Should we reject SDT in favour of Intentionalist Representationalism(on Jackson, Byrne and others)?

Defining intentionalism

Alex Byrne has argued that sense datum theory is compatible with intentionalist representationalism ("Intentionalism Defended", ... ). He introduces intentionalism as that view that:

... the sensational component of a perceptual experience cannot vary independently of its intentional component: the phenomenal character of a perceptual experience is entirely determined by the experience's propositional content-that is, by what it represents. 199

Byrne offers some variant summary statements of the position as follows:

All mental facts are representational facts. (Dretske 1995, xiii) 10

Phenomenal character (or what it is like) is one and the same as a certain sort of intentional content. (Tye 1995, 137)

[T] he mind has no special properties that are not exhausted by its representational properties, along with or in combination with the functional organization of its components. (Lycan, 1996b, 11)

He observes:

Despite their differences, there is a basic claim that all these philosophers wish to defend. It is that the propositional content of perceptual experiences in a particular modality (for example, vision) determines their phenomenal character. In other words: there can be no difference in phenomenal character without a difference in content. So if two (metaphysically possible) visual experiences differ in phenomenal character, then they differ in content. 204

The impossibility of the existence of experiences with divergent phenomenologies but identical intentional contents is known as the ‘supervenience’ of phenomenology on intentional content.

We see, therefore, two important components of the position: one is that mental factors are representations, and the other that phenomenal character is determined by what is represented as being there. We shall return to the issue of what it means to be a representation.

Byrne’s Argument Sketch

Noting that the claim that intentionalism and sense datum theory are compatible sets him apart from many intentionalists, Byrne briefly states his argument as follows:

It is true that defenders of intentionalism are no friends of sense-data and, moreover, sometimes they give the impression that an intentionalist cannot also be a sense-datum theorist.

Indeed, for many supporters of intentionalism, much of the appeal of the theory is that it offers a means of avoiding items like sense data that they judge to be problematic metaphysically, epistemologically, and explanatorily.

As Byrne himself has noted (p. 212, n. 22) other philosophers have argued on grounds that make intentionalism as they understand it and SDT irreconcilable. Their arguments have assumed the so-called 'diaphaneity of experience'.

G.E. Moore originally introduced the notion of diaphaneity as the near-undetectability in conscious awareness of the relation being-conscious-of, holding between a sensation and its object. Many proponents of intentionalism, and others who discuss diaphaneity in the modern literature, have used the same terminology for a logically distinct notion. It is that what is given in our phenomenology is a way the external world is or could be (give or take a few phosphenes, after-images, and migraine castellations, for example). Such a view expressly asserts what SDT denies. If the appeal to such "latter-day" diaphaneity is essential for the cogency of arguments for intentionalism, then SDT and intentionalism are incompatible.

Byrne does not fall into the Diaphaneity Intentionalist camp; for him, core Intentionalism is more permissive. Indeed, the position puts little or no constraint on content at all. How far does the permissiveness extend? Byrne's own position is that the true Intentionalism is one where the subject is aware of a world outside our bodies, the world of distal stimuli – what he calls “environmental” Intentionalism, in common with Diaphaneity Intentionalists. But Byrne's key point is that the barebones position, that intentional content determines phenomenal character, lays down no special constraint on what the intentional content is. The content could be as of the external world or as of sense data, and embracing intentionalism by itself does not imply the rejection of SDT. As he states,

... intentionalism and the sense-datum theory are perfectly compatible. A red' sense-datum seems or appears red' . So it is represented as red' . The sense-datum theorist simply has a strange view about the content of experience and any view about the content of experience is compatible with intentionalism. 224-225. [The reader may, for present purposes, read “red’” as “red” . – JMN]

On this view, if sense datum theorists' account of perceptual experience appeals to the notion of a (visual, say) sense datum's looking a certain way, then it follows that the experience represents it as being that way.

