Cartesian Conceptions? Sellars (and Husserl) on

Perceptual Consciousness

May 17, 2013

1 Introduction

Wilfrid Sellars, recounting his philosophical history, notes:

...[I]n 1933, I went to Bu�alo as a teaching assistant...Marvin Farberled me through my �rst careful reading of the Critique of Pure Rea-son and introduced me to Husserl. His combination of utter respectfor the structure of Husserl's thought with the equally �rm convic-tion that this structure could be given a naturalistic intepretationwas undoubtedly a key in�uence on my own subsequent philosophi-cal strategy.

We are all of us, I am sure, aware of Sellar's strategy of fusing the two images,Manifest and Scienti�c. And this passage speaks to the early formation of thatstrategy. And while I am interested in the fusing of the images in Sellars, Iam more interested in the fact that Husserl in�uenced Sellars. That may seemunlikely, but nonetheless, I claim, it is so. The Husserlian in�uence is displayedmost clearly in a relatively undiscussed paper of Sellars', �Some Re�ectionson Perceptual Consciousness�. Sellars gave that paper to an audience mostlycomposed of phenomenologists, and although he does not mention Husserl byname, it seems very likely that he was thinking of Husserl when he writes,

...[F]or longer than I care to remember I have conceived of philo-sophical analysis (and synthesis) as akin to phenomenology. I wouldtherefore expect this audience to be more sympathetic to what I saythan many of my colleagues would expect.

What's more�in this paper, Sellars o�ers his own �phenomenological reduction�,using just that term. This makes the paper a particularly interesting paperfor anyone who is interested in phenomenology. But I should now point outthat my e�orts in this paper are not going to be devoted to convincing anyonethat Sellars was generally in�uenced by and that he did a bit of something hecalled �phenomenology�. I take that to be beyond dispute. No, what I wantto show is that a key feature of Sellars' view in SRPC is itself an out�owing of

1

his phenomenological commitment, a commitment to phenomenology in a morethan Pickwickian sense of that term.

I want to show more than this, however. I also want to show that Sellars isnot guilty�at least not in this essay�of a crypto-Cartesianism in his account ofperceptual consciousness. He has been accused of such, as I will discuss. Con-nected to that, I also want to show that the account of perceptual consciousnessthat Sellars gives is not, as it might seem, retrograde, yeilding territory he hadwon in EPM. I also want to voice a couple of suspicions and worries of my ownabout Sellars' account, but I will get to those in due time�actually at the end.I should warn you that my �nal purpose is not to come down hard in favorof or against Sellars' account, but instead to acheive a better understanding ofwhat it is and why it is what it is. I am concerned with providing a perspicuouspresentation of Sellars' view.

2 Recapping the Re�ections

Sellars judges the primary datum of his view to be the �paradigm case" of seeinga physical object from a point of view in physical space. His working example isseeing a red brick over there facing me edgewise. Sellars begins his `reduction'by distinguishing between the object seen�a red brick�and what we see ofthe object�certain of its surfaces. What we see of the object Sellars categorizesas dependent particulars, as �parts" on a careful and self-conscious use of thatterm. Now, Sellars is not particularly interested in trying further to clarify whatit means to call what we see of the object a dependent particular, but he doeswant to stress it; in particular, he stresses that what we see of the object is aparticular. His concern her is to mark the di�erence seeing parts of the objectand seeing universals, for these too are among what we see of an object, andmust be kept distinct from what we see of an object that is a constituent of anobject. Universals, Sellars notes, are in no useful sense constituents of objects.

Sellars limits his focus to peceptual takings�occurrent believings which aperceiver (given her perceptual set) is caused to have in visual perceptual situa-tions. And, although Sellars concedes that it is di�cult to analyze the relevantsense of `cause', the object seen is the external cause of her perceptual, vi-sual taking. Familiarly, Sellars understands occurrent believings as mental acts;their appropriate expression is the tokening of a sentence. If the sentence isa subject-predicate sentence, we may talk of the subject and predicate of thecorresponding believing.

Commonly, accounts of the sort that interest Sellars treat perceptual tak-ings as having a subject-predicate form, but as moreover having a subject thatis to be expressed by a demonstrative. Sometimes, the subject is a complexdemonstrative, like

This red brick is larger than that one

Such a sentence stands in an obvious relation to a compound sentence

This is a brick and it is read and it is larger than that one

2

But should we treat the thought expressed by the �rst of these sentences ashaving the structure of the second? If we do so treat it, then we will treat thesubject of perceptual takings as a bare `this', and will shunt all characterizinginto a predicative position. So Sellars' working example would become

This is a red brick facing me edgewise

Now, the referent of `this' in perceptual contexts is understood to be the ob-ject seen. So Sellars construes his working example (which he supposes to beveridical) as a case in which the perceiver sees a certain red brick facing himedgewise and sees that it is a red brick facing him edgewise. This is because theperceivers perceptual believing has the form

This is a red brick facing me edgewise

But this sort of account dissatis�es Sellars. He is worried about the divingof visual takings into (1) a pure demonstrative subject and (2) a predicativeconstituent in an explicit predicative position. Sellars reckons this will not do.But realizing why takes work.

Are all perceptual takings rightly understood as sentences with demonstra-tive subjects? Perhaps they may�even, must�have some demonstrative com-ponent, but does that have to be a demonstrative subject? At the same time,however, we may wonder whether distinguishing the situation we see and whatwe see of it might make us tend to believe that the demonstative analysis isprimary, in some gut feeling sense of `primary'. We may believe that visualperception always somehow or other includes the perception of an object, andthat the visual taking must involve a demonstrative referring to that object.

Sellars doesn't decide this issue; he instead adds another complication. Evenif we work with a demonstrative account, we will need a distinction betweenwhat we see and what we see it as. A demonstrative account notoriously hasdi�culty incorporating the distinction. However, an account that feeaturescomplex demonstrative phrases shows promise.

Consider a perciever who sees a bush as a bear has a perceptual taking ofwhich the subject constituent is the complex demonstrative, �This large blackbear...� And if the taking can be said to be a peceptual believing, thus

This large black bear moving toward me

then we can distinguish between takings that and takings as and further suggestthat what a perceiver sees an object as is what keeps company with `this' inthe complex demonstrative. We could observe of this distinction that strictudictu perceptual takings just are the complex demonstrative constituents ofperceptual beliefs and that the explicit predicative constituent is not part ofwhat is taken, but rather, and simply, what is believed about what is taken.This would make takings then much like Strawson's presuppositions: and wecan accomodate this by extending our concept of occurrent belief by markingthe di�erence between believing that and believing in. Perceptual believings inthen could be treated as the subject constituent of the believing expressed by

3

This red brick facing me edgewise is too large to �t that gap.

