suggested Sellars adapts from Kant in modified form? In an important sense,
Sellars rejected Kants notion of the synthetic a priori, as Rottschaefer correctly
notes (although Rottschaefer does not register the sense in which Sellars alsoclearly
sought to preserve a pragmatic and framework-relativized version of something
akin to the synthetic a priori, but this is not my main concern in this paper); and
Sellars also rejected, as Rottschaefer notes, what he calls the this or nothing
justification of epistemic principles that he finds in Chisholm and says is familiar
to the Kantian tradition (SK, III, 43).
Sellarss view is that Kant justified his synthetic a priori meta-principles by
embeddingthem within what is ultimately an analysisof the conceptually neces-
sary conditions on any conceptual representation of temporal states-of-affairs in a
world, as explained earlier. Like the good non-traditional, Kantian empiricist that
he is, Sellars does not think that anything other than (broadly) logico-conceptual
and (broadly) empirical warrants are ultimately required for this Kantian analysis, at
bottom. But Kants analysis has a complex and illuminating structure, according to
Sellars, and hence is in its own way explanatory. It is in fact one of the two dimen-
sions of explanatory coherencethat is embodied in the complex that is theory T,
to be discussed in the next section in relation to Sellarss MGEC. And contrary to
what Rottschaefer suggests, Sellars thinks that such an analysis does carry weight
against scepticism, despite not engaging in the futile task of attempting to provide
apresuppositionlessproof that there exists such a thing as empirical knowledge
(see TTC, VIII, e.g., 53). Here is how Sellars puts these subtle methodological
and interpretive points concerning Kant in KTE (see also TTC, part VIII):
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
21/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 347
(10) It is obvious to the beginning student that the truths of transcenden-
tal logic cannot themselves be synthetic a priori. If they were, then any
transcendental demonstration that objects of empirical knowledge conform
to synthetic universal principles in the modality of necessity [JOS: whichSellars himselfholds, too] would be question-begging. It must in a tough
sense be an analytictruth that objects of empirical knowledge conform to
logically synthetic universal principles. It must however, also be an illumi-
nating analytic truth, far removed from the trivialities established by the
unpacking of body in to extended substance and brother into male
sibling. [In a footnote here Sellars says that the concluding chapters of the
Critique show that Kant himself understood that transcendental logic
as knowledge about knowledge could consist of analytic knowledge about
synthetic knowledge.](11) It is also obvious, on reflection, that Kant is not seeking toprovethat
there is empirical knowledge, but only to show that the concept is a coher-
ent one and that it is such as to rule out the possibility that there could be
empirical knowledge not implicitly of the form such and such a state of
affairs belongs to a coherent system of states of affairs of which my percep-
tual experiences are a part. By showing this, he undercuts both the skeptic
and the problematic idealist who, after taking as paradigms of empirical
knowledge items that seem to involve no intrinsic commitment to such a
larger context, raise the illegitimate question of how one can justifiably movefrom these items to the larger context to which we believethem to belong.
What Sellars rejects are unexplicated, primitive, dogmatic appeals to the synthetic
a priori, to self-evidence, or to the this or nothing move.
In the next section I argue that, contrary to Rottschaefers criticism, the mod-
ified-Kantian aspects of my reading of Sellarss article, More on Givenness and
Explanatory Coherence (MGEC), are strongly confirmed by a closer look at the
argument of that article.
V. MORE ON SELLARS ON
GIVENNESS AND EXPLANATORY COHERENCE
More on Givenness and Explanatory Coherence is one of Sellarss most im-
portant and complex articles in epistemology. Rottschaefer argues in some detail
that my emphasis on the Kantian aspects of Sellarss theory of knowledge is not
supported by this text. In particular, Rottschaefer argues that what Sellars in
MGEC calls theory T is exclusively a scientificexplanatory theory, and not in
any respect a Kantian transcendental justification that is warranted in some way
independent of the sorts of warrant that derive from ongoing scientific theoriz-ing. I argue, however, that a further examination of MGEC strongly reconfirms
my original interpretation and also explains why as sensitive a reader of Sellars
as Rottschaefer might have been led to misconstrue certain key moves in Sel-
larss argument.