In his sketch, Byrne seems to assume that an experience's being a vehicle for intentional content is enough for intentionalism to be satisfied. Supervenience is not quite front and centre. Perhaps he intended, with one fell swoop, to establish both the key elements, first, that something's looking a certain way implies that there is representation of its being that way and, second, that its looks (phenomenal character) supervene on the content of the representation. How the supervenience is implied is not made explicit. For that matter, how representation follows from looks is not made explicit either.

Given that Byrne presents a suggestive argument that there is no inconsistency between intentional representationalism and sense datum theory, it will be useful for us to consider the paraticular reasons given by Frank Jackson for ditching SDT in favour of IR. We shall return to Byrne's argument later, below.

Should one ditch SDT for IR?

Frank Jackson, a prominent advocate of SDT in his Perception: A Representative Theory (1977) (henceforth, PART), came to reject the theory and took up Intentional Representationalism instead. His stated grounds for the move are of intrinsic interest and they also provide a point of reference from which to consider claims that SDT and IR are compatible, including Byrne's alleged reconciliation outlined above.

Jackson notes that SDT relies on an act-object relation where the object is a sense datum which instantiates the apparent properties. He makes the following comparison of SDT and IR

Both theories see the nature of experience as lying in the properties of the objects of experience, with the big difference that,
for representationalists, the properties of the objects of experience reside in the way that
experience represents things as being. There need be nothing actually having the properties; the ‘objects’ are intentional objects. (“Some Reflections on Representationalism”, Draft at 5 April 2000 for NYU seminar on 25 April 2000., click here, downloaded 25 January 2017 3:47PM.)

SDT -- "a classic piece of buck passing"?

Jackson's dissatisfaction with SDT is that "it fails to capture the representational nature of perceptual experience. ...". There are two complaints that Jackson levels at SDT in this respect.

The "central objection" is that SDT constitutes "a classic piece of buck passing". When one experiences something that looks red, he has proposed, "one represents that there is, in the world, something red in front of one" because it is "definitive of perceptual experience that it represents how our world is".

With hindsight from 2000, Jackson insists that the SDT analysis of a visual experience of a red external object-- which rests on 'seeing' a sense datum which is red -- fails to capture the experience's representational character. The problem, says Jackson, is that:

the key relation of direct awareness to something mental must be explicated as involving representing that there is something red in the world, and we have gained nothing. It is like analysing knowledge as belief that constitutes knowledge.

So, I suppose, seeing a red tomato must be to represent it. And, what it takes to do that is to see a suitable sense datum -- with Jackson's insistent hindsight, it must be to represent a sense datum. Represention is explained by representation. The difficulty is that this elucidation is circular, and therefore fails to achieve the essential thing.

I shall argue, below, that Jackson's own sense datum theory is free of circularity, but before making that argument, I wish to challenge the view that being guilty of circularity of the kind that Jackson identifies should always be taken to be a significant difficulty.

Familiar “compound” or “nested” representations

I understand Jackson's 'buck-passing' criticism to assume that the point of appealing to sense data is to give an account of our perception of distal stimuli. So, to explain our having an experience of a red tomato SDT invokes our awareness of a red sense datum. Representing that there is something red in the world (the tomato) is explained by representing that there is something red in the world (the sense datum).

But is there a logical difficulty in envisaging an experience of distal objects and properties that mobilizes intentional content about both a sense datum and, derivatively, the external object? The idea that representations may be ‘compound’ or ‘nested’ is familiar to common sense. Consider using cutlery, cruets, knapkins, and fragments of dinner roll at a table in order to represent the configurations of units and terrain at the battle of El-Alamein. It does not undermine the adequacy of the representation of the battle by the items available on the dining table to note that the viewers have to have visual representations of those knives, cruets, and knapkins. Represented things are not, by virtue of their being represented, disqualified from serving as representations in their own right. To complain that we have no adequate account of how the battle is represented because we have had to presuppose a prior representation of cutlery, cruet, and so on, seems to be not to the point.

Analogously, in the experience of the red tomato, the awareness of and putative representation of a sense datum has a crucial role for the experience's representation of the tomato, according to the SDT account. Must we say that the representation of the tomato fails because there is representation of the sense datum?