We thus treat the complex demonstrative as presupposing that the referent of`this' is a red brick facing the perceiver edgewise, and because of this presup-posing we could argue that the perceiver perceptually takes the referent of `this'as a red brick facing her edgewise.

So far, this would be so good. But the problem that faces our perciever,along with the red brick, is accounting for the sense in which `this' in the com-plex demonstrative can be understood to have a reference independent of thepredicates which keep it company. In short, can we make sense of the claim that`this' refers to something that is not a red brick? Putting the question in a waythat re�ects all that we have so far considered, we can ask: should we say thatJones' seeing a certain object as a red brick facing her edgewise consists in herbelieving in a red brick facing her edgewise, where this believing in is visual inthe sense that her having this belief is (given her perceptual set) brought aboutby the action of that object on her visual apparatus?

At this point, Sellars takes us back to the issue of the universals that areincluded in what we see of an object. He does so in order to re�ne the distinctionbetween the object seen and what we see of it. That universals are included inwhat we see of an object strikes Sellars as obvious enough, but how it includesthem is not obvious enough, is indeed highly problematic. Sellars now shifts hisworking example from the red brick to his best beloved pink ice cube. It consistsof ice largely because of its causal properties, its cooling tea and its being meltedby �re. So when we see the pink ice cube in normal conditions, we see that thetransparent cube is mad of ice. We see the pink ice cube as a cube of pink ice.But do we see of the pink ice cube the cause properties involved in its beingmade of ice? Do we see of the object these causal properties? Admittedly, thequestion is grammatically lame, but even so its seems that the answer is No.Consider the contrast: We see that the ice cube is pink, see it as pink, we seethe very pinkness of the object�also its very shape, albeit perspectivally. Do wesee its very melted-by-�reness?

That we seem not too becomes more fascinating when we notice that inseeing the cube as a cube of ice, we are also seeing it as cool. But do we seeof the cube its very coolness? The question induces an oscillation: It seemsthat phenomenologically the very pinkness of the cube and its very coolness areequal: we see them both. (The idea that we do not is a precipitate of scienti�ctheorizing, not of phenomenologizing.) But if we are asked straight up whetherwe really see its coolness, we cannot quite answer that we do�we grope forsomething else to say, something like we believe in its coolness, we take it ascool, but do not see its coolness. However, even if we do say something likethis, we also know it will not quite do. The cubes very coolness is not merelybelieved in, even if it is not seen. Clearly, we do not feel or imagine a coolnesswhen seeing the cube. We are considering its coolness, not a coolness.

Sellars leaves this issue here, turning to the more limited but still di�cultquestion of the sense in which the perceiver sees of the ice cube its very pinkness.This question is one that phenomenology can only take us so far in answering. It

4

assures us that something, somehow a cube of pink in physical space is present inthe perception other than as merely believed in. How are we to understand thisodd locution? Like this: When we see a pink ice cube we see something thatis somehow a-cube-of-pink-ice-in-physical-space is present, and present otherthan merely as believed in. Sellars tries further to clarify what he means bya brief comparison with traditional terminology. What he calls the somehow

presence of a cube of pink does not consist in its intentional in-existence asthe content of a conceptual act. Nor is its character as cube of pink in spacefacing the perceiver edgewise a matter of its actually being a cube of pink inphysical space. Rather, it is somehow a cube of pink in physical space facingthe perceiver edgewise without actually being a cube of pink in physical spacefacing me edgewise, and without being merely the content of a belief in a cube ofpink in physical space facing the perceiver edgewise. Clearly, Sellars is walkinga tightrope here.

He denies that the somehow presence of a cube of pick consists in the inten-tional inexistence of a cube of pink as the content of a conceptual act. In otherwords, Sellars denies that there is a cube of pink ice `in' my mind, a cube of pinkice made of the stu� that dreams are made of, not made of water or extendedin physical space. And he denies that the somehow presence of a cube of pinkice is a matter of its actually being a cube of pink ice. In other words, Sellarsdenies that there is a cube of pink ice in my mind, made of water and extendedin physical space. Rather, it is somehow a cube of pink in physical space facingme edgewise whout actually being a cube of pink in physical space facing meedgewise. Even so, it cannot be reduced to the content of a belief in a cube ofpink faceing me edgewise. What I have is a somehow presence of something, nota mere belief in a presence of something. As Sellars puts it: seeing of the cubeits very pinkness is to be analyzed as the somehow, other than merely believedin presence of a pink cube in visual experience.

Sellars notes that he introduced that �nal phrase, `visual experience' to makeroom for the fact that we can seem to see a cube of pink ice when there is infact no cube in the o�ng. So Sellars mobilizes the phrase, `ostensible seeing of acube of pink ice...� to refer to a visual experience which would be a case of seeinga cube of pink ice if there was such a cube in the o�ng and causally responsiblefor the ostensible seeing. The somehow, other than merely believed in presenceof the pink cube is thus what is common to veridical and non-veridical ostensibleseeings of a cube of pink ice facing one edgewise. Ok: somehow!

At this point, Sellars takes leave of phenomenology and begins to theo-rize. He brings into the discussion the theoretical or proto-theoretical states ofperceives that he calls `sensations'. These are meant to explain the results ofphenomenology, to explain the somehow presences. The fact that a perceiverostensibly (veridically or non-veridically) sees the very pinkness of the cube isexplain by proto-theorizing the occurrence in perceivers of a sensation of a pinkcube. So, a sensation of a pink cube is a sensation of a certain kind, the kindnormally brought about by causal action on the visual apparatus of the percieverby physical cubes of pink ice in the o�ng.

But these sensations are not literally cubes of pink. �Still, it is not simply

5

false to say that they are cubes of pink. To say why it is not simply false wemight appeal to the tradition of analogy and say that the pinkness of a sensationwhich belongs to the `of a cube of pink' kind is analogous to the pinkness of itsstandard cause. The trouble with saying this is cashing in or spelling ou theanalogy. Sellars does so in the following way:

According to [my] version of the adverbial theory of sensing, then,sensing a-pink-cube-ly is sensing in a way which is normally broughtabout by the physical presence to the senses of a pink...object, butwhich can be brought about in abnormal circumstances by objectthat [are not] pink...and, �nally, according to [my] form of the adver-bial theory, the manners of sensing are analogous to the common andproper sensibles in that they have a common conceptual structure.Thus, the color manners of sensing form a family of incompatibles,where the incompatibilities involved are understood in terms of theincompatibilities involved in the family of ordinary physical colorattributes...