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
22/33
348 JAMES R. OSHEA
In sections III and IV of his article Rottschaefer helpfully highlights the central
features of my account of Sellarss views on the status epistemic principles, rightly
focusing upon what I called Sellarssperceptual reliabilityprinciple:
[PR] Ss perceptual judgment [P] thatx, over there, is redconstitutes a case
of perceptual knowledge if and only if there is a generally reliable
connection between cases of Ss judging that [P] and its being in fact
true that there is a red physical object over there. (Cf. Rottschaefer
2011, 299; and OShea 2007, 126)
[PR] is a meta-judgment, an epistemic principle, concerning the reliability, ce-
teris paribus, of first-order perceptual judgments such as [P]. In the terminology
of Sellarss MGEC, [P] is an example of a first-order IPM judgment, i.e., an
ostensible introspection, perception, or memory, and [PR] would be an instanceof Sellarss meta-judgment MJ
5, i.e., IPM judgments are likely to be true
(MGEC, 83). Sellars uses Roderick Firths Coherence, Certainty, and Epis-
temic Priority as a starting point for raising questions about the inferentialor
non-inferential warrantfor judgments such as [PR] and [P] (using my terminol-
ogy). In paragraphs 3339, Sellars uses Firth to make essentially the following
points, one of which is particularly important for my diagnosis of Rottschaefers
mistaken reading of MGEC.
Sellars explains Firths distinction between two modes of inferentialwarrant-
increasing properties (33) that a judgment may possess. Kind-1 inferentialwarrant is the sort of straightforward case in which [P] is validly inferable from
certain other statements of a specified kind (ibid.). Kind-2inferential warrant is
more indirect: in this case [P] is inferentially warranted if a certain meta-judgment
about [P]for example, [P] is believed by the relevant expertsis itselfinfer-
entially warranted in the straightforward kind-1 way, for instance by instantial
induction from the past successes of the experts. By contrast, a judgment has a
non-inferentialwarrant-increasing property if, put negatively, it is not warranted
in either of the two kinds of inferential way; and just what non-inferential warrant
might positively amount to, withoutappealing (as Firth does) to the Given, is oneof the central questions of Sellarss essay.
The point that is important not to miss, however, is that in sections 34 and
39 (and implicitly invoked again in 6061 and 76) Sellars appeals to Firths
use of coheres with in connection with inferable from to broadenthe scope of
inductive inference to include other [i.e., non-instantial] modes of non-deductive
explanatory reasoning (34): as a matter of fact, even if one counts the acquisi-
tion of a theoryby a substantial degree of confirmation as a variety of acquiring
inductive support, the distinctions remain reasonably straightforward (39). Sellars
is here broadening the scope of inductive inferential warrant (relevant to bothkind-1 and kind-2 inferential warrant) to include more complex accounts of the
inductive confirmation of scientific theoriesimplicitly including his own account
of induction in terms of the rationality of theory-replacement in ongoing scientific
inquiry relative to certain broader ends of explanation. (See Sellarss Induction
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
23/33
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
24/33
350 JAMES R. OSHEA
Sellarss response to this circularity threat involves a further subtle distinction
betweentwo kinds of explanatory coherence, both of which are aspects of what in
MGEC he calls a more encompassing version of what I have been calling theory
T (74, italics added), and which embodies epistemic principles such as [PR]. Sel-
larss first anticipation of what (in 61) he calls Theory T occurs in 40, directly
after having included the broadernotions of inductive confirmation by theories
within Firths kind-1 and kind-2 inferential warrants:
But suppose that P'[i.e., the inferential warranting principle (38)] is the
property of belonging to a theory of persons as representers of themselves-
in-the-world, which, although it has good explanatory power and is capable
of refinement by inductive procedures, was not(and, indeed, could not have
been) arrived atby inferences guided by inductive canons however broadly
construed. Would P'be an inferentialWP [warrant-increasing property] or