Note that the Newtonian explanation of the gravity of a terrestial body by appeal to the gravity of its microscopic parts is not vitiated by gravity's explaining gravity. The point of the explanation is not to explain the essential nature of gravity but to clarify the details of its operation and show how the new information constrains the phenomena of falling bodies, planetary motions in longitude and so on. Newton recognized that any explanation of what the essence of gravity is would involve attending to causes that are "occult" and, at that time, were beyond the scope of authoritative science. The theory of gravitation did not pretend to explain gravity, only to show how it applied in the relevant phenomena.

Similarly, SDT's picture of the visual perception of the red tomato, contrary to Jackson's supposition, need not be proposed to explain what representation is essentially. It may, rather, set out the fine structure of a hierarchy in which awareness of the sense datum, in the context of certain other functional constraints, enables the sense datum to serve as a proxy for the red tomato -- the ultimate object of the experience. That the awareness of the sense datum may imply a certain sort of representation, as Jackson -- and Byrne -- insist, does not undermine the explanatory force of the picture. Explanations which involve "compound" or "nested" representations do not, thereby, vitiate themselves.

Leaving these general considerations to one side for the present, let us quickly look at Jackson's version of SDT (henceforth, JSDT) and see if there are circularities of the kind that Jackson now finds problematic.

Is Jackson's JSDT (PART) guilty of circularity as he later charged?

Is JSDT committed to a circularity of representation, and, if not of representation, then of seeing, or anything else? Certainly, SDT is often thought to assume that seeing external things is to be explained by seeing seeing sense data. Michael Tye has aptly captured this conception in the following terms:

You see various objects; and you see these objects by seeing their facing surfaces. Sense-datum theorists claimed that the facing surfaces of the objects are themselves seen by seeing further immaterial surfaces or sense data. (Emphasis added. C, C, and C, pp. 45-46)

It is also not uncommon to see SDT described as explaining "indirectly" seeing external things by "directly" seeing sense data. That legitimately suggests questions about circularity if univocality can be assumed for seeing. And, indeed, at first sight, Jackson's formulation of JSDT conforms with see/see Tye's description. As we see below, JSDT seems to explain seeing distal stimuli by seeing sense data, and one is compelled to ask whether this implies that there is a pernicious circularity. Further, if there were such a circularity, and if seeing were assumed to be representational, then Jackson's complaint about 'buck-passing' would hold up. However, I shall argue that, despite the appearances, we find no circularity of seeing in JTSD. and hence no circularity of representation to be established by that route.

Is Jackson's JSDT "buck-passing"?

Let us take as our starting point Jackson's contrast between immediate and mediate objects of perception. Mediate objects are those that are perceived in virtue of the subject's perceiving immediate objects of perception that belong to the mediate ones. Jackson explains in virtue of by appeal to a selection of paradigm cases such as "my car being red in virtue of its body being red" and "someone's living in Australia in virtue of living in Melbourne". Belonging has as a necessary condition that the mediate object -- a material object in the external world --and the immediate object -- a mental sense datum --have a causal link satisfying certain conditions. The link is not directly between the objects, but rather between an event implicating the material object and the event of having the sense datum. To determine a sufficient condition, to supplement the above necessary condition for belonging we must fix an “essential feature” of the material object event, which Jackson proposes to be that it causes the having of the sense datum ‘in the normal way’, although he gives no general conditions for normality. Ultimately, for Jackson, belonging embodies a “functional spatial dependence of the sense-datum on the object, a dependence which is consequent on the causal connexion between the object and the sense-datum”. (170)

Jackson summarizes the key elements as follows:

we noted that opaque physical objects of reasonable volume are not immediate objects of perception. They are seen in virtue of seeing something else, and we analysed this situation in terms of a relation between the physical object and the something else. We called this relation that of belonging to and the position can be summed up, using our definition of sense-datum, by saying that seeing something physical may be analysed as immediateley seeing a sense-datum which belongs to that physical object. That is, where X is a physical object, 'S sees X' is analysed as 'S immedately sees a sense-datum which belongs to X.' Moreover, on the SDT, the immediate object bears the apparent properties, so that 'X looks F to S' is analysed as 'S immediately sees a sense-datum which is F and which belongs to X.' (101)