Sellars recognizes that these bare statement of the theory needs to be supple-mented. What needs to be enforced in it is the idea that the analogy strictlypreserves the conceptual content of the predicates involved, even while trans-posing the content into the �radically di�erent categorial framework to whichmanners of sensing belong�. Only such enforcement, whatever speci�cally itwould involve, whether providing greater detail or adding additional dimen-sions to the analogy, will allow Sellars to explain the way that color conceptspreserve their content as they migrate from the manifest to the scienti�c image.If Sellars is going to give meaning to the idea that the very pinkness of the cubecan be somehow present in an ostensible seeing of pink cube as a pink cube, hemust be able to show how content is preserved through a change of category.

A sensing of the sort Sellars postulates is not a sensing as. To sense (acube of pink)ly is not to sense something as a cube of pink, but it is a statedesigned to explain what it is to ostensibly see a cube of pink as a cube of pink.The important point is: �sensing, though a constituent of seeing something assomething, is not itself a case of seeing something as something. �

At this point, Sellars has in hand tow constituents of an ostensible seeing ofa pink cube as a pink cube.

1. The taking or believing in, construed on the model of a complex demon-strative phrase, like �This cube of pink ice�; and

2. The sensing (a cube of pink)ly

And the key question now is how these constituents are related. What kindof togetherness do they have? We might be tempted to declare their kind oftogetherness causal: given a certain perceptual set, (2) immediately causes (1).Sellars assents to this declaration, but only as right in part. He regards the kindof togetherness as more intimate than merely causal.

6

To show this, he asks what, in veridical ostensible seeing, is the referentof the demonstrative phrase, �This cube of pink ice�. The obvious answer, heo�ers, is that the referent is a certain cube of pink ice. But imagine that thereis instead a colorless cube of ice in the o�ng, instead of a pink one. Doesreference fail in that case? Sellars says that perhaps we would judge that thereis a referent�the colorless cube�but that the perceiver takes it to be a pinkcube. Would we also be willing to judge that she sees the referent as a pinkcube? Well, we cannot say that she sees its very pinkness, but can rather saythat she ostensibly sees it very pinkness. Imagine now that there is nothingthere at all; the perceiver is hallucinating. Does reference fail? Sellars knowsthat many philosophers will say Yes. After all, there is nothing in the o�ng ofwhich it can be said that the perceiver takes it to be a pink cube. Is that Yesthe Last Word?

No. Sellars contemplates another possibility. What if, after all the presup-positions packed into the complex demonstrative phrase have been unpackedinto explicit propositional form, it makes sense to preserve the referent of `this'as the sensation of a cube of pink facing one edgewise? In other words, if weunpack

This cube of pink ice facing me edgewise...

as

This is a cube and this is pink and this is facing me edgewise

could we treat the 'this' as referring to the sensation of a cube of pink ice...?The serious di�culty here is that the referent of the complex demonstrativeseems like it must be the sort of thing that could be in the o�ng, over there,in physical space, facing me. And no sensation is in the o�ng, over there, inphysical space, facing me. Categorically, no sensation is such.

Still, Sellars argues, it would be wrong to say that the perceiver sees nothing,to deny that he sees anything. Typically, denying that someone sees anythingimplies that the perceiver is imagining something. So, if we are willing tocountenance the idea that we can thin out perceptual commitments via phe-nomenological reduction, then the reduction would end not with

This cube of pink ice facing me edgewise

but with

This somehow (a cube of pink over facing me edgewise)ly

And this would give us a chance to rescue reference by contstruing it to be thesensation, �for the sensation is indeed that in the experience with is somehow acube of pink over there facing one edgewise.�

It is worth noting, �rst, that making this move does not require that theperceiver conceptualize her sensation as a sensation, and second, that `some-how' admits as a special case, `straightforwardly'. So while the referent of themost cautious of perceptual takings can be treated as a sensation, we need

7

not conclude that the referent of all is to be treated as a sensation. Perhapsthe ultimate�the terminal�referent is always a sensation; but, by treating thecomplex demonstrative phrase as phenomenologically reducible to

This somehow (a cube of pink facing me edgewise)ly which is a cubeof pink facing me edgewise

the intial stages of reference rescue can proceed without interpreting the referentas a sensation.

2.1 Sellarsian Phenomenological Reduction

We can think of phenomenological reduction generally as the method, in some

ways analogous to Descartes' Method of Doubt, of shelving or putting out of playor as bracketing the commitments of our ordinary or scienti�c frames of mind,our �natural attitude�. SPR is Sellars' perceptual re�nement of the method. Itis supposed to `thin out' our perceptual commitments until, in its ultimate orterminal phase, we are left with a bare minimal commitment, one presumablyrequired in some way if we are going to maintain our claim that what we aredoing is reducing perceptual consciousness. We will come back to that. So howdoes SPR work? Recall that early in the essay, Sellars warns that we shouldnot construe the perceptual believing expressed by

This is a red brick facing me edgewise is too large

as if it were identical with the perceptual believing expressed as

This is a red brick and it faces me edgewise and it is too large

But he leaves the signi�cance of this non-identity to be shown, not really said,in later stages of the essay. And there what he attempts to show is that ifwe fail to appreciate the signi�cance of this non-identify, we are likely to thinout perceptual commitments (moving in the direction of minimal perceptualtakings) like this

This dangerous black bear standing on its hind legs...

This dangerous black bear....

This black bear...

This black object... [etc]

and thus will be left by our series construing the referent of `this' as, in a minimalperceptual taking, the sort of thing that could be in the o�ng, over there, facingme. But since sensations, as we have seen, are categorically not in the o�ing,over there, facing me, we would rule out sensation as the referent of a minimalperceptual taking. By contrast, SPR culminates in

This somehow (a cube of pink over there facing me edgewise)

8

that is, in a sensation, something precisely not over there, facing me edgewise.So SPR appreciates this non-identity and proceeds along a path di�erent fromthe one that the identity would suggest. So how should we then �esh out acharacterization of SPR? Well, I take it that we should see it as more closelysimilar to, say, Husserlian Phenomenological Reduction (HPR) than we mighthave expected. It is internal to HPR that the reduction has to preserve theinherent sense or meaning of the act reduced throughout the reduction. Whatthat mean is this: if it belongs to the inherent sense of a perceptual takingthat it is a taking of the object itself�the object facing me over there edgewise,say�then that inherent sense must be preserved throughout the reduction. If itis lost, then the reduction has, somewhere along its path, lost track of the act tobe reduced. In practice, this preservation is an extremely sensitive occupation�and Sellars' essay shows that, in its way.

It is also crucial to appreciate that part of the signi�cance of the non-identityis that it will allow us to identify something common to veridical and non-veridical ostensible seeings. If we accept the identity, then we will need a referentfor `this' that is really-and-for-true over there, facing me edgewise. Where thereisn't, as in hallucination, we will be unable to rescue reference for `this'. Butthis would turn the hallucinator into a mere imaginer�something Sellars wantsnot to do, and believes we should not do. Sellars is quite sure that we shouldnot deny that the hallucinator sees anything. The hallucinator does, in a sense,see something�she has a sensation.