an explanatory but not inferentialWP? (MGEC, 40)
It turns out (5965) that if such a theory of persons as representers of
themselves-in-the-worldwhich Sellars in 61 calls theory T, and which is sup-
posed to provide the warrantfor such epistemic meta-judgments as the reliability
principle [PR]is itself inferentially warranted only by the standard means of
empirical confirmation, including the broadest sense involved in ongoing scientific
theorizing, then we are still stuck inside the vicious circle. Sellars indicates that
the circle threatens not only narrowly inductively supported claims, but empiri-cally confirmed claims in the broadened sense (which as we have seen, includes
scientific-theoretical hypotheses). Thus, of epistemic meta-judgments such as [PR]
(or in Sellarss terminology, meta-judgments MJ1,3,4
about the reliability of our
IPM judgments), his answer to the following question is negative if it is construed
empirically (in however broad a sense):
(58) . . . Might not these [principles] be both principles which provide
criteria for adjudicating certain empirical knowledge claims andempirical
knowledge claims in their own right? (59) Now if an affirmative response
took the form of a claim that MJ1, MJ3, and MJ4are empirically confirmedknowledge claimsthus putting them in a box with MJ
2[i.e., the instantial
induction, Statements which are accepted by the AGS [experts] are likely
to be true], a sensitive nerve would be struck. Would not such a claim
involve a vicious circularity? (60) Since it is obvious that they cannot be
empirical generalizations which owe their epistemic authority to confirma-
tion by instances, one might look for a less direct mode of confirmation by
experience [cf. 3439]. (61) Even if indirectly, however, an appeal must
ultimately be made [that is, if such meta-judgments are to be construed as
empiricalclaims even in the broadersense] to the fruits of introspection,
perception, and memory. (MGEC, 5861)
Sooner or later, Sellars continues, we would be confronted by such pairs of
statements as (here I will paraphrase by substituting my own abbreviations):
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
25/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 351
It is reasonable to accept epistemic principle or meta-judgment [PR] because
it is an element in a theory Twhich coheres with our perceptions [P]. And
[P] is likely to be true because it falls under [PR]. (Cf. MGEC, 61)
In 66 (cf. 58) Sellars consequently makes the crucial claim that in order to
break out of the circle what we need is a way in which it could be independently
reasonable to accept principles such as [PR] in spite of the fact that aground
for accepting such meta-judgments is the fact that they belong to T, which we
suppose to be an empirically well-confirmed theory. That is, to escape from the
circle we need a warranting ground for the theory T (of persons as represent-
ers of themselves-in-the-world) that is independent of and additional to the sort
of empirically based inferential warrant that theory Tmay also(now, coherently)
possess, i.e., of the broadly empirical sort paradigmatically exemplified by ongoing
scientific theoretical explanation.
In the final sections of the article (6689), Sellars accordingly sketches an
account according to which there is a more encompassing version of what I have
been calling theory T (74). Sellars indicates that this more encompassing theory
Tis something like theory T (88, 8081, 74) in that, as he concludes in the
final sentences of the article:
(88) . . . as it exists at any one time, theory Tis a complex [i.e., this is the
more encompassing version] which includes MJ5[e.g., [PR]] andattempts
to explain why IPM judgments [e.g., [P]] are likely to be true. The latterenterprise is still unfinished business [i.e., of ongoing empirical scientific
theorizing].
(89) It is in the former respect that it constitutesthe conceptual framework
which spells out the explanatory coherence which is the ultimate criterion
of truth. (MGEC, interpolations and final emphasis added)
That is, the complex that is the more encompassing theory Tincludes two dif-
ferent sources of warrant, only one of which is a matter of the ongoing process of
scientific explanation:
(84) Now for the linchpin. We must carefully distinguish between having
good reason to accept MJ5[PR] and having good reason to accept a proposed
explanationof whyIPM judgments [P] are likely to be true.