Now, nowhere in his text is there a suggestion that any of this implies or suggests a concern with issues of representation, other than in the title: Perception: A Representative Theory. No account of representation per se appears to be presupposed, and there is no entry for 'representation' in the -- admittedly, rather thin -- index. (The only use of a related term that I can find is at page 35 and is in a quotation from Hume. This may be the only explicit appeal to the concept of representation in book.) So, it does not look as if Jackson's analysis commits any circularity in the application of an explicit concept of representation. Nevertheless, Jackson's complaint that SDT doesn't adequately capture the concept of representation seems warranted. It does not explicitly address it at all.

Whether or not circularity arises implicitly in his analysis of the concept of seeing mediate objects in terms of seeing immediate objects is a distinct question. Despite his appeal to the concept of seeing a material external object in virtue of seeing something else distinct from it, namely a mental object, it turns out that the appearance of circularity (of seeing) is illusory.

Jackson's account is deliberately designed to be a conceptual analysis, an exercise in a particular philosophical idiom, with explicit analysandum and explicit analysans. Now, if the analysis does its job, the analysandum should  be everywhere eliminable in terms of the analysans. What we mean or what we should mean by the analysandum is the analysans, so that language of “mediate” objects of perception must be taken to be eliminable in favour of language of “immediate” objects of perception, sense data. Let us look at how this applies to the treatment of the notion of mediate object of perception.

Jackson defines mediate object, immediate object and immediate perception as follows:

first, mediate object:

x is a mediate object of (visual) perception (for S at t) iff S sees x at t, and there is a y such that (xy and) S sees x in virtue of seeing y.

next, immediate object:

An immediate object of perception is one that is not mediate;

and, immediately perceiving:

and we can define the relation of immediately perceiving thus: y immediately perceives x and t iff x is an immediate object of perception for S at t (as just defined).(PART, 19-20)

On JSDT's analysis, we see that talk of seeing material objects in the world is (just!) talk of seeing sense data, though not just any sense data. They are from a delimited class which instantiate causally based relations of ‘belonging’ to the material objects in question. Talk of mediate seeing of cats, tables, and cars is talk of immediate seeing those sense data which are, in the appropriate ‘normal’ way, the effects of events implicating the cats, tables and cars. 'Belonging to' means 'being a certain kind of effect of'.

Consequently, there is no plausible claim of circularity in the repeated use of ‘see’. It may be argued that ‘see’ is not being used univocally in the mediate and immediate cases. That would rebut any allegation of circularity, although it would raise issues of the felicitousness of such a terminology. But, if Jackson’s analysis is to be accepted, then talk of mediate seeing is properly eliminable and questions of the meaning of “mediate” seeing of external objects are, we might say, philosophically moot. There is no literal seeing of external objects.

In JSDT, 'seeing a mediate object of perception' is parsed in a way that allows us to see that only 'seeing an immediate object of perception' is in play, making the former expression an inessential façon de parler. There is only one kind of seeing and no regress.

Jackson, not entirely unambiguously, sets out the situation as follows:

I have argued that though we see opaque, material things of reasonable volume, they are never the immediate objects of perception for a person at a time. But this is not to say that we see material objects in a different sense from that in which we see the immediate objects; it is, rather to advance as an analytical thesis that to see a reasonable-sized, opaque material object is to see something distinct from that object, the relevant immediate object of perception (whatever the ontological status of the latter may turn out to be). (PART, p. 28)

On a quick reading, Jackson seems here to be saying that expressions such as "we see opaque material things" and "see a sense datum" use "see" univocally, and the reader might suppose that there is literal "seeing" of the material object. But Jackson makes it plain that such talk in relation to material objects is talk only about sense data. Speaking literally, we must say that, in respect of external material objects, we see only the remote effects of events surrounding material objects. The reason that 'seeing' does not have different senses with 'material object' and 'sense datum' is that 'seeing' only applies to sense data.