So what SPR needs to culminate in, when it culminates with the most catiousof perceptual takings (it need not and typically will not in veridical ostensibleseeing), is something that preserves the inherent sense of the perceptual actwithout doing so in a way that more-or-less guarantess reference failure for'this' in the case of non-veridical ostensible seeings, or without construing thehallucinator as a mere imaginer. Consider the following dilemma: Either wecan understand perceptual consciousness as a relation to a really-and-for-trueover there, facing me edgewise referents, in which case the reference failure ofnon-veridical ostensive seeings de�es explanation, or we can understand percep-tual consciousness in terms of perceived mental contents (images or somesuchwhatnot), and in e�ect lose the distinction between verical ostensible seeingsand non-veridical ones (`in e�ect': if we do not lose the distinction, then wewill only be able to retain it by calling upon something `external' to perceptualconsciousness and our account of its workings). What we need to escape thedilemma is an account of perceptual consciousness that accounts for it in termsof structures through which perceptual consciousness is directed, rather thanthe real objects, if any, toward which it is directed or the actual mental con-tents perceived. The structure will be neither a really-and-for-true over, there,facing me edgewise object, nor a real subjective mental content. This is, I takeit, a generic description of which Husserl's account of perceptual consciousnessis one species and Sellars' is another. Husserl structures his phenomenologicalre�ections on the distinction between act, content and object, where the contentis the quality and the matter act (the quality might be perceiving, say, and thematter a red brick, but where both are `moments' of the act). Sellars structures

9

his phenomenological re�ection on the distinction between act, adverbial `con-tent' and object, where the adverbial `content', precisely because it is adverbial,is on the act-side of the structure and not on the object-side.

2.2 Sensations, What and Where?

To answer these questions, it makes sense to expand our horizons a bit andto consider Sellars' largely congruent discussion of perceptual consciousness in�The Structure of Knowledge�. There, Sellars writes

So far we would be little better o� than if we simply said that os-tensibly seeing that there is in front of one an object which is redand triangular on the facing side di�ers from merely thinking thatthere is an object in front of one which is red and triangular on thefacing side, by virtue of being a thinking which is also an ostensibleseeing. But we can say more. For, phenomenologically speaking,the descriptive core [the nonpropositional feature or compontent ofthe experience] consists in the fact that something in some way redand triangular is in some way present to the perceiver other than asmerely thought of.

Sellars notes that the �multiple inde�niteness� of this description is troublesome�it poses problems. But he thinks that the description is helpful in its de�nitenegative aspect: �The mode of presence is not that of being thought of.�

So, we might say, there is a multiply inde�nite something-or-other present tothe perceiver in a perception, and present other than as being thought of. Or, ifwe shift back to the terminology of �Some Re�ections...�, the multiply inde�nitesomething or other is present other than as merely believed (in). Sellars toyswith the comparison of what he has in mind to the scholastic distinction betweensomething's having being for sense as well as being for thought, but he does notpress the comparison.

In �Some Re�ections...�, Sellars concentrates the inde�niteness that concernshim to something, somehow inde�niteness, i.e., a something that is in some way(Sellars will shift to how), say, a red brick facing me edgewise. He does notomit the worry about the `some way present', but he does not thematize it bypresenting it as, say a something, somehow, someway.

2.3 The Ultimate Referent?

As the SRPC closes, Sellars says the following:

Thus while the referent of the most cautious perceptual taking can beconstrued as a sensation, we need not conclude that the referent of allperceptual takings is a sensation. For while it could be argued thatthe ultimate referent is always a sensation, by construing our originalcomplex demonstrative phrase along the lines of phenomenologicalreduction as

10

This somehow (a cube of pink facing me edgewise)[ly] which is acube of pink facing me edgewise...

the initial stages of reference saving can proceed without interpretingthe referent as a sensation.

There is a great deal in this section�and I have omitted its crucial opening sen-tence, one on which I will comment momentarily�that needs to be sorted out.Perhaps the most puzzling claim in it is the claim about the �ultimate� referent.What does Sellars mean by this claim? Note, �rst, that the claim is carefullymodulated: �For while it could be argued...� So what we are getting here isnot an assertion that the ultimate referent is always a sensation, but ratherthe ticking o� of an argumentative possibility. Note, second, the ambiguity ofSellars' talk of ultimacy: it could mean that it could be argued that the referentof any perceptual taking is ultimately to a sensation, or it could mean that itcould be argued that the �nal stage in a SPR, its terminal or ultimate phase,the �nal step in the stepwise thinning out of perceptual commitments, is onein which the referent is to a sensation�but that would leave open the questionof the referent in the earlier steps of the SPR. In fact, we might think that incases of veridical ostensible seeing, there is going to be no need to construe thereferent as a sensation at all. In such cases, we might say that the referent is tothe pink ice cube facing me edgewise.

Part of what is puzzling here is what motivates the stepwise progressionthrough the SPR. Is an SPR fated, so to speak, to step stepwise from its open-ing much-committed perceptual taking to its terminal least-possibly committedperceptual taking, so that it would be true that the actual referent of my �rsttaking would be decided by the referent of the last? Is alph held hostage byomega? Or is it rather that, although I may not know whether my ostensibleseeing is veridical or non-veridical, and so am unsure where in the SPR thereduction ends for me, it need not be the case that it ends with a minimaltaking? If this second possiblity is right, then although I may not be able fromthe inside, so to speak, i.e., phenomenologically, tell whether I am veridicallyostensibly seeing or not, if I am veridically ostensibly seeing there is no need toworry about less perceptually committed takings. My referent in such a case isthe object itself and not my sensation.

Here, I think, we have to face up to the crucial opening sentence of the sec-tion: `Notice also that �somehow�' admits as a special case �straightforwardly�.'This is what the �nal embedded sentence of the essay is supposed to show, I takeit: �This somehow (a cube of pink facing me edgewise) which is a cube of pinkfacing me edgewise...� What is that sentence showing? I am tempted to say thatit shows how it can be that, when veridically ostensibly seeing, the perceivercan have a sensation that conduces to the physical pink ice cube in such a waythat reference should be construed as to the physical pink ice cube and not tothe sensation. (What is seen in the veridical case is also what is referred to, i.e.,something that is straightforwardly (and not merely somehow) a pink ice cube.)It is important not to neglect the elipsis that ends the showing sentence. Thesentence does not end where it ends, we might say, but continues to something

11

like (using an earlier continuation) �...is to large for the opening�. In this case,what the perciever is referring to is something that can categorically be too largefor a [physical] opening. So why not take the showing sentence to show us howperceptual takings work in the veridical ostensible seeing case, and as showinghow, although a sensation remains involved in the account of the experience,the sensation is not referred to by the demonstrative in the perceptual taking?