(85) To explain why IPM judgments [P] are likely to be true doesinvolve in-
ductive support [i.e., including the broadersense] for hypotheses concerning
the mechanisms involved and how they evolved in response to evolutionary
pressures. And thisobviously presupposes the reasonableness of accepting
IPM judgments [P]. (MGEC, interpolations added)
So ongoing scientific explanatory proposals will concern, for example, par-ticular empirically based theories as to how natural selection generated perceptual
mechanisms that are reliably truth-conducive. This aspect of the more encompass-
ing theory T, i.e., the ongoing scientific explanatory attempt to provide a causal
answer to the question How did we get into the framework? that is constituted
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
26/33
352 JAMES R. OSHEA
by epistemic meta-principles such as [PR] (78),presupposes the validity of [P]-
judgments in general in the effort to provide observational support for proposed
scientific-evolutionary theories as to why [P]-judgments are likely to be true.
But since Sellars has just argued that this ordinary scientific interplay between
hypotheses and observations does notbreak us out of the vicious circle, [PR]
[P], unlessthere is also a source of warrant for epistemic principles such as [PR]
that is independentof the ongoing process of empirical confirmatory warrant, the
burden of the entire article rests on the latter, non-empirically generateddimension
of explanatory coherence that Sellars argues is possessed by the more encompass-
ing version of theory Tand hence by epistemic reliability principles such as [PR].
So the final question is: what is Sellarss account of the latter dimension of non-
empiricallybased epistemic warrant for such epistemic meta-principles as [PR]?
Here the modified Kantian dimension of Sellarss theory of knowledge that I have
outlined in the previous sections and in the book breaks us out of the threatening
vicious circle. It does so by providing a substantive philosophical-conceptual analy-
sis of the conceptual framework which defines, i.e., independently of ongoing
scientific theorizing, what it is to be a finite knower in a world one never made
(73, emphasis added). Here philosophy makes a contribution that ongoing scientific
theorizing does not makeand should we really be surprised that Sellars thinks
that philosophy can play such a role?by analysing the general structure of the
systematically coherent framework of epistemic principles that must be exhibited
by anyparticular conceptual framework within which ongoing scientific or any
other epistemic activity is possible in the first place. Once this is established we can
now coherently hold that to be in this framework is to appreciate the interplay of
the reasonablenesses of inductive hypotheses [in the broadest sense] and of IPM
judgments [P] (MGEC, 75, interpolations added). That is, once we have in this
manner revealed the way in which it could be independentlyreasonable to accept
epistemic meta-principles such as [PR], then we have philosophically resolved the
appearance of vicious circularity that seemed to threaten the ordinary scientific
interplay (75) between warranted principles and warranted observations that
is characteristic of all ongoing theoreticalexplanation (76).
In MGEC Sellars entitles this account of the permanent dimension of explana-
tory coherence that is required for all possible human knowledge, and is explicated
by philosophical analysis rather than by ongoing scientific explanation, Epistemic
Evaluation as Vindication (68). As we have already seen, in KTE he had char-
acterized essentially the same project as transcendental linguistics and in TTC
as pure pragmatics or transcendental logic. In The Structure of Knowledge
(SK) he characterized it this way:
(45) . . . We have to be in this framework [of epistemic principles] to bethinking and perceiving beings at all. I suspect that it is this plain truth which
is the real underpinning of the [Chisholmian, this or nothing] idea that
the authority of epistemic principles rests on the fact that unless they were
true we could not see that a cat is on the roof.
-
8/12/2019 How to Be a Kantian and a Naturalist About Human Knowledge
27/33
BEING KANTIAN ANDNATURALIST ABOUT HUMAN KNOWLEDGE 353
As we have seen, t
top related