Given that talk of seeing external material objects is, on his analysis, eliminable as I have described, it is plausible to suggest that his assumed truth conditions for talk of seeing material objects do not include being in a seeing relationship with the objects.

Jackson very transparently illustrates this plausible interpretation with his remarks on his proposed truth conditions for the relevant statements about coloured material objects and coloured sense data.

When I say that colour is not a property of material things I am not, of course, saying that every statement that some material thing is a certain colour is false. I am saying, rather, that the truth conditions of a statement like ‘That material thing is red’ are not to be given in terms of that material thing having the property of being red, but in some quite different way, in particular, in terms of the sense data normally belonging to that thing having the property of being red. (PART, p. 128, emphasis added.)

Exactly because the truth conditions for seeing a material object are given in a similar "different way" to the colour case, in terms of the sense data normally belonging to that thing being seen, rather than in terms of the material object being seen, it makes it proper to say, strictly speaking, that we do not see the material object.

In sum, in Jackson's 1977 sense datum theory issues of circularity do not arise for the concept of seeing despite there being an initial semblance of the explaining of seeing material things by ("in virtue of") the seeing of sense data. Whether it is explanation or analysis, the details reveal that the account is an eliminativist one for seeing mediate objects. I conclude that Jackson's (2000) criticism of the circularity of SDT -- that it's like "analysing knowledge as belief that constitutes knowledge" -- does not hold up for Jackson's own SDT, at least in respect of the notion of seeing. It also means that descriptions of SDT, like Tye's given above, are misleading about the meaning, if not the terminology of SDT.

If that is right, then we should conclude that SDT is not essentially circular in the way that Jackson identifies, involving a regress of representation. While it is of logical interest whether there exist sense datum theories that commit a regress in that way, the fact that at least one does not is enough to void the general criticism of SDT.

Nevertheless, even if Jackson's (2000) criticism of the circularity of JSDT is misplaced, his claim that the theory fails to capture the representational character of experience is correct, for reasons that we turn to now.



But does direct awareness even represent?

In a way that, at first sight, seems to sit rather uncomfortably with the above criticism, Jackson offers a second way in which SDT fails to capture representation in experience: invoking direct awareness of the sense datum -- the "immediate visual object" in PART's vocabulary -- "cuts no representational ice".

SDT, says Jackson, falsely supposes that

the requirement that there be something which is red of which the subject is directly aware, automatically captures, or part way captures, the key representational notion.

But, he argues, such direct awareness doesn't achieve that, because just presenting something to the perceiver doesn't achieve representation.

The problem is that simply being aware of the sense datum just lacks resources to oonstitute representation. Jackson allows that "I can represent how I am representing something to be by using the actual way something is". For example, I can show you a sample of a colour to pick out "the colour I remember the murderer’s coat to be". But the "crucial point" is that

the fact that I am using an actual sample of the colour cuts no representational ice per se. I could be using the sample to represent the one colour I do not think the murderer’s coat to be. Or I could be following the convention of holding up a sample with the colour complementary to that I remember the murder’s coat to be.

To put it crudely, someone shown a coat of a certain colour might well respond: "OK, what about it?"

As for the coat, so for the sense datum:

standing in a certain direct-awareness relationship to a mental item with such and such properties says nothing, represents nothing, per se, about how the world is. (Emphasis added)

Putting the two criticisms together, then, we see, then, that Jackson is concerned that 1) SDT's assuming that awareness of the sense datum is representation makes the account of the representational nature of experience circular 'buck-passing'. and 2) simply being aware of the sense datum "cuts no philosophical ice" and fails to constitute the representational character of perceptual experience. and 2) Although Jackson does not put the point explicitly, his two criticims make up a nasty dichotomy for SDT. If awareness of the sense datum is representation, then the the acount of the experience's representation of distal objects is regressive. If awareness of the sense datum is not representation, then SDT fails to capture the repesentational character of experience. Either way, SDT fails to capture the representational character of experience.

(to be continued)