If we accept the �rst possibility, then it turns out that even in the case ofveridical ostensible seeing, the referent is to the sensation. But this makes theveridical case and the non-veridical case too close for my comfort. In particular,it looks like a way of impaling oneself on something like a horn of my earlierdilemma, namely on something like the horn that ruins the distinction betweenthe veridical and non-veridical by understanding perceptual consciousness interms of perceived mental contents. Now, it is true that the sensation hereis not exactly to be construed as an object in an act/object structure, butthe parallel remains bothersome. Why run together the veridical and the non-veridical in this way? Why deprive the veridical ostensible seer of the physicalobject as the referent in his perceptual taking?

2.4 The Hallucinator

How are we to understand what is happening with the hallucinator? As Sellarsmakes clear, there is nothing physically present to serve as the referent of thehallucinator's demonstrative. Even so, Sellars insists, it would be �incorrect�to deny that the hallucinator sees anything: �After all, the [hallucinator] is notimagining anything, which is what the denial that he is seeing anything wouldnormally imply.� I am willing to concede this to Sellars�at least for argument'ssake. So what Sellars is trying to do is to situate the hallucinator simultane-ously in relation to the mere thinker (believer, conceiver) and in relation to theveridical perceiver.

Sellars takes it that the hallucinator is `more' than a mere thinker and `less'than a veridical ostensible seer. How so?

Let's concentrate on the `more' that the hallucinator has than the merethinker. The mere thinker's has no (relevant) sensation. He or she thinks of apink ice cube over there without a something, somehow (a pink ice cube overthere) occuring. The pink ice cube over there has, we might guardedly say, onlybeing for thought and no being for sense.

But the hallucinator does have a sensation�a something, somehow (a pinkice cube over there)�occuring, present to her. And the hallucinator's perceptualdemonstrative refers to that sensation, that something, somehow (...). However,her sensation is not caused by any pink ice cube in her environment. Its causemay be a drug or a kick in the head or both or neither but anyway not a pinkice cube over there. Her sensation has a cause, sure enough; but just not theright (kind of) cause. So the hallucinator has a sensation but nothing for it torelate to in its peculiar adverbial way. In such a plight, it is the sensation itselfthat functions as the referent of the perceptual demonstrative.

Now, it will not do, given the shape of Sellars' thinking, to in ordinary cases

12

of hallucination to treat the hallucinator as hip�as aware of his or her hallu-cination as such. No, she is not aware in ordinary cases that she hallucinates.She believes herself to be a veridical ostensible seer. This is why, presumably,Sellars insists in Sec. 60 that treating the referent of `this' as the sensation neednot require the hallucinator to conceptualize her sensation as a sensation. Sheinstead conceptualizes her sensation as the presentation of a pinck ice cube overthere. That is, she is mistaken, in the ordinary non-hip case, about the natureof the referent of her `this'. She takes it to be something straightforwardly apink ice cube over there.

Her non-veridical ostensible seeing thus involves two constituents:

(1) This cube of pink ice over there... [The taking or believing in]

(2) a something, somehow (a pink ice cube over there) [the sensing]

So the hallucinator, in having (2) has `more' than the mere thinker, but insofaras her `this' in (1) refers merely to her sensation, since her sensing is wronglycaused, she has `less' than the veridical seer.

2.5 The Veridical Seer

How should we understand this case? Well, we have already been describing itinter alia as we discussed the hallucinator. But oddly enough, as I understandSellars' essay, the primary puzzle of it is understanding the veridical seer. Likethe hallucinator, the veridical seer has both a perceptual taking and a sensation,where the sensation is the immediate cause of the taking. Moreover, the veridicalseers sensation is rightly causesd, i.e., caused by a pink ice cube over there. Sofar, no surprises. But what is the referent of the perceptual demonstrative ofthe veridical seer?

It is tempting to claim that the referent must be the same as the referentof the the hallucinator's `this'. And indeed, Anil Gupta has so claimed. Imentioned this earlier. Here is what Gupta says.

The object demondstrated in perception is the thing before the sub-ject, and on the Sellarsian analysis, the thing present in experi-ence is invariably a sensing...No matter how hard philosophers tryto evade Cartesian conceptions�and none tried harder than Sellars�their deeper re�ections, it appears, invariably entangle them withthese conceptions.

But is that right? I have already suggested that I do not think it is. Whatis most striking about Gupta's claim here is his simultaneous concession thatSellars tried harder than any philosopher to avoid Cartesian conceptions andhis willingness to treat Sellars' e�orts as a failure. But are they? Does Sellarsyeild to Descartes here, in the very bowels of the account of perceptual con-sciousness? Is Sellars an intestine Cartesian, a crypto-Cartesian, at the end?�And, keep in mind, Sellars takes himself to be doing phenomenology, at leastin part, Husserlian phenomenology, and Husserl, despite his many references to

13

Descartes and even comparisons of himself to Descartes, is another philosopherwho strained every nerve attempting to evade Cartesian considerations. �Areboth Sellars and Husserl �nally blind guides, guilty of obsessively straining atgnats only eventually to swallow the camel?

One puzzling feature of Gupta's essay is that he has no comment on Sec. 61of �Some Re�ections...�, its ultimate section. This is a ringing silence. There,Sellars notes that (just to repeat)

...[W]hile the referent of the most cautious perceptual taking can beconstrued as a sensation, we need not conclude that the referent ofall perceptual taking is a sensation.

I reckon this to tell decisively against Gupta: it is not the case that the demon-strative of every perceptual taking refers to a sensation. The veridical seer hasa sensation, surely; but he is not referring to it. Rather, he refers to the pink icecube over ther�the one in his environment, causing his sensation. The pressingquestion now is this: what is the sensation doing, as it were, what is it for, inthe veridical seeing?

Let's take Sellars at his word. Sensations, as we have seen, are to be un-derstood as being pink adverbially, they are somehow (pink). The sensation isnot literally�and let's here specify that as meaning adjectivally�pink. The sen-sation is adverbially pink but not adjectivally so. Now, to be adverbially pinkis to be such as to somehow preserve the content of being adjectivally pink,but across a transcategorial analogy. For the veridical perceiver, as well as forthe hallucinator, there is the presence of this something, somehow (pink). Foreach, this is the something more that is had that the mere thinker about thepink ice cube lacks�the mere thinker, lacking the sensation, does not rate as aputative ostensible seer. Further, we know that the hallucinator's perceptualdemonstrative refers to her sensation, but that she does not conceptualize itas doing so. She conceptualizes her referent as the really-and-for-true pink icecube over there. Presumably, the veridical perceiver conceptualizes in just thesame way. But does he do so mistakenly, as the hallucinator does? Unlike thehallucinator, the veridical perceiver's sensation is actually caused by the really-and-for-true pink ice cube. Is that causal fact�along, perhaps, with the factabout the veridical perceiver's conceptualization (and perceptual set)�enoughto make it the case that the veridical perceiver's perceptual demonstrative refersto the really-and-for-true pink ice cube?

Go back for a moment to Sellar's idea that doing phenomenology is to beunderstood as part of the exploration of the Manifest Image. Now, I submitthat the Manifest Image understands perceptual consciousness to be a way ofbeing open to the world: we are open to the world of objects and it is opento us; perceptual consciousness is no enclosure. So when we begin re�ect in aManifest Image way, we begin by taking perceptual consciousness as opennessto the world. We begin with a perception, say, whose sense crucially involves itspresenting us with an object out there, an object that transcends consciousness.We do not start toward objects out there, transcendent objects, by startingfrom within, particularly not by starting from something within an enclosure.

14

If anything like that enters our re�ection on MI-perception (Manifest Im ageperception), it does so later. We do not, just to put the point pointedly, startwith sensations. We instead start with the perceptual act as a whole and workour way in, and only so far in as is necessary. For our re�ection to maintainitself as re�ection on MI-perception, we must not lose our outside-in orientation,we must preserve the sense of the MI-perception as presenting us with an objectout there. At any rate, this is how I believe that Sellars sees things, Manifestly.It is how he attempts to do them. So the job of the sensation in Sellars isto not, �rst and foremost, reference saving, but is instead to put the veridicalperceiver into contact with a transcendent object, say, a red brick or pink icecube. In a crucial sense, the something, somehow (pink) in the perception is notpresent to the veridical perceiver in Manifest moments of veridical perception.The pink ice cube is. That is what is demonstrated�what is present, the pinkice cube, the really-and-for-true one. The something, somehow (pink) presentsitself in the reduced perception, not in the perception as such.1 The something,somehow (pink) is immanent to perceptual consciousness, but not in a way thatmakes it automatically the referent of the perceptual demonstrative nor in away that prevents the sensation from presenting the perceptual object itself,because it presents itself instead (as if its ego got in the way of its attemptto praise another). To say that 'somehow' takes 'straighforwardly' as a specialcase is not to say that it does so only rarely, but rather to say that 'somehow'is, as it were, made to be 'straightforwardly'; the case is special in that it isthe very case that is to be expected, that is proper, to 'somehow'. 'Somehow'with anything else as a case rings hollow, improper�the destiny of 'somehow'has been short-circuited. A hollow 'somehow' is the sad sack sensation of thehallucinator.

2.6 `Somehow' and 'Straightforwardly': Sellars' (ManifestImage) Direct Realist Disjunctivism

Perhaps the most gnomic comment in Sellars's essay is the comment that Itake to be its punchline: the comment that somehow takes as a special case,straightforwardly. Sellars illustrates what he has in mind with the �nal embed-ded sentence of the essay.

This somehow (a red brick facing me edgewise) which is a red brickfacing me edgewise...

It is not easy to see exactly what Sellars accomplishes here. A fairly naturalreading (to the extent that anything can be natural here) is to take Sellars to besuggesting that in the case of veridical ostensible seeing the sensation su�ered bythe veridical seer is such that it serves to present the physical object seen. Thesensation is neither conceptualized as the referent of the veridical seer's `this'nor does it prevent the referent from being the really-and-for-true red brick. In

1Gupta fails to see this, and talks as if the something, somehow (pink) were simply presentin the perception, and not as present only under reduction.

15

one sense�a causal sense�the sensation mediates the perception; but it does notdo so by getting in the seer's line of sight, as it were: it is not obturation. �Itdoes not epistemically mediate the perception. Here again it is crucial to recallthat sensing somehow (a red brick facing me edgewise) is not to sense an objectwhich, while not a red brick, is adverbially red, brick, etc. But sensing somehow(a red brick facing me edgewise) may be sensing something which is a red brick:`Somehow' takes as a special case, `straightforwardly'. So, from the point ofview of the Manifest Image, and setting aside any (further) concern with theScienti�c Image, Sellars' account of perceptual consciousness leaves open roomfor what can be called an adverbial direct realism.

In fact, again considered qua the Manifest Image, it looks like Sellars' accountis a form of disjuctivism. From the perciever's place, from `the inside', she willtake herself to see a red brick facing her edgewise�or she will be very confused.And so this is how Sellars' view is supposed to work (whether succeeds ornot). From the inside, things will seem to the hallucinator as they seem tothe veridical perceiver: the hallucinator takes herself to be referring to a pinkice cube over there or to a red brick over there or to whatever over there. Butthe hallucinator's reference falls short, we might say, of the pink ice cube orof the red brick or of whatever, landing on the something, somehow (pink) or(red) or (whatever), even while all seems to the hallucinator as if the over-theretarget has been hit.

Sellars' averbial direct realism�at least it is such considered qua the ManifestImage�makes it clear how it is that Sellars can argue for the position he doesin SRPC without taking himself to undermine the achievements of EPM, and itmakes clear how it is that Sellars evades the Cartesian considerations to whichGupta sees him as prisoner. Sellars does not, after all, make the reference ofall perceptual demonstratives sensings. He does not �reconstruct perceptualjudgments to be about the subjective, mental realm� (Gupta)�or at least hedoes not reconstruct them all in that way. And not only does Sellars leaveintact his thinking about perceptual consciousness in EPM, not only does henot become prisoner to Cartesian considerations, but his adverbial direct realismallows him to do all of this while enjoying the relative ontological austerity thatnon-relational theories, like adverbial theories, enjoy over relational ones. �Doesall this actually work, though? Should we buy it? I do not know; I guess I doubtit. But we do need to see the view of SRPC for what Sellars crafted it to be,and I think that is hard to do.

2.7 Husserl's Non-Relationalism

Although there is no doubt that Sellars plays to a degree idiosyncatically withphenomenological categories in the essay, I submit that there is also no doubtthat he takes seriously his indebtedness to phenomenology and his claim thatwhat he has to say will fall upon friendly ears, ears of phenomenologists. Mybelief in this plays and has played a role in how I understand the essay as awhole, and especially in how I understand its ending. I take the basic structureof Sellars' view to be generically similar to and to be at least in part generated

16

by the basic structure of Husserl's.Two things often missed in accounts of Husserlian phenomenology, or at

least two things often not appropriately emphasized, is the fact that Husserlearly rejects what Chisholm called the Sense Datum Fallacy (SDF) and thatHusserl can be seen, modulo a few details perhaps, as himself an early non-relationalist (as Sellars would later be). These facts are often missed because ofthe way in which commentators handle the so-called `intentional object'. Now,I am not claiming that �guring out how to account for the intentional objectin Husserl is easy, not in the least. But, if we limit ourselves to perceptualconsciousness, I believe we can begin to account for it and in a non-relationalistway, in a way that allows Husserl to reject the SDF�as he surely does, or meansto do.

But before we do that, I want to get the SDF clearly and brie�y into view.Chisholm long ago observed that to infer from the fact that someone has asensation of red that there is something red which the person senses is a fallacy,the SDF. To have a sensation of red is not to have a red sensation. Chisholmtalked of the seer as �being appeared to redly� or as �sensing redly�. So, hecharacterized acts of sensing non-relationally and �adverbially�, not in terms ofobjects sensed but rather in terms of features of the acts themselves. This is afoundational move for any non-relational account, since it prevents the undesiredrelational acount by preventing inferences from the occurrences of acts of sensingto the (gotta be) existence of ojects which are sensed. To understand why thisis so, here is a useful comment of James Cornman's.

...I have claimed that the state of having an experience of rednessis identical with the state of red-sensing which is a non-relationalstate. I maintain that having an experience of redness is analogousto dancing a waltz: it is not that there are two entities, myself and awaltz, and there is a relationship of dancing between them. Ratherthere is just myself waltz-dancing (that is, waltzing), which is anonrelational state of me.

Putting the SFD in place should make it obvious enough why it is tricky toconstrue Husserl as a non-relationalist, as a rejector of the SFD. It might seem asthough Husserl's commitment to intentional objects in consciousness manifestshis relationalism by showing that he cannot really be a rejector of the SFD. Butthis is a decidedly super�cial reading of Husserl. Husserl is a non-relationalistabout consciousness.

One useful way of appreciating non-relationalism, and of seeing that Husserlis a proponent of it, is to consider the following trenchant passage from DallasWillard's �Perceptual Realism�:

Now, it appears to me that there are at least two sound and rathersimple arguments against the view that perceiving (or, more gen-erally, cognizing or thinking) is a relation, both of which rest onfundamental features of relations and upon familiar aspects of hu-man consciousness. (1) The �rst argument goes as follows: It very

17

often happens that what I think of, or am conscious of (i.e., what Iperceive, imagine, believe and the like), does not exist. This must beso, for otherwise, I would only have to wish for something to produceit. Wishes would be horses...for beggars to ride...But now I cannotbe related to what does not exist. From Rab it follows that (Vx)Rax, so that Rab & -(Vx) Rax is a self-contradiction. Consequently,I am often conscious of that to which I cannot be related, whichcould not be so if consciousness were a relation. (2) The secondargument is: a has relation R to be i� b has the converse relationof R to a. That is, from Rab it follows that Fba, where F is theconverse of R. The sentence, Rab & (VU) -Uba, is, therefore a self-contradiction...But, if this is so, then which I see this paper there isno relation belonging to it and consisting of its being seen by me; forto be seen by me is not a fact about the paper but is wholly a factabout me. To �nd out whether or not the table is seen by me, I, andnot the table, must be examined...To treat my seeing the table...asbeing [a relation] is, therefore, tantamount to claiming that thereare relations with no converse, which would be absurd.

Without deciding whether we should accept what Willard says, I want to useit to frame the fact that Husserl's phenomenology not only is non-relationalist,but that its very possibility rests on its being non-relationalist. To simplify abit, it is hard to see how phenomenological reduction as Husserl understoodit could be possible if Husserl construed consciousness relationally. When wereduce, we bracket out the transcendent object of cognition�it is put out ofplay, falls away. But its doing so would, if consciousness were a relation, destroythe very relational state that is the datum clari�candum of the reduction. If,instead, consciousness is understood non-relationally, then the possibility (whichHusserl takes himself to have shown an actuality) of the reduction becomesunderstandable. Consciousness of something or other is a non-relational stateof mine, of me (my intending is just me waltz-dancing), and so my state is opento phenomenological study qua itself even when its transcendent object fallsaway, and even if, as in hallucination, the transcendent object does not exist.This is what Husserl is trying above all to make clear, I submit, in the otherwisewholly bizarre section 49of Ideas, �Absolute Consciousness as the Residuum afterthe Nulli�cation of the World�. Now, I will not here try to provide a readingof that section�an All-Day Philosophical Sucker if ever there was one�but I dowant to treat its title as o�ering tuition. How could we have consciousness asa residuum after the nulli�cation of the world�if consciousness is a relation?We cannot. Willard's arguments seem to show that to be absurd. But Husserldoes not take it to be absurd; and I submit that is because he does not thinkof consciousness as a relation, a relation to the world.

(An aside for some of you who know Husserl's text. Yes, he often enoughuses the word 'relation'. But we need to remember, �rst, that it is possible touse that word without actually intending for its use to be such that, on that use,the instancing of a relation requires the existence of all of the relation's �terms�.

18

Perhaps that will seem incautious, if not intentionally misleading�but we allknow that philosophers often force words to carry peculiar crosses. Perhapsthat is what Husserl is doing here. But, let me also add, that Husserl reliesheavily and with considerable ceremony on contextual de�nitions of his crucialterms, often using a word in what he knows is a decontextualized misleadingway, but one that he thinks not misleading when contextualized. And �nally, itis worth keeping in mind that many times when Husserl is talking of relations heis talking in the natural attitude, where perhaps for him such talk is appropriatein certain ways, after all.)

One more quick thought about Husserl's Section 49. That section containsa notoriously Monadological description of the consciousness, a more Leibizian-than-Leibniz sort of passage, in which Husserl claims that consciousness is asystem of being into which nothing can penetrate and from which nothing canescape�and so on. But if we bear in mind Husserl's non-relationalism, this,while perhaps still seeming outrageous, can seem coherently outrageous. (Andanyway, we should also recall that Leibniz himself, in his Discourse on Meta-physics (28), makes clear that he takes seeing, say, non-relationally. That is notto say that he denies that our senses enjoy some causal commerce with objects,but it is to say that that causal commerce is not the seeing.)

The point here is that we need to see Husserl as a non-relationalist, despitehis talk of intentional objects. Intentional objects are not relata in a relationship.That realization is perhaps the �rst and most crucial realization in readingHusserl.

Simplifying, we can understand what is immanent in perception to be whatHusserl calls a noema. Now, noema are themselves mightily mysterious, andtheir mysteries shift, if you read Husserl in the United States, depending onwhich coast you favor, east or west. I want to avoid as much of that debateas I can. Instead, I want to fasten on a few features of the noema as Husserldescribes them in Ideas. Consider the following passage.

From the natural standpoint [a blossoming apple-tree] is somethingthat exists in the transcendent reality of space, and the perception[of it] as well as the pleasure [we take in it is] a psychical state withwe enjoy as real human beings...

Let us now pass over to the phenomenological standpoint. The tran-scendent world enters its �bracket�; in respect of its real being weuse the disconnecting epoche...

Here in regard to the perception, and also to any arbitrarily contin-ued nexus of perceptions (e.g., if we were to observe the blossomingtree ambulando), we have no such question to put as to whetheranything corresponds to it in �the� real world. This posited reality,if our judgment is to be the measure of it, is simply not there for us.

And yet everything remains, so to speak, as of old. Even the phe-nomenologically reduced perceptual experience is a perception of�this apple-tree in bloom, in this garden. . . �. The tree has not for-

19

feited the least shade [a rare Husserlian pun?] of content from allthe phases, qualities, characters with which it appeared in this per-ception. . .

[T]he described itself, though it �gures as �the same exactly�, isstill something radically other than it was, in virtue, so to speak, ofan inverting change of signature. �In� the reduced perception. . . we�nd, as belonging to its essence indissolubly, the perceived as such,and under such titles as �material thing�, �plant�, �tree�, �blossom-ing�, and so forth. The inverted commas are clearly signi�cant; theyexpress that change of signature, the corresponding radical modi�-cation of the meaning of the words. The tree plain and simple, thething in nature, is as di�erent as it can be from this perceived treeas such, which as perceptual meaning belongs to the perception, andthat inseparably. The tree plain and simple can burn away, resolveitself into chemical elements, and so forth. But the meaning�themeaning of this perception, something that belongs necessarily to itsessence�cannot burn away; it has no chemical elements, no forces,no real properties.

One thing we need to continue to keep in mind is that, for Husserl, to shiftinto the phenomenological attitude, to put out of play or reduce, is not to lowerthe veil of ideas, to, as it were, usher into view something that is immanentand is red when the experience being reduced is an experience of a red thing,a stop sign, for example. Shifting into the phenomenological attitude does notmake it the case that we are now seeing something other than the apple-tree,some new object of our act of seeing�but it does mean that we are now ableto consider our seeing in a new way, since we can now consider that throughwhich our consciousness of, our reference to, the apple-tree passes; a somethingthat is present, in a particular sense, as background in our natural attitude, andbecomes foregrounded in the shift into the phenomenological standpoint. Butwhat is foregrounded is not something that itself bears the adjectival propertieswe also ascribe to the apple-tree. Husserl will have no truck with the SFD.Moreover, even though Husserl allows himself to talk of both the tree plain andsimple and the perceived tree as such, it is crucial to remember that he is notcounting two trees, or even two things that look like trees, one because it is, theother because of its relationship to the one that is. No, there is one tree and onenoema of a tree, a meaning of a tree. The second is not an in�ammable tree,but is rather in�ammable because it is the perceptual meaning of a tree. AndHusserl is not here counting two di�erent objects of an act, one transcendent,the other immanent. No, what is immanent, the noemata, is not an object inthe relevant sense at all.

Much that we will say, when we begin our phenomenological description ofthe perception of the apple-tree, will be very much like what we would have saiddescriptively in the natural standpoint. But the phenomenologically descriptivelanguage will have �undergone a radical modi�cation of meaning�.

I take it that the parallels with Sellars here are reasonably clear, even if

20

working them out in detail would be a trouble-causing and time-consumingexercise. I am not going to work out the detail, since my main point in talkingabout Husserl, and indeed my point in the various phenomenological comments Ihave made along the way, is that Sellars thinking about perceptual consciousnessis guided in important ways by phenomenological, Husserlian lessons. While notstraightforwardly adverbialist, Husserl's non-relational account of perceptualconsciousness clearly involves something analogous to Sellars' transcategorialanalogy between the adjectival and the adverbial. Husserl's noema of the apple-tree is not green, but it is �green�. We can call this, borrowing a term from J.M. Mohanty, Husserl's �phenomenological ascent�. The adjectival green of thetree trans�gures into the noematic �green� of the reduced perception. This iswhat Husserl calls the inverting change of signature, a radical modi�cation ofmeaning. Crucially, the only thing green, adjectivally green, is the apple-tree,not its noema. The noema functions as the referential conduit to the apple-tree,if there is an apple-tree there. But that function remains, even if there is notan apple-tree there.

.....�...[A]ccording to Brentano (at least late in his career) this directedness is

not a relation to the object. Rather, it is a non-relational property of the subjectthat bears some similarities to relations and is to that extent `relational-ish,' asBrentano puts it.�

�To that extent, the fact that the experience or phenomenology of beingin an intentional state involves a phenomenology of standing in a relation tosomething does not tell against a non-relational account of intentionality. Thenon-relational account can readily admit that intentional states involve sucha phenomenology but insist that this phenomenology need not be veridical.From a phenomenological standpoint, then, we can describe intentionality aspurported directedness or purported aboutness. The idea is that intentionalitypurports to involve a directedness or aboutness relation to something, insofaras it involves a phenomenology as of standing in such a relation. This does notimply actually standing in such a relation, since having the phenomenology doesnot require it being veridical.�

Of course, it is also clear that despite the parallels between Husserl andSellars, there are also deep di�erences.

When we move into the phenomenological standpoint, we suspend the dog-matism of the natural standpoint: but we are still reducing the perception ofthe apple-tree, the noema involved is the noema of the perception of the apple-tree. We cannot be focused phenomenologically on the right noema, we mightsay, if the apple-tree were not understood as the transcendent object. (Noemaare individuated by their �objects�, by what we intend through them.) So donot let our further talk of the noema as the object of the phenomenologicalstandpoint make you think that phenomenological description is description ofdenuded noemata, noemata pruned of any directedness to the �object� they areof.

(As I understand Sellars, there is a tension between his Scienti�c Imagetheorizing and his Manifest Image re�ection, a tension that shows itself with

21

particular urgency when he tries to see perceptual consciousness stereoscopically.I have just been trying to make clear the outside-in character of Sellars' re�ectionon MI-perception. I fear though that Sellars SI-perception theorizing worksinside-out. This makes keeping a clear and resolute picture of the job of thesensation extremely di�cult. The reference-saving job of the sensation and itsobject-presenting job are not easily simultaneously dischargeable. DEVELOP.)

We seem to face a dilemma of sorts here. If the perceptual demonstrativeof the veridical perceiver refers to the pink ice cube, and that is because itis the really-and-for-true pink ice cube that causes it, then why wouldn't theperceptual demonstrative of the hallucinator refer to drugs or to her injury, atany rate to whatever it is, other than the really-and-for-true pink ice cube, thatcauses her sensation? In other words, it seems we should expect a uni�ed storyabout the referent of the perceptual demonstrative, one that makes the referentin each case the same or suitably similar. Either the perceptual demonstrativeof the perciever and the hallucinator both refer to the sensation (for whateverreason) or the perceptual demonstrative refers to its cause in both cases, eventhough the causes are not identical.